2024
Matey-Sanz, Miguel
Human Activity Recognition with Consumer Devices and Real-Life Perspectives PhD Thesis
2024.
Abstract | Links | BibTeX | Tags: activity recognition, machine learning, smartphone app, smartwatch
@phdthesis{Matey2024c,
title = {Human Activity Recognition with Consumer Devices and Real-Life Perspectives},
author = {Miguel Matey-Sanz},
doi = {http://dx.doi.org/10.6035/14101.2024.663821},
year = {2024},
date = {2024-10-30},
abstract = {During the last decade, research on human activity recognition has grown due to its applications in diverse fields such as video surveillance, exercise monitoring or health monitoring systems. In the latter case, researchers are putting their efforts into using human activity recognition in monitoring elderly people, for example, for fall prevention and detection applications. Existing research usually has drawbacks regarding their requirements regarding sensing devices (e.g., cost, quantity, location). Therefore, research needs to keep these drawbacks in mind to have a real impact on society. This thesis addresses the abovementioned issue by focusing on the feasibility of the use of consumer devices such as smartphones and smartwatches, and cheap devices like microcontrollers, for human activity recognition and its application in real-life problems.},
keywords = {activity recognition, machine learning, smartphone app, smartwatch},
pubstate = {published},
tppubtype = {phdthesis}
}
During the last decade, research on human activity recognition has grown due to its applications in diverse fields such as video surveillance, exercise monitoring or health monitoring systems. In the latter case, researchers are putting their efforts into using human activity recognition in monitoring elderly people, for example, for fall prevention and detection applications. Existing research usually has drawbacks regarding their requirements regarding sensing devices (e.g., cost, quantity, location). Therefore, research needs to keep these drawbacks in mind to have a real impact on society. This thesis addresses the abovementioned issue by focusing on the feasibility of the use of consumer devices such as smartphones and smartwatches, and cheap devices like microcontrollers, for human activity recognition and its application in real-life problems.
Matey-Sanz, Miguel; González-Pérez, Alberto; Casteleyn, Sven; Granell-Canut, Carlos
Implementing and Evaluating the Timed Up and Go Test Automation Using Smartphones and Smartwatches Journal Article
In: IEEE Journal of Biomedical and Health Informatics, vol. 28, iss. 11, pp. 6594 - 6605, 2024, ISSN: 2168-2208.
Abstract | Links | BibTeX | Tags: activity recognition, machine learning, Mobile apps, symptoms, wearables
@article{Matey2024b,
title = {Implementing and Evaluating the Timed Up and Go Test Automation Using Smartphones and Smartwatches},
author = {Miguel Matey-Sanz and Alberto González-Pérez and Sven Casteleyn and Carlos Granell-Canut},
doi = {https://doi.org/10.1109/JBHI.2024.3456169},
issn = {2168-2208},
year = {2024},
date = {2024-09-09},
urldate = {2024-09-09},
journal = {IEEE Journal of Biomedical and Health Informatics},
volume = {28},
issue = {11},
pages = {6594 - 6605},
abstract = {Physical performance tests aim to assess the physical abilities and mobility skills of individuals for various healthcare purposes. They are often driven by experts and usually performed at their practice, and therefore they are resource-intensive and time-demanding. For tests based on objective measurements (e.g., duration, repetitions), technology can be used to automate them, allowing the patients to perform the test themselves, more frequently and anywhere, while alleviating the expert from supervising the test. The well-known Timed Up and Go (TUG) test, typically used for mobility assessment, is an ideal candidate for automation, as inertial sensors (among others) can be deployed to detect the various movements constituting the test without expert supervision. To move from expert-led testing to self-administered testing, we present a mHealth system capable of automating the TUG test using a pocket-sized smartphone or a wrist smartwatch paired with a smartphone, where data from inertial sensors are used to detect the activities carried out by the patient while performing the test and compute their results in real time. All processing (i.e., data processing, machine learning-based activity inference, results calculation) takes place on the smartphone. The use of both devices to automate the TUG test was evaluated (w.r.t. accuracy, reliability and battery consumption) and mutually compared, and set off with a reference method, obtaining excellent Bland-Altman agreement results and Intraclass Correlation Coefficient reliability. Results also suggest that the smartwatch-based system performs better than the smartphone-based system.},
keywords = {activity recognition, machine learning, Mobile apps, symptoms, wearables},
pubstate = {published},
tppubtype = {article}
}
Physical performance tests aim to assess the physical abilities and mobility skills of individuals for various healthcare purposes. They are often driven by experts and usually performed at their practice, and therefore they are resource-intensive and time-demanding. For tests based on objective measurements (e.g., duration, repetitions), technology can be used to automate them, allowing the patients to perform the test themselves, more frequently and anywhere, while alleviating the expert from supervising the test. The well-known Timed Up and Go (TUG) test, typically used for mobility assessment, is an ideal candidate for automation, as inertial sensors (among others) can be deployed to detect the various movements constituting the test without expert supervision. To move from expert-led testing to self-administered testing, we present a mHealth system capable of automating the TUG test using a pocket-sized smartphone or a wrist smartwatch paired with a smartphone, where data from inertial sensors are used to detect the activities carried out by the patient while performing the test and compute their results in real time. All processing (i.e., data processing, machine learning-based activity inference, results calculation) takes place on the smartphone. The use of both devices to automate the TUG test was evaluated (w.r.t. accuracy, reliability and battery consumption) and mutually compared, and set off with a reference method, obtaining excellent Bland-Altman agreement results and Intraclass Correlation Coefficient reliability. Results also suggest that the smartwatch-based system performs better than the smartphone-based system.
Novak, Robert; Delgado, Marcos; García-Sipols, Ana Elizabeth; Trilles-Oliver, Sergio; de Blas, Clara Simón; Gallego, Micael; Rodríguez-Sánchez, Maria Cristina
A Real-Time Framework for Enhancing Emergency Response Effectiveness in Firefighting Contexts Proceedings Article
In: Seminario Anual de Automática, Electrónica Industrial e Instrumentación, pp. 1-7, Granada, 2024.
Abstract | BibTeX | Tags: machine learning, Real time analysis, Smart Cities
@inproceedings{Novak2024a,
title = {A Real-Time Framework for Enhancing Emergency Response Effectiveness in Firefighting Contexts},
author = {Robert Novak and Marcos Delgado and Ana Elizabeth García-Sipols and Sergio Trilles-Oliver and Clara Simón de Blas and Micael Gallego and Maria Cristina Rodríguez-Sánchez},
year = {2024},
date = {2024-07-03},
urldate = {2024-07-03},
booktitle = {Seminario Anual de Automática, Electrónica Industrial e Instrumentación},
pages = {1-7},
address = {Granada},
abstract = {This study presents a real-time framework designed to enhance emergency response effectiveness, initially applied in firefighting contexts but potentially generalizable to other emergency scenarios. Integrating advanced sensors with a comprehensive mathematical framework significantly enhances immediate situational awareness and substantially improves operational decision-making. Deployed and tested at Fire Station 9 in Chamartín, the system utilizes strategically placed sensors with variable transmission rates to simulate diverse emergency scenarios. The core achievement of this research is the demonstration of the framework’s capacity to provide real-time predictions, enabling emergency responders to act swiftly and accurately in dynamic situations. The results highlight the significant potential of real-time data analytics in revolutionizing emergency response strategies, offering a path towards safer and more efficient firefighting operations.},
keywords = {machine learning, Real time analysis, Smart Cities},
pubstate = {published},
tppubtype = {inproceedings}
}
This study presents a real-time framework designed to enhance emergency response effectiveness, initially applied in firefighting contexts but potentially generalizable to other emergency scenarios. Integrating advanced sensors with a comprehensive mathematical framework significantly enhances immediate situational awareness and substantially improves operational decision-making. Deployed and tested at Fire Station 9 in Chamartín, the system utilizes strategically placed sensors with variable transmission rates to simulate diverse emergency scenarios. The core achievement of this research is the demonstration of the framework’s capacity to provide real-time predictions, enabling emergency responders to act swiftly and accurately in dynamic situations. The results highlight the significant potential of real-time data analytics in revolutionizing emergency response strategies, offering a path towards safer and more efficient firefighting operations.
Macias, Juan Emilio Zurita; Trilles-Oliver, Sergio
Machine learning-based prediction model for battery levels in IoT devices using meteorological variables Journal Article
In: Internet of Things, vol. 25, pp. 101109, 2024, ISSN: 2542-6605.
Abstract | Links | BibTeX | Tags: battery level prediction, Internet of things, machine learning
@article{Zurita2024a,
title = {Machine learning-based prediction model for battery levels in IoT devices using meteorological variables},
author = {Juan Emilio Zurita Macias and Sergio Trilles-Oliver},
doi = {10.1016/j.iot.2024.101109},
issn = {2542-6605},
year = {2024},
date = {2024-04-01},
journal = {Internet of Things},
volume = {25},
pages = {101109},
abstract = {Efficient energy management is vital for the sustainability of IoT devices employing solar harvesting systems, particularly to circumvent battery depletion during periods of diminished solar incidence. Embracing the structured methodology of CRISP-DM, this study introduces machine learning (ML) models that utilise meteorological data to predict battery charge levels in solar-powered IoT devices. These models enable proactive adjustments to the devices’ data sampling frequencies, ensuring effective energy utilisation. The proposed ML models were evaluated using authentic battery charge data and weather forecast records. The empirical results of this study corroborate the predictive prowess of the models, with an average accuracy reaching as high as 94.09% in specific test cases. This substantiates the potential of the developed methodology to significantly enhance the energy autonomy of IoT devices through predictive analytics.},
keywords = {battery level prediction, Internet of things, machine learning},
pubstate = {published},
tppubtype = {article}
}
Efficient energy management is vital for the sustainability of IoT devices employing solar harvesting systems, particularly to circumvent battery depletion during periods of diminished solar incidence. Embracing the structured methodology of CRISP-DM, this study introduces machine learning (ML) models that utilise meteorological data to predict battery charge levels in solar-powered IoT devices. These models enable proactive adjustments to the devices’ data sampling frequencies, ensuring effective energy utilisation. The proposed ML models were evaluated using authentic battery charge data and weather forecast records. The empirical results of this study corroborate the predictive prowess of the models, with an average accuracy reaching as high as 94.09% in specific test cases. This substantiates the potential of the developed methodology to significantly enhance the energy autonomy of IoT devices through predictive analytics.
Klus, Lucie; Klus, Roman; Torres-Sospedra, Joaquín; Lohan, Elena Simona; Granell-Canut, Carlos; Nurmi, Jari
EWOk: Towards Efficient Multidimensional Compression of Indoor Positioning Datasets Journal Article
In: IEEE Transactions on Mobile Computing, vol. 25, no. 5, pp. 3589-3604, 2024, ISSN: 1558-0660.
Abstract | Links | BibTeX | Tags: A-wear, machine learning, prediction algorithms, Wi-Fi fingerprint
@article{Klus2024a,
title = {EWOk: Towards Efficient Multidimensional Compression of Indoor Positioning Datasets},
author = {Lucie Klus and Roman Klus and Joaquín Torres-Sospedra and Elena Simona Lohan and Carlos Granell-Canut and Jari Nurmi},
doi = {10.1109/TMC.2023.3277333},
issn = {1558-0660},
year = {2024},
date = {2024-03-01},
journal = {IEEE Transactions on Mobile Computing},
volume = {25},
number = {5},
pages = {3589-3604},
abstract = {Indoor positioning performed directly at the end-user device ensures reliability in case the network connection fails but is limited by the size of the RSS radio map necessary to match the measured array to the device’s location. Reducing the size of the RSS database enables faster processing, and saves storage space and radio resources necessary for the database transfer, thus cutting implementation and operation costs, and increasing the quality of service. In this work, we propose EWOk, an Element-Wise cOmpression using k-means, which reduces the size of the individual radio measurements within the fingerprinting radio map while sustaining or boosting the dataset’s positioning capabilities. We show that the 7-bit representation of measurements is sufficient in positioning scenarios, and reducing the data size further using EWOk results in higher compression and faster data transfer and processing. To eliminate the inherent uncertainty of k-means we propose a data-dependent, non-random initiation scheme to ensure stability and limit variance. We further combine EWOk with principal component analysis to show its applicability in combination with other methods, and to demonstrate the efficiency of the resulting multidimensional compression. We evaluate EWOk on 25 RSS fingerprinting datasets and show that it positively impacts compression efficiency, and positioning performance.},
keywords = {A-wear, machine learning, prediction algorithms, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {article}
}
Indoor positioning performed directly at the end-user device ensures reliability in case the network connection fails but is limited by the size of the RSS radio map necessary to match the measured array to the device’s location. Reducing the size of the RSS database enables faster processing, and saves storage space and radio resources necessary for the database transfer, thus cutting implementation and operation costs, and increasing the quality of service. In this work, we propose EWOk, an Element-Wise cOmpression using k-means, which reduces the size of the individual radio measurements within the fingerprinting radio map while sustaining or boosting the dataset’s positioning capabilities. We show that the 7-bit representation of measurements is sufficient in positioning scenarios, and reducing the data size further using EWOk results in higher compression and faster data transfer and processing. To eliminate the inherent uncertainty of k-means we propose a data-dependent, non-random initiation scheme to ensure stability and limit variance. We further combine EWOk with principal component analysis to show its applicability in combination with other methods, and to demonstrate the efficiency of the resulting multidimensional compression. We evaluate EWOk on 25 RSS fingerprinting datasets and show that it positively impacts compression efficiency, and positioning performance.
2023
Bravenec, Tomás
Exploiting Wireless Communications for Localization: Beyond Fingerprinting PhD Thesis
Universitat Jaume I. INIT, 2023.
Abstract | Links | BibTeX | Tags: A-wear, data analysis methods, geoprivacy, Indoor positioning, machine learning
@phdthesis{Bravenec2023d,
title = {Exploiting Wireless Communications for Localization: Beyond Fingerprinting},
author = {Tomás Bravenec},
url = {http://hdl.handle.net/10803/689593},
doi = {http://dx.doi.org/10.6035/14124.2023.868082},
year = {2023},
date = {2023-12-18},
school = {Universitat Jaume I. INIT},
abstract = {The field of Location-based Services (LBS) has experienced significant growth over the past decade, driven by increasing interest in fitness tracking, robotics, and eHealth. This dissertation focuses on evaluating privacy measures in Indoor Positioning Systems (IPS), particularly in the context of ubiquitous Wi-Fi networks. It addresses non-cooperative user tracking through the exploitation of unencrypted Wi-Fi management frames, which contain enough information for device fingerprinting despite MAC address randomization. The research also explores an algorithm to estimate room occupancy based on passive Wi-Fi frame sniffing and Received Signal Strength Indicator (RSSI) measurements. Such room occupancy detection has implications for energy regulations in smart buildings. Furthermore, the thesis investigates methods to reduce computational requirements of machine learning and positioning algorithms through optimizing neural networks and employing interpolation techniques for IPS based on RSSI fingerprinting. The work contributes datasets, analysis scripts, and firmware to improve reproducibility and supports advancements in the LBS field.},
keywords = {A-wear, data analysis methods, geoprivacy, Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {phdthesis}
}
The field of Location-based Services (LBS) has experienced significant growth over the past decade, driven by increasing interest in fitness tracking, robotics, and eHealth. This dissertation focuses on evaluating privacy measures in Indoor Positioning Systems (IPS), particularly in the context of ubiquitous Wi-Fi networks. It addresses non-cooperative user tracking through the exploitation of unencrypted Wi-Fi management frames, which contain enough information for device fingerprinting despite MAC address randomization. The research also explores an algorithm to estimate room occupancy based on passive Wi-Fi frame sniffing and Received Signal Strength Indicator (RSSI) measurements. Such room occupancy detection has implications for energy regulations in smart buildings. Furthermore, the thesis investigates methods to reduce computational requirements of machine learning and positioning algorithms through optimizing neural networks and employing interpolation techniques for IPS based on RSSI fingerprinting. The work contributes datasets, analysis scripts, and firmware to improve reproducibility and supports advancements in the LBS field.
Matey-Sanz, Miguel; Casteleyn, Sven; Granell-Canut, Carlos
Dataset of inertial measurements of smartphones and smartwatches for human activity recognition Journal Article
In: Data in Brief, vol. 51, pp. 109809, 2023, ISSN: 2352-3409.
Abstract | Links | BibTeX | Tags: activity recognition, dataset, machine learning, smartphone app, smartwatch, symptoms
@article{Matey2023c,
title = {Dataset of inertial measurements of smartphones and smartwatches for human activity recognition},
author = {Miguel Matey-Sanz and Sven Casteleyn and Carlos Granell-Canut},
doi = {https://doi.org/10.1016/j.dib.2023.109809},
issn = {2352-3409},
year = {2023},
date = {2023-12-15},
journal = {Data in Brief},
volume = {51},
pages = {109809},
abstract = {This article describes a dataset for human activity recognition with inertial measurements, i.e., accelerometer and gyroscope, from a smartphone and a smartwatch placed in the left pocket and on the left wrist, respectively. Twenty-three heterogeneous subjects (μ = 44.3, σ = 14.3, 56% male) participated in the data collection, which consisted of performing five activities (seated, standing up, walking, turning, and sitting down) arranged in a specific sequence (corresponding with the TUG test). Subjects performed the sequence of activities multiple times while the devices collected inertial data at 100 Hz and were video-recorded by a researcher for data labelling purposes. The goal of this dataset is to provide smartphone- and smartwatch-based inertial data for human activity recognition collected from a heterogeneous (i.e., age-diverse, gender-balanced) set of subjects. Along with the dataset, the repository includes demographic information (age, gender), information about each sequence of activities (smartphone's orientation in the pocket, direction of turns), and a Python package with utility functions (data loading, visualization, etc). The dataset can be reused for different purposes in the field of human activity recognition, from cross-subject evaluation to comparison of recognition performance using data from smartphones and smartwatches.},
keywords = {activity recognition, dataset, machine learning, smartphone app, smartwatch, symptoms},
pubstate = {published},
tppubtype = {article}
}
This article describes a dataset for human activity recognition with inertial measurements, i.e., accelerometer and gyroscope, from a smartphone and a smartwatch placed in the left pocket and on the left wrist, respectively. Twenty-three heterogeneous subjects (μ = 44.3, σ = 14.3, 56% male) participated in the data collection, which consisted of performing five activities (seated, standing up, walking, turning, and sitting down) arranged in a specific sequence (corresponding with the TUG test). Subjects performed the sequence of activities multiple times while the devices collected inertial data at 100 Hz and were video-recorded by a researcher for data labelling purposes. The goal of this dataset is to provide smartphone- and smartwatch-based inertial data for human activity recognition collected from a heterogeneous (i.e., age-diverse, gender-balanced) set of subjects. Along with the dataset, the repository includes demographic information (age, gender), information about each sequence of activities (smartphone's orientation in the pocket, direction of turns), and a Python package with utility functions (data loading, visualization, etc). The dataset can be reused for different purposes in the field of human activity recognition, from cross-subject evaluation to comparison of recognition performance using data from smartphones and smartwatches.
Matey-Sanz, Miguel; Torres-Sospedra, Joaquín; González-Pérez, Alberto; Casteleyn, Sven; Granell-Canut, Carlos
Analysis and Impact of Training Set Size in Cross-Subject Human Activity Recognition Proceedings Article
In: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications, pp. 391–405, Springer, Cham, 2023, ISBN: 978-3-031-49018-7.
Abstract | Links | BibTeX | Tags: activity recognition, machine learning, smartphone app, smartwatch, symptoms
@inproceedings{Matey2023b,
title = {Analysis and Impact of Training Set Size in Cross-Subject Human Activity Recognition},
author = {Miguel Matey-Sanz and Joaquín Torres-Sospedra and Alberto González-Pérez and Sven Casteleyn and Carlos Granell-Canut},
doi = {https://doi.org/10.1007/978-3-031-49018-7_28},
isbn = {978-3-031-49018-7},
year = {2023},
date = {2023-12-01},
booktitle = {Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications},
volume = {14469},
pages = {391–405},
publisher = {Springer, Cham},
series = {Lecture Notes in Computer Science},
abstract = {The ubiquity of consumer devices with sensing and computational capabilities, such as smartphones and smartwatches, has increased interest in their use in human activity recognition for healthcare monitoring applications, among others. When developing such a system, researchers rely on input data to train recognition models. In the absence of openly available datasets that meet the model requirements, researchers face a hard and time-consuming process to decide which sensing device to use or how much data needs to be collected. In this paper, we explore the effect of the amount of training data on the performance (i.e., classification accuracy and activity-wise F1-scores) of a CNN model by performing an incremental cross-subject evaluation using data collected from a consumer smartphone and smartwatch. Systematically studying the incremental inclusion of subject data from a set of 22 training subjects, the results show that the model’s performance initially improves significantly with each addition, yet this improvement slows down the larger the number of included subjects. We compare the performance of models based on smartphone and smartwatch data. The latter option is significantly better with smaller sizes of training data, while the former outperforms with larger amounts of training data. In addition, gait-related activities show significantly better results with smartphone-collected data, while non-gait-related activities, such as standing up or sitting down, were better recognized with smartwatch-collected data.},
keywords = {activity recognition, machine learning, smartphone app, smartwatch, symptoms},
pubstate = {published},
tppubtype = {inproceedings}
}
The ubiquity of consumer devices with sensing and computational capabilities, such as smartphones and smartwatches, has increased interest in their use in human activity recognition for healthcare monitoring applications, among others. When developing such a system, researchers rely on input data to train recognition models. In the absence of openly available datasets that meet the model requirements, researchers face a hard and time-consuming process to decide which sensing device to use or how much data needs to be collected. In this paper, we explore the effect of the amount of training data on the performance (i.e., classification accuracy and activity-wise F1-scores) of a CNN model by performing an incremental cross-subject evaluation using data collected from a consumer smartphone and smartwatch. Systematically studying the incremental inclusion of subject data from a set of 22 training subjects, the results show that the model’s performance initially improves significantly with each addition, yet this improvement slows down the larger the number of included subjects. We compare the performance of models based on smartphone and smartwatch data. The latter option is significantly better with smaller sizes of training data, while the former outperforms with larger amounts of training data. In addition, gait-related activities show significantly better results with smartphone-collected data, while non-gait-related activities, such as standing up or sitting down, were better recognized with smartwatch-collected data.
Hammad, Sahibzada Saadoon; Iskandaryan, Ditsuhi; Trilles-Oliver, Sergio
An unsupervised TinyML approach applied to the detection of urban noise anomalies under the smart cities environment Journal Article
In: Internet of Things, vol. 23, pp. 100848, 2023, ISSN: 2542-6605.
Abstract | Links | BibTeX | Tags: environmental monitoring, machine learning, TinyML
@article{Saadoon2023a,
title = {An unsupervised TinyML approach applied to the detection of urban noise anomalies under the smart cities environment},
author = {Sahibzada Saadoon Hammad and Ditsuhi Iskandaryan and Sergio Trilles-Oliver},
doi = {10.1016/j.iot.2023.100848},
issn = {2542-6605},
year = {2023},
date = {2023-10-01},
journal = {Internet of Things},
volume = {23},
pages = {100848},
abstract = {Artificial Intelligence of Things (AIoT) is an emerging area of interest, and this can be used to obtain knowledge and take better decisions in the same Internet of Things (IoT) devices. IoT data are prone to anomalies due to various factors such as malfunctioning of sensors, low-cost devices, etc. Following the AIoT paradigm, this work explores anomaly detection in IoT urban noise sensor networks using a Long Short-Term Memory Autoencoder. Two autoencoder models are trained using normal data from two different sensors in the sensor network and tested for the detection of two different types of anomalies, i.e. point anomalies and collective anomalies. The results in terms of accuracy of the two models are 99.99% and 99.34%. The trained model is quantised, converted to TensorFlow Lite format and deployed on the ESP32 microcontroller (MCU). The inference time on the microcontroller is 4 ms for both models, and the power consumption of the MCU is 0.2693 W ± 0.039 and 0.3268 W ± 0.015. Heap memory consumption during the execution of the program for sensors TA120-T246187 and TA120-T246189 is 528 bytes and 744 bytes respectively.},
keywords = {environmental monitoring, machine learning, TinyML},
pubstate = {published},
tppubtype = {article}
}
Artificial Intelligence of Things (AIoT) is an emerging area of interest, and this can be used to obtain knowledge and take better decisions in the same Internet of Things (IoT) devices. IoT data are prone to anomalies due to various factors such as malfunctioning of sensors, low-cost devices, etc. Following the AIoT paradigm, this work explores anomaly detection in IoT urban noise sensor networks using a Long Short-Term Memory Autoencoder. Two autoencoder models are trained using normal data from two different sensors in the sensor network and tested for the detection of two different types of anomalies, i.e. point anomalies and collective anomalies. The results in terms of accuracy of the two models are 99.99% and 99.34%. The trained model is quantised, converted to TensorFlow Lite format and deployed on the ESP32 microcontroller (MCU). The inference time on the microcontroller is 4 ms for both models, and the power consumption of the MCU is 0.2693 W ± 0.039 and 0.3268 W ± 0.015. Heap memory consumption during the execution of the program for sensors TA120-T246187 and TA120-T246189 is 528 bytes and 744 bytes respectively.
Matey-Sanz, Miguel; Torres-Sospedra, Joaquín; Moreira, Adriano
Temporal Stability on Human Activity Recognition based on Wi-Fi CSI Proceedings Article
In: 2023 IEEE 13th International Conference on Indoor Positioning and Indoor Navigation (IPIN), pp. 1-6, IEEE, 2023, ISBN: 979-8-3503-2012-1.
Abstract | Links | BibTeX | Tags: activity recognition, CSI, machine learning
@inproceedings{Matey2023a,
title = {Temporal Stability on Human Activity Recognition based on Wi-Fi CSI},
author = {Miguel Matey-Sanz and Joaquín Torres-Sospedra and Adriano Moreira},
doi = {https://doi.org/10.1109/IPIN57070.2023.10332214},
isbn = {979-8-3503-2012-1},
year = {2023},
date = {2023-09-25},
booktitle = {2023 IEEE 13th International Conference on Indoor Positioning and Indoor Navigation (IPIN)},
pages = {1-6},
publisher = {IEEE},
abstract = {Over the last years, numerous studies have emerged using Wi-Fi channel state information, enabling device-free (passive) sensing for applications such as motion detection, indoor positioning or human activity recognition. More recently, the development framework for the low-cost ESP32 microcontrollers has added support for obtaining channel state information data. In this work, we collected channel state information data for human activity recognition, where activities are relatively localized with respect to the Wi-Fi infrastructure. The data was collected in several runs, duly spaced in time, and a convolutional neural network model was used for the classification of activities. Classification performance evaluation showed a clear degradation when a model evaluated with data collected 10 minutes after the data used for training showed a 52% relative loss in the accuracy of the classification.},
keywords = {activity recognition, CSI, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Over the last years, numerous studies have emerged using Wi-Fi channel state information, enabling device-free (passive) sensing for applications such as motion detection, indoor positioning or human activity recognition. More recently, the development framework for the low-cost ESP32 microcontrollers has added support for obtaining channel state information data. In this work, we collected channel state information data for human activity recognition, where activities are relatively localized with respect to the Wi-Fi infrastructure. The data was collected in several runs, duly spaced in time, and a convolutional neural network model was used for the classification of activities. Classification performance evaluation showed a clear degradation when a model evaluated with data collected 10 minutes after the data used for training showed a 52% relative loss in the accuracy of the classification.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
A set of deep learning algorithms for air quality prediction applications Journal Article
In: Software Impacts, vol. 17, pp. 100562, 2023, ISSN: 2665-9638.
Abstract | Links | BibTeX | Tags: geospatial analysis, machine learning, spatiotemporal prediction
@article{Iskandaryan2023d,
title = {A set of deep learning algorithms for air quality prediction applications},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.1016/j.simpa.2023.100562},
issn = {2665-9638},
year = {2023},
date = {2023-08-10},
journal = {Software Impacts},
volume = {17},
pages = {100562},
abstract = {This paper presents a set of machine learning algorithms, including grid-based (Bidirectional Convolutional Long Short-Term Memory) and graph-based (Attention Temporal Graph Convolutional Network) algorithms to predict air quality. The methods were implemented on a spatiotemporal combination of air quality, meteorological and traffic data of the city of Madrid. The two methods are exposed to be reused for prediction in other scenarios and different air quality phenomena.},
keywords = {geospatial analysis, machine learning, spatiotemporal prediction},
pubstate = {published},
tppubtype = {article}
}
This paper presents a set of machine learning algorithms, including grid-based (Bidirectional Convolutional Long Short-Term Memory) and graph-based (Attention Temporal Graph Convolutional Network) algorithms to predict air quality. The methods were implemented on a spatiotemporal combination of air quality, meteorological and traffic data of the city of Madrid. The two methods are exposed to be reused for prediction in other scenarios and different air quality phenomena.
Pascacio-de-los-Santos, Pavel
Collaborative Techniques for Indoor Positioning Systems PhD Thesis
Universitat Jaume I. INIT, 2023, ISBN: 978-952-03-2905-1.
Abstract | Links | BibTeX | Tags: A-wear, Bluetooth Low Energy, Indoor positioning, machine learning, Wi-Fi fingerprint
@phdthesis{Pascacio2023a,
title = {Collaborative Techniques for Indoor Positioning Systems},
author = {Pavel Pascacio-de-los-Santos},
url = {http://hdl.handle.net/10803/688489},
doi = {http://dx.doi.org/10.6035/14124.2023.821144},
isbn = {978-952-03-2905-1},
year = {2023},
date = {2023-06-09},
school = {Universitat Jaume I. INIT},
abstract = {This doctoral thesis focuses on developing and evaluating mobile device-based collaborative techniques to enhance the position accuracy of traditional indoor positioning systems based on RSSI (i.e., lateration and fingerprinting) in real-world conditions. During the research, first, a comprehensive systematic review of Collaborative Indoor Positioning Systems (CIPSs) was conducted to obtain a state-of-the-art; second, extensive experimental data collections considering mobile devices and collaborative scenarios were performed to create a mobile device-based BLE database and BLE and Wi-Fi radio maps for testing our collaborative and non-collaborative indoor positioning approaches; third, traditional methods to estimate distance and position were evaluated to present their limitations and challenges and two novel approaches to improve distance and positioning accuracy were proposed; finally, our proposed CIPSs using Multilayer Perceptron Artificial Neural Networks were developed to enhance the accuracy of BLE–RSSI lateration and fingerprinting-KNN methods and evaluated under real-world conditions to demonstrate its feasibility and benefits.},
keywords = {A-wear, Bluetooth Low Energy, Indoor positioning, machine learning, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {phdthesis}
}
This doctoral thesis focuses on developing and evaluating mobile device-based collaborative techniques to enhance the position accuracy of traditional indoor positioning systems based on RSSI (i.e., lateration and fingerprinting) in real-world conditions. During the research, first, a comprehensive systematic review of Collaborative Indoor Positioning Systems (CIPSs) was conducted to obtain a state-of-the-art; second, extensive experimental data collections considering mobile devices and collaborative scenarios were performed to create a mobile device-based BLE database and BLE and Wi-Fi radio maps for testing our collaborative and non-collaborative indoor positioning approaches; third, traditional methods to estimate distance and position were evaluated to present their limitations and challenges and two novel approaches to improve distance and positioning accuracy were proposed; finally, our proposed CIPSs using Multilayer Perceptron Artificial Neural Networks were developed to enhance the accuracy of BLE–RSSI lateration and fingerprinting-KNN methods and evaluated under real-world conditions to demonstrate its feasibility and benefits.
Chukhno, Nadezhda; Chukhno, Olga; Moltchanov, Dmitri; Molinaro, Antonella; Gaidamaka, Yuliya; Samouylov, Konstantin; Koucheryavy, Yevgeni; Araniti, Giuseppe
Optimal Multicasting in Millimeter Wave 5G NR With Multi-Beam Directional Antennas Journal Article
In: IEEE Transactions on Mobile Computing, vol. 22, no. 6, pp. 3572 - 3588, 2023, ISSN: 1558-0660.
Abstract | Links | BibTeX | Tags: A-wear, machine learning, wearables
@article{Chukhno2023a,
title = {Optimal Multicasting in Millimeter Wave 5G NR With Multi-Beam Directional Antennas},
author = {Nadezhda Chukhno and Olga Chukhno and Dmitri Moltchanov and Antonella Molinaro and Yuliya Gaidamaka and Konstantin Samouylov and Yevgeni Koucheryavy and Giuseppe Araniti},
doi = {10.1109/TMC.2021.3136298},
issn = {1558-0660},
year = {2023},
date = {2023-06-01},
journal = {IEEE Transactions on Mobile Computing},
volume = {22},
number = {6},
pages = {3572 - 3588},
abstract = {The support of multicast communications in the fifth-generation (5G) New Radio (NR) system poses unique challenges to system designers. Particularly, the highly directional antennas do not allow to serve all the user equipment devices (UEs) that belong to the same multicast session in a single transmission. The capability of modern antenna arrays to utilize multiple beams simultaneously, with potentially varying half-power beamwidth, adds a new degree of freedom to the UE scheduling. This work addresses the challenge of optimal multicasting in 5G millimeter wave (mmWave) systems by presenting a globally optimal solution for multi-beam antenna operation. The optimization problem is formulated as a special case of multi-period variable cost and size bin packing problem that allows to not impose any constraints on the number of the beams and their configurations. We also propose heuristic solutions having polynomial time complexity. Our results show that for small cell radii of up to 100 meters, a single beam is always utilized. For higher cell coverage and practical ranges of the number of users (5-50), the optimal number of beams is upper bounded by 3.},
keywords = {A-wear, machine learning, wearables},
pubstate = {published},
tppubtype = {article}
}
The support of multicast communications in the fifth-generation (5G) New Radio (NR) system poses unique challenges to system designers. Particularly, the highly directional antennas do not allow to serve all the user equipment devices (UEs) that belong to the same multicast session in a single transmission. The capability of modern antenna arrays to utilize multiple beams simultaneously, with potentially varying half-power beamwidth, adds a new degree of freedom to the UE scheduling. This work addresses the challenge of optimal multicasting in 5G millimeter wave (mmWave) systems by presenting a globally optimal solution for multi-beam antenna operation. The optimization problem is formulated as a special case of multi-period variable cost and size bin packing problem that allows to not impose any constraints on the number of the beams and their configurations. We also propose heuristic solutions having polynomial time complexity. Our results show that for small cell radii of up to 100 meters, a single beam is always utilized. For higher cell coverage and practical ranges of the number of users (5-50), the optimal number of beams is upper bounded by 3.
Klus, Lucie
From Compression of Wearable-based Data to Effortless Indoor Positioning PhD Thesis
Tampere University. Faculty of Information Technology and Communication Sciences, 2023, ISBN: 978-952-03-2832-0.
Abstract | Links | BibTeX | Tags: A-wear, Indoor positioning, machine learning, wearables
@phdthesis{Klus2023a,
title = {From Compression of Wearable-based Data to Effortless Indoor Positioning},
author = {Lucie Klus},
url = {http://hdl.handle.net/10803/688947},
doi = {http://dx.doi.org/10.6035/14124.2023.45900046},
isbn = {978-952-03-2832-0},
year = {2023},
date = {2023-04-27},
school = {Tampere University. Faculty of Information Technology and Communication Sciences},
abstract = {In recent years, wearable devices have become ever-present in modern society. They
are typically defined as small, battery-restricted devices, worn on, in, or in very close
proximity to a human body. Their performance is defined by their functionalities as
much as by their comfortability and convenience. As such, they need to be compact
yet powerful, thus making energy efficiency an extremely important and relevant
aspect of the system. The market of wearable devices is nowadays dominated by
smartwatches and fitness bands, which are capable of gathering numerous sensorbased
data such as temperature, pressure, heart rate, or blood oxygen level, which
have to be processed in real-time, stored, or wirelessly transferred while consuming
as little energy as possible to ensure long battery life. Implementing compression
schemes directly at the wearable device is one of the relevant methods to reduce the
volume of data and to minimize the number of required operations while processing
them, as raw measurements include plenty of redundancies that can be removed
without damaging the useful information itself.},
keywords = {A-wear, Indoor positioning, machine learning, wearables},
pubstate = {published},
tppubtype = {phdthesis}
}
In recent years, wearable devices have become ever-present in modern society. They
are typically defined as small, battery-restricted devices, worn on, in, or in very close
proximity to a human body. Their performance is defined by their functionalities as
much as by their comfortability and convenience. As such, they need to be compact
yet powerful, thus making energy efficiency an extremely important and relevant
aspect of the system. The market of wearable devices is nowadays dominated by
smartwatches and fitness bands, which are capable of gathering numerous sensorbased
data such as temperature, pressure, heart rate, or blood oxygen level, which
have to be processed in real-time, stored, or wirelessly transferred while consuming
as little energy as possible to ensure long battery life. Implementing compression
schemes directly at the wearable device is one of the relevant methods to reduce the
volume of data and to minimize the number of required operations while processing
them, as raw measurements include plenty of redundancies that can be removed
without damaging the useful information itself.
are typically defined as small, battery-restricted devices, worn on, in, or in very close
proximity to a human body. Their performance is defined by their functionalities as
much as by their comfortability and convenience. As such, they need to be compact
yet powerful, thus making energy efficiency an extremely important and relevant
aspect of the system. The market of wearable devices is nowadays dominated by
smartwatches and fitness bands, which are capable of gathering numerous sensorbased
data such as temperature, pressure, heart rate, or blood oxygen level, which
have to be processed in real-time, stored, or wirelessly transferred while consuming
as little energy as possible to ensure long battery life. Implementing compression
schemes directly at the wearable device is one of the relevant methods to reduce the
volume of data and to minimize the number of required operations while processing
them, as raw measurements include plenty of redundancies that can be removed
without damaging the useful information itself.
Quezada-Gaibor, Darwin
Cloud-based Indoor Positioning Platform for Context-adaptivity in GNSS-denied Scenarios PhD Thesis
Universitat Jaume I. INIT, 2023.
Abstract | Links | BibTeX | Tags: A-wear, Cloud computing, Indoor positioning, machine learning, Wi-Fi fingerprint
@phdthesis{Quezada2023a,
title = {Cloud-based Indoor Positioning Platform for Context-adaptivity in GNSS-denied Scenarios},
author = {Darwin Quezada-Gaibor},
doi = {http://dx.doi.org/10.6035/14124.2023.821275},
year = {2023},
date = {2023-03-31},
school = {Universitat Jaume I. INIT},
abstract = {The demand for positioning, localisation and navigation services is on the rise, largely owing to the fact that such services form an integral part of applications in areas such as agriculture, robotics, and eHealth. Depending on the field of application, these services must accomplish high levels of accuracy, flexibility, and integrability. This dissertation focuses on improving computing efficiency, data pre-processing, and software architecture for indoor positioning solutions without leaving aside position and location accuracy. The dissertation begins by presenting a systematic review of current cloud-based indoor positioning solutions. Secondly, we focus on the study of data optimisation techniques such as data cleansing and data augmentation. The third contribution suggests two algorithms to group similar fingerprints into clusters. The fourth contribution explores the use of Machine Learning (ML) models to enhance position estimation accuracy. Finally, this dissertation summarises the key findings in an open-source cloud platform for indoor positioning.},
keywords = {A-wear, Cloud computing, Indoor positioning, machine learning, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {phdthesis}
}
The demand for positioning, localisation and navigation services is on the rise, largely owing to the fact that such services form an integral part of applications in areas such as agriculture, robotics, and eHealth. Depending on the field of application, these services must accomplish high levels of accuracy, flexibility, and integrability. This dissertation focuses on improving computing efficiency, data pre-processing, and software architecture for indoor positioning solutions without leaving aside position and location accuracy. The dissertation begins by presenting a systematic review of current cloud-based indoor positioning solutions. Secondly, we focus on the study of data optimisation techniques such as data cleansing and data augmentation. The third contribution suggests two algorithms to group similar fingerprints into clusters. The fourth contribution explores the use of Machine Learning (ML) models to enhance position estimation accuracy. Finally, this dissertation summarises the key findings in an open-source cloud platform for indoor positioning.
Iskandaryan, Ditsuhi
Universitat Jaume I. INIT, 2023.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning, spatiotemporal prediction
@phdthesis{Iskandaryan2023c,
title = {Study and Prediction of Air Quality in Smart Cities through Machine Learning Techniques Considering Spatiotemporal Components},
author = {Ditsuhi Iskandaryan},
doi = {http://dx.doi.org/10.6035/14101.2023.726676},
year = {2023},
date = {2023-03-07},
school = {Universitat Jaume I. INIT},
abstract = {Air quality is considered one of the top concerns. Information and knowledge about air quality can assist in effectively monitoring and controlling concentrations, reducing or preventing its harmful impacts and consequences. The complexity of air quality dependence on various components in spatiotemporal dimensions creates additional challenges to acquire this information. The current dissertation proposes machine learning and deep learning technologies that are capable of capturing and processing multidimensional information and complex dependencies controlling air quality. The following components come together to formulate the novelty of the current work: spatiotemporal forecast of the defined prediction target (nitrogen dioxide); incorporation and integration of air quality, meteorological and traffic data with their features/variables in spatiotemporal dimensions within a certain spatial extent and temporal interval; the consideration of coronavirus disease 2019 as an external key factor impacting air quality level; and provision of the code and data implemented to incentivise and guarantee reproducibility.},
keywords = {air quality prediction, machine learning, spatiotemporal prediction},
pubstate = {published},
tppubtype = {phdthesis}
}
Air quality is considered one of the top concerns. Information and knowledge about air quality can assist in effectively monitoring and controlling concentrations, reducing or preventing its harmful impacts and consequences. The complexity of air quality dependence on various components in spatiotemporal dimensions creates additional challenges to acquire this information. The current dissertation proposes machine learning and deep learning technologies that are capable of capturing and processing multidimensional information and complex dependencies controlling air quality. The following components come together to formulate the novelty of the current work: spatiotemporal forecast of the defined prediction target (nitrogen dioxide); incorporation and integration of air quality, meteorological and traffic data with their features/variables in spatiotemporal dimensions within a certain spatial extent and temporal interval; the consideration of coronavirus disease 2019 as an external key factor impacting air quality level; and provision of the code and data implemented to incentivise and guarantee reproducibility.
Chukhno, Nadezhda; Chukhno, Olga; Moltchanov, Dmitri; Gaydamaka, Anna; Samuylov, Andrey; Molinaro, Antonella; Koucheryavy, Yevgeni; Iera, Antonio
The Use of Machine Learning Techniques for Optimal Multicasting in 5G NR Systems Journal Article
In: IEEE Transactions on Broadcasting, vol. 69, no. 1, pp. 201-214, 2023, ISSN: 1557-9611.
Abstract | Links | BibTeX | Tags: A-wear, machine learning, wearables
@article{Chukhno2023b,
title = {The Use of Machine Learning Techniques for Optimal Multicasting in 5G NR Systems},
author = {Nadezhda Chukhno and Olga Chukhno and Dmitri Moltchanov and Anna Gaydamaka and Andrey Samuylov and Antonella Molinaro and Yevgeni Koucheryavy and Antonio Iera},
doi = {10.1109/TBC.2022.3206595},
issn = {1557-9611},
year = {2023},
date = {2023-03-01},
journal = {IEEE Transactions on Broadcasting},
volume = {69},
number = {1},
pages = {201-214},
abstract = {Multicasting is a key feature of cellular systems, which provides an efficient way to simultaneously disseminate a large amount of traffic to multiple subscribers. However, the efficient use of multicast services in fifth-generation (5G) New Radio (NR) is complicated by several factors, including inherent base station (BS) antenna directivity as well as the exploitation of antenna arrays capable of creating multiple beams concurrently. In this work, we first demonstrate that the problem of efficient multicasting in 5G NR systems can be formalized as a special case of multi-period variable cost and size bin packing problem (BPP). However, the problem is known to be NP-hard, and the solution time is practically unacceptable for large multicast group sizes. To this aim, we further develop and test several machine learning alternatives to address this issue. The numerical analysis shows that there is a trade-off between accuracy and computational complexity for multicast grouping when using decision tree-based algorithms. A higher number of splits offers better performance at the cost of an increased computational time. We also show that the nature of the cell coverage brings three possible solutions to the multicast grouping problem: (i) small-range radii are characterized by a single multicast subgroup with wide beamwidth, (ii) middle-range deployments have to be solved by employing the proposed algorithms, and (iii) BS at long-range radii sweeps narrow unicast beams to serve multicast users.},
keywords = {A-wear, machine learning, wearables},
pubstate = {published},
tppubtype = {article}
}
Multicasting is a key feature of cellular systems, which provides an efficient way to simultaneously disseminate a large amount of traffic to multiple subscribers. However, the efficient use of multicast services in fifth-generation (5G) New Radio (NR) is complicated by several factors, including inherent base station (BS) antenna directivity as well as the exploitation of antenna arrays capable of creating multiple beams concurrently. In this work, we first demonstrate that the problem of efficient multicasting in 5G NR systems can be formalized as a special case of multi-period variable cost and size bin packing problem (BPP). However, the problem is known to be NP-hard, and the solution time is practically unacceptable for large multicast group sizes. To this aim, we further develop and test several machine learning alternatives to address this issue. The numerical analysis shows that there is a trade-off between accuracy and computational complexity for multicast grouping when using decision tree-based algorithms. A higher number of splits offers better performance at the cost of an increased computational time. We also show that the nature of the cell coverage brings three possible solutions to the multicast grouping problem: (i) small-range radii are characterized by a single multicast subgroup with wide beamwidth, (ii) middle-range deployments have to be solved by employing the proposed algorithms, and (iii) BS at long-range radii sweeps narrow unicast beams to serve multicast users.
Torres-Sospedra, Joaquín; Quezada-Gaibor, Darwin; Nurmi, Jari; Koucheryavy, Yevgeni; Lohan, Elena Simona; Huerta-Guijarro, Joaquín
Scalable and Efficient Clustering for Fingerprint-Based Positioning Journal Article
In: IEEE Internet of Things Journal, vol. 10, no. 4, pp. 3484 - 3499, 2023, ISSN: 2327-4662.
Abstract | Links | BibTeX | Tags: Bluetooth Low Energy, Indoor localization, machine learning, Wi-Fi fingerprint
@article{Torres-Sospedra2023a,
title = {Scalable and Efficient Clustering for Fingerprint-Based Positioning},
author = {Joaquín Torres-Sospedra and Darwin Quezada-Gaibor and Jari Nurmi and Yevgeni Koucheryavy and Elena Simona Lohan and Joaquín Huerta-Guijarro},
doi = {10.1109/JIOT.2022.3230913},
issn = {2327-4662},
year = {2023},
date = {2023-02-13},
journal = {IEEE Internet of Things Journal},
volume = {10},
number = {4},
pages = {3484 - 3499},
abstract = {Indoor positioning based on IEEE 802.11 wireless LAN (Wi-Fi) fingerprinting needs a reference data set, also known as a radio map, in order to match the incoming fingerprint in the operational phase with the most similar fingerprint in the data set and then estimate the device position indoors. Scalability problems may arise when the radio map is large, e.g., providing positioning in large geographical areas or involving crowdsourced data collection. Some researchers divide the radio map into smaller independent clusters, such that the search area is reduced to less dense groups than the initial database with similar features. Thus, the computational load in the operational stage is reduced both at the user devices and on servers. Nevertheless, the clustering models are machine-learning algorithms without specific domain knowledge on indoor positioning or signal propagation. This work proposes several clustering variants to optimize the coarse and fine-grained search and evaluates them over different clustering models and data sets. Moreover, we provide guidelines to obtain efficient and accurate positioning depending on the data set features. Finally, we show that the proposed new clustering variants reduce the execution time by half and the positioning error by ≈7 % with respect to fingerprinting with the traditional clustering models.},
keywords = {Bluetooth Low Energy, Indoor localization, machine learning, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {article}
}
Indoor positioning based on IEEE 802.11 wireless LAN (Wi-Fi) fingerprinting needs a reference data set, also known as a radio map, in order to match the incoming fingerprint in the operational phase with the most similar fingerprint in the data set and then estimate the device position indoors. Scalability problems may arise when the radio map is large, e.g., providing positioning in large geographical areas or involving crowdsourced data collection. Some researchers divide the radio map into smaller independent clusters, such that the search area is reduced to less dense groups than the initial database with similar features. Thus, the computational load in the operational stage is reduced both at the user devices and on servers. Nevertheless, the clustering models are machine-learning algorithms without specific domain knowledge on indoor positioning or signal propagation. This work proposes several clustering variants to optimize the coarse and fine-grained search and evaluates them over different clustering models and data sets. Moreover, we provide guidelines to obtain efficient and accurate positioning depending on the data set features. Finally, we show that the proposed new clustering variants reduce the execution time by half and the positioning error by ≈7 % with respect to fingerprinting with the traditional clustering models.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Graph Neural Network for Air Quality Prediction: A Case Study in Madrid Journal Article
In: IEEE Access, vol. 11, pp. 2729-2742, 2023, ISSN: 2169-3536.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning, spatiotemporal prediction
@article{Iskandaryan2023a,
title = {Graph Neural Network for Air Quality Prediction: A Case Study in Madrid},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {10.1109/ACCESS.2023.3234214},
issn = {2169-3536},
year = {2023},
date = {2023-01-04},
journal = {IEEE Access},
volume = {11},
pages = {2729-2742},
abstract = {Air quality monitoring, modelling and forecasting are considered pressing and challenging topics for citizens and decision-makers, including the government. The tools used to achieve the above goals vary depending on the opportunities provided by technological development. Much attention is currently being paid to machine learning and deep learning methods, which, compared to domain knowledge methods, often perform better in terms of capturing, computing and processing multidimensional information and complex dependencies. The technique introduced in this work is an Attention Temporal Graph Convolutional Network based on a combination of Attention, a Gated Recurrent Unit and a Graph Convolutional Network. In the framework of the current study, it is initially suggested to use the presented approach in the domain of air quality prediction. The proposed method was tested using air quality, meteorological and traffic data obtained from the city of Madrid for the periods January-June 2019 and January-June 2022. The evaluation metrics, including Root Mean Square Error, Mean Absolute Error and Pearson Correlation Coefficient, confirmed the proposed model’s advantages compared with the reference models (Temporal Graph Convolutional Network, Long Short-Term Memory and Gated Recurrent Unit).},
keywords = {air quality prediction, machine learning, spatiotemporal prediction},
pubstate = {published},
tppubtype = {article}
}
Air quality monitoring, modelling and forecasting are considered pressing and challenging topics for citizens and decision-makers, including the government. The tools used to achieve the above goals vary depending on the opportunities provided by technological development. Much attention is currently being paid to machine learning and deep learning methods, which, compared to domain knowledge methods, often perform better in terms of capturing, computing and processing multidimensional information and complex dependencies. The technique introduced in this work is an Attention Temporal Graph Convolutional Network based on a combination of Attention, a Gated Recurrent Unit and a Graph Convolutional Network. In the framework of the current study, it is initially suggested to use the presented approach in the domain of air quality prediction. The proposed method was tested using air quality, meteorological and traffic data obtained from the city of Madrid for the periods January-June 2019 and January-June 2022. The evaluation metrics, including Root Mean Square Error, Mean Absolute Error and Pearson Correlation Coefficient, confirmed the proposed model’s advantages compared with the reference models (Temporal Graph Convolutional Network, Long Short-Term Memory and Gated Recurrent Unit).
2022
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Spatiotemporal Prediction of Nitrogen Dioxide Based on Graph Neural Networks Proceedings Article
In: Advances and New Trends in Environmental Informatics. ENVIROINFO 2022. , pp. 111–128, Springer, Cham, 2022, ISBN: 978-3-031-18311-9.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning
@inproceedings{Iskandaryan2022e,
title = {Spatiotemporal Prediction of Nitrogen Dioxide Based on Graph Neural Networks},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.1007/978-3-031-18311-9_7},
isbn = {978-3-031-18311-9},
year = {2022},
date = {2022-11-10},
booktitle = {Advances and New Trends in Environmental Informatics. ENVIROINFO 2022. },
pages = {111–128},
publisher = {Springer, Cham},
series = {Progress in IS},
abstract = {Air quality prediction, especially spatiotemporal prediction, is still a challenging issue. Considering the impact of numerous factors on air quality causes difficulties in integrating these factors in a spatiotemporal dimension and developing a model to make efficient predictions. At the same time, machine learning and deep learning development bring advanced approaches to addressing these challenges and propose novel solutions. The current work introduces one of the most advanced methods, an attention temporal graph convolutional network, which was implemented on datasets constructed by combining air quality, meteorological and traffic data on a spatiotemporal axis. The datasets were obtained from the city of Madrid for the periods January-June 2019 and January–June 2020. The evaluation metrics, the Root Mean Square Error and the Mean Absolute Error confirmed the proposed model’s advantages compared with long short-term memory (reference model). Particularly, it outperformed the latter method by 14.18% and 3.78%, respectively.},
keywords = {air quality prediction, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Air quality prediction, especially spatiotemporal prediction, is still a challenging issue. Considering the impact of numerous factors on air quality causes difficulties in integrating these factors in a spatiotemporal dimension and developing a model to make efficient predictions. At the same time, machine learning and deep learning development bring advanced approaches to addressing these challenges and propose novel solutions. The current work introduces one of the most advanced methods, an attention temporal graph convolutional network, which was implemented on datasets constructed by combining air quality, meteorological and traffic data on a spatiotemporal axis. The datasets were obtained from the city of Madrid for the periods January-June 2019 and January–June 2020. The evaluation metrics, the Root Mean Square Error and the Mean Absolute Error confirmed the proposed model’s advantages compared with long short-term memory (reference model). Particularly, it outperformed the latter method by 14.18% and 3.78%, respectively.
Pascacio-de-los-Santos, Pavel; Torres-Sospedra, Joaquín; Casteleyn, Sven; Lohan, Elena Simona
A Collaborative Approach Using Neural Networks for BLE-RSS Lateration-Based Indoor Positioning Proceedings Article
In: 2022 International Joint Conference on Neural Networks (IJCNN), pp. 1-9, IEEE, 2022, ISBN: 978-1-7281-8671-9.
Abstract | Links | BibTeX | Tags: Bluetooth Low Energy, Indoor positioning, machine learning
@inproceedings{Pascacio2022b,
title = {A Collaborative Approach Using Neural Networks for BLE-RSS Lateration-Based Indoor Positioning},
author = {Pavel Pascacio-de-los-Santos and Joaquín Torres-Sospedra and Sven Casteleyn and Elena Simona Lohan},
doi = {https://doi.org/10.1109/IJCNN55064.2022.9892484},
isbn = {978-1-7281-8671-9},
year = {2022},
date = {2022-09-30},
booktitle = {2022 International Joint Conference on Neural Networks (IJCNN)},
pages = {1-9},
publisher = {IEEE},
abstract = {In daily life, mobile and wearable devices with high computing power, together with anchors deployed in indoor en-vironments, form a common solution for the increasing demands for indoor location-based services. Within the technologies and methods currently in use for indoor localization, the approaches that rely on Bluetooth Low Energy (BLE) anchors, Received Signal Strength (RSS), and lateration are among the most popular, mainly because of their cheap and easy deployment and accessible infrastructure by a variety of devices. Never-theless, such BLE- and RSS-based indoor positioning systems are prone to inaccuracies, mostly due to signal fluctuations, poor quantity of anchors deployed in the environment, and/or inappropriate anchor distributions, as well as mobile device hardware variability. In this paper, we address these issues by using a collaborative indoor positioning approach, which exploits neighboring devices as additional anchors in an extended positioning network. The collaborating devices' information (i.e., estimated positions and BLE- RSS) is processed using a multilayer perceptron (MLP) neural network by taking into account the device specificity in order to estimate the relative distances. After this, the lateration is applied to collaboratively estimate the device position. Finally, the stand-alone and collaborative position estimates are combined, providing the final position estimate for each device. The experimental results demonstrate that the proposed collaborative approach outperforms the stand-alone lateration method in terms of positioning accuracy.},
keywords = {Bluetooth Low Energy, Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
In daily life, mobile and wearable devices with high computing power, together with anchors deployed in indoor en-vironments, form a common solution for the increasing demands for indoor location-based services. Within the technologies and methods currently in use for indoor localization, the approaches that rely on Bluetooth Low Energy (BLE) anchors, Received Signal Strength (RSS), and lateration are among the most popular, mainly because of their cheap and easy deployment and accessible infrastructure by a variety of devices. Never-theless, such BLE- and RSS-based indoor positioning systems are prone to inaccuracies, mostly due to signal fluctuations, poor quantity of anchors deployed in the environment, and/or inappropriate anchor distributions, as well as mobile device hardware variability. In this paper, we address these issues by using a collaborative indoor positioning approach, which exploits neighboring devices as additional anchors in an extended positioning network. The collaborating devices' information (i.e., estimated positions and BLE- RSS) is processed using a multilayer perceptron (MLP) neural network by taking into account the device specificity in order to estimate the relative distances. After this, the lateration is applied to collaboratively estimate the device position. Finally, the stand-alone and collaborative position estimates are combined, providing the final position estimate for each device. The experimental results demonstrate that the proposed collaborative approach outperforms the stand-alone lateration method in terms of positioning accuracy.
Quezada-Gaibor, Darwin; Torres-Sospedra, Joaquín; Nurmi, Jari; Koucheryavy, Yevgeni; Huerta-Guijarro, Joaquín
SURIMI: Supervised Radio Map Augmentation with Deep Learning and a Generative Adversarial Network for Fingerprint-based Indoor Positioning Proceedings Article
In: 2022 IEEE 12th International Conference on Indoor Positioning and Indoor Navigation (IPIN), IEEE, 2022, ISBN: 978-1-7281-6218-8.
Abstract | Links | BibTeX | Tags: deep learning, Indoor positioning, machine learning
@inproceedings{Quezada2022d,
title = {SURIMI: Supervised Radio Map Augmentation with Deep Learning and a Generative Adversarial Network for Fingerprint-based Indoor Positioning},
author = {Darwin Quezada-Gaibor and Joaquín Torres-Sospedra and Jari Nurmi and Yevgeni Koucheryavy and Joaquín Huerta-Guijarro},
doi = {10.1109/IPIN54987.2022.9918146},
isbn = {978-1-7281-6218-8},
year = {2022},
date = {2022-09-06},
booktitle = {2022 IEEE 12th International Conference on Indoor Positioning and Indoor Navigation (IPIN)},
number = {1-8},
publisher = {IEEE},
abstract = {Indoor Positioning based on Machine Learning has drawn increasing attention both in the academy and the industry as meaningful information from the reference data can be extracted. Many researchers are using supervised, semi-supervised, and unsupervised Machine Learning models to reduce the positioning error and offer reliable solutions to the end-users. In this article, we propose a new architecture by combining Convolutional Neural Network (CNN), Long short-term memory (LSTM) and Generative Adversarial Network (GAN) in order to increase the training data and thus improve the position accuracy. The proposed combination of supervised and unsupervised models was tested in 17 public datasets, providing an extensive analysis of its performance. As a result, the positioning error has been reduced in more than 70% of them.},
keywords = {deep learning, Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Indoor Positioning based on Machine Learning has drawn increasing attention both in the academy and the industry as meaningful information from the reference data can be extracted. Many researchers are using supervised, semi-supervised, and unsupervised Machine Learning models to reduce the positioning error and offer reliable solutions to the end-users. In this article, we propose a new architecture by combining Convolutional Neural Network (CNN), Long short-term memory (LSTM) and Generative Adversarial Network (GAN) in order to increase the training data and thus improve the position accuracy. The proposed combination of supervised and unsupervised models was tested in 17 public datasets, providing an extensive analysis of its performance. As a result, the positioning error has been reduced in more than 70% of them.
Matey-Sanz, Miguel; González-Pérez, Alberto; Casteleyn, Sven; Granell-Canut, Carlos
Instrumented Timed Up and Go Test Using Inertial Sensors from Consumer Wearable Devices Proceedings Article
In: Artificial Intelligence in Medicine. AIME 2022, pp. 144-154, Springer, Cham, 2022, ISBN: 978-3031093418.
Abstract | Links | BibTeX | Tags: machine learning, Mobile apps, mobile computing, symptoms, wearables
@inproceedings{Matey2022a,
title = {Instrumented Timed Up and Go Test Using Inertial Sensors from Consumer Wearable Devices},
author = {Miguel Matey-Sanz and Alberto González-Pérez and Sven Casteleyn and Carlos Granell-Canut},
doi = {https://doi.org/10.1007/978-3-031-09342-5_14},
isbn = {978-3031093418},
year = {2022},
date = {2022-07-09},
booktitle = {Artificial Intelligence in Medicine. AIME 2022},
volume = {13263},
pages = {144-154},
publisher = {Springer, Cham},
series = {Lectures Notes in Artificial Intelligence},
abstract = {Precision medicine pursues the ambitious goal of providing personalized interventions targeted at individual patients. Within this vision, digital health and mental health, where fine-grained monitoring of patients form the basis for so-called ecological momentary assessments and interventions, play a central role as complementary technology-based and data-driven instruments to traditional psychological treatments. Mobile devices are hereby key enablers: consumer smartphones and wearables are ubiquitously present and used in daily life, while they come with the necessary embedded physiological, inertial and movement sensors to potentially recognise user’s activities and behaviors. In this article, we explore whether real-time detection of fine-grained activities - relevant in the context of wellbeing - is feasible, applying machine learning techniques and based on sensor data collected from a consumer smartwatch device. We present the system architecture, whereby data collection is performed in the wearable device, real-time data processing and inference is delegated to the paired smartphone, and model training is performed offline. Finally, we demonstrate its use by instrumenting the well-known Timed Up and Go (TUG) test, typically used to assess the risk of fall in elderly people. Experiments show that consumer smartwatches can be used to automate the assessment of TUG tests and obtain satisfactory results, comparable with the classical manually performed version of the test.},
keywords = {machine learning, Mobile apps, mobile computing, symptoms, wearables},
pubstate = {published},
tppubtype = {inproceedings}
}
Precision medicine pursues the ambitious goal of providing personalized interventions targeted at individual patients. Within this vision, digital health and mental health, where fine-grained monitoring of patients form the basis for so-called ecological momentary assessments and interventions, play a central role as complementary technology-based and data-driven instruments to traditional psychological treatments. Mobile devices are hereby key enablers: consumer smartphones and wearables are ubiquitously present and used in daily life, while they come with the necessary embedded physiological, inertial and movement sensors to potentially recognise user’s activities and behaviors. In this article, we explore whether real-time detection of fine-grained activities - relevant in the context of wellbeing - is feasible, applying machine learning techniques and based on sensor data collected from a consumer smartwatch device. We present the system architecture, whereby data collection is performed in the wearable device, real-time data processing and inference is delegated to the paired smartphone, and model training is performed offline. Finally, we demonstrate its use by instrumenting the well-known Timed Up and Go (TUG) test, typically used to assess the risk of fall in elderly people. Experiments show that consumer smartwatches can be used to automate the assessment of TUG tests and obtain satisfactory results, comparable with the classical manually performed version of the test.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Comparison of Nitrogen Dioxide Predictions During a Pandemic and Non-pandemic Scenario in the City of Madrid using a Convolutional LSTM Network Journal Article
In: International Journal of Computational Intelligence and Applications, vol. 21, no. 2, pp. 2250014, 2022, ISSN: 1757-5885.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning
@article{Iskandaryan2022d,
title = {Comparison of Nitrogen Dioxide Predictions During a Pandemic and Non-pandemic Scenario in the City of Madrid using a Convolutional LSTM Network},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.1142/S1469026822500146},
issn = {1757-5885},
year = {2022},
date = {2022-06-21},
journal = {International Journal of Computational Intelligence and Applications},
volume = {21},
number = {2},
pages = {2250014},
abstract = {Traditionally, machine learning technologies with the methods and capabilities available, combined with a geospatial dimension, can perform predictive analyzes of air quality with greater accuracy. However, air pollution is influenced by many external factors, one of which has recently been caused by the restrictions applied to curb the relentless advance of COVID-19. These sudden changes in air quality levels can negatively influence current forecasting models. This work compares air pollution forecasts during a pandemic and non-pandemic period under the same conditions. The ConvLSTM algorithm was applied to predict the concentration of nitrogen dioxide using data from the air quality and meteorological stations in Madrid. The proposed model was applied for two scenarios: pandemic (January–June 2020) and non-pandemic (January–June 2019), each with sub-scenarios based on time granularity (1-h, 12-h, 24-h and 48-h) and combination of features. The Root Mean Square Error was taken as the estimation metric, and the results showed that the proposed method outperformed a reference model, and the feature selection technique significantly improved the overall accuracy.},
keywords = {air quality prediction, machine learning},
pubstate = {published},
tppubtype = {article}
}
Traditionally, machine learning technologies with the methods and capabilities available, combined with a geospatial dimension, can perform predictive analyzes of air quality with greater accuracy. However, air pollution is influenced by many external factors, one of which has recently been caused by the restrictions applied to curb the relentless advance of COVID-19. These sudden changes in air quality levels can negatively influence current forecasting models. This work compares air pollution forecasts during a pandemic and non-pandemic period under the same conditions. The ConvLSTM algorithm was applied to predict the concentration of nitrogen dioxide using data from the air quality and meteorological stations in Madrid. The proposed model was applied for two scenarios: pandemic (January–June 2020) and non-pandemic (January–June 2019), each with sub-scenarios based on time granularity (1-h, 12-h, 24-h and 48-h) and combination of features. The Root Mean Square Error was taken as the estimation metric, and the results showed that the proposed method outperformed a reference model, and the feature selection technique significantly improved the overall accuracy.
Quezada-Gaibor, Darwin; Torres-Sospedra, Joaquín; Nurmi, Jari; Koucheryavy, Yevgeni; Huerta-Guijarro, Joaquín
Lightweight Hybrid CNN-ELM Model for Multi-building and Multi-floor Classification Proceedings Article
In: 2022 International Conference on Localization and GNSS (ICL-GNSS), pp. 1-6, IEEE, 2022.
Abstract | Links | BibTeX | Tags: Indoor positioning, machine learning
@inproceedings{Quezada2022b,
title = {Lightweight Hybrid CNN-ELM Model for Multi-building and Multi-floor Classification},
author = {Darwin Quezada-Gaibor and Joaquín Torres-Sospedra and Jari Nurmi and Yevgeni Koucheryavy and Joaquín Huerta-Guijarro},
doi = {https://doi.org/10.1109/ICL-GNSS54081.2022.9797021},
year = {2022},
date = {2022-06-19},
booktitle = {2022 International Conference on Localization and GNSS (ICL-GNSS)},
pages = {1-6},
publisher = {IEEE},
abstract = {Machine learning models have become an essential tool in current indoor positioning solutions, given their high capa-bilities to extract meaningful information from the environment. Convolutional neural networks (CNNs) are one of the most used neural networks (NNs) due to that they are capable of learning complex patterns from the input data. Another model used in indoor positioning solutions is the Extreme Learning Machine (ELM), which provides an acceptable generalization performance as well as a fast speed of learning. In this paper, we offer a lightweight combination of CNN and ELM, which provides a quick and accurate classification of building and floor, suitable for power and resource-constrained devices. As a result, the proposed model is 58% faster than the benchmark, with a slight improvement in the classification accuracy (by less than 1 %).},
keywords = {Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Machine learning models have become an essential tool in current indoor positioning solutions, given their high capa-bilities to extract meaningful information from the environment. Convolutional neural networks (CNNs) are one of the most used neural networks (NNs) due to that they are capable of learning complex patterns from the input data. Another model used in indoor positioning solutions is the Extreme Learning Machine (ELM), which provides an acceptable generalization performance as well as a fast speed of learning. In this paper, we offer a lightweight combination of CNN and ELM, which provides a quick and accurate classification of building and floor, suitable for power and resource-constrained devices. As a result, the proposed model is 58% faster than the benchmark, with a slight improvement in the classification accuracy (by less than 1 %).
Klus, Lucie; Quezada-Gaibor, Darwin; Torres-Sospedra, Joaquín; Lohan, Elena Simona; Nurmi, Jari; Granell-Canut, Carlos; Huerta-Guijarro, Joaquín
Towards Accelerated Localization Performance Across Indoor Positioning Datasets Proceedings Article
In: 2022 International Conference on Localization and GNSS (ICL-GNSS), pp. 1-7, IEEE, 2022.
Abstract | Links | BibTeX | Tags: Indoor localization, machine learning
@inproceedings{Klus2022a,
title = {Towards Accelerated Localization Performance Across Indoor Positioning Datasets},
author = {Lucie Klus and Darwin Quezada-Gaibor and Joaquín Torres-Sospedra and Elena Simona Lohan and Jari Nurmi and Carlos Granell-Canut and Joaquín Huerta-Guijarro},
doi = {https://doi.org/10.1109/ICL-GNSS54081.2022.9797035},
year = {2022},
date = {2022-06-19},
booktitle = {2022 International Conference on Localization and GNSS (ICL-GNSS)},
pages = {1-7},
publisher = {IEEE},
abstract = {he localization speed and accuracy in the indoor scenario can greatly impact the Quality of Experience of the user. While many individual machine learning models can achieve comparable positioning performance, their prediction mechanisms offer different complexity to the system. In this work, we propose a fingerprinting positioning method for multi-building and multi-floor deployments, composed of a cascade of three models for building classification, floor classification, and 2D localization regression. We conduct an exhaustive search for the optimally performing one in each step of the cascade while validating on 14 different openly available datasets. As a result, we bring forward the best-performing combination of models in terms of overall positioning accuracy and processing speed and evaluate on independent sets of samples. We reduce the mean prediction time by 71% while achieving comparable positioning performance across all considered datasets. Moreover, in case of voluminous training dataset, the prediction time is reduced down to 1% of the benchmark's},
keywords = {Indoor localization, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
he localization speed and accuracy in the indoor scenario can greatly impact the Quality of Experience of the user. While many individual machine learning models can achieve comparable positioning performance, their prediction mechanisms offer different complexity to the system. In this work, we propose a fingerprinting positioning method for multi-building and multi-floor deployments, composed of a cascade of three models for building classification, floor classification, and 2D localization regression. We conduct an exhaustive search for the optimally performing one in each step of the cascade while validating on 14 different openly available datasets. As a result, we bring forward the best-performing combination of models in terms of overall positioning accuracy and processing speed and evaluate on independent sets of samples. We reduce the mean prediction time by 71% while achieving comparable positioning performance across all considered datasets. Moreover, in case of voluminous training dataset, the prediction time is reduced down to 1% of the benchmark's
Iskandaryan, Ditsuhi; Sabatino, Silvana Di; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Exploratory Analysis and Feature Selection for the Prediction of Nitrogen Dioxide Proceedings Article
In: AGILE GIScience Series (Proceedings of the 25th AGILE Conference on Geographic Information Science), Copernicus Publications, 2022.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning
@inproceedings{Iskandaryan2022c,
title = {Exploratory Analysis and Feature Selection for the Prediction of Nitrogen Dioxide},
author = {Ditsuhi Iskandaryan and Silvana Di Sabatino and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = { https://doi.org/10.5194/agile-giss-3-6-2022},
year = {2022},
date = {2022-06-15},
booktitle = {AGILE GIScience Series (Proceedings of the 25th AGILE Conference on Geographic Information Science)},
volume = {3},
number = {6},
publisher = {Copernicus Publications},
abstract = {Nitrogen dioxide is one of the most hazardous pollutants identified by the World Health Organisation. Predicting and reducing pollutants is becoming a very urgent task and many methods have been used to predict their concentration, such as physical or machine learning models. In addition to choosing the right model, it is also critical to choose the appropriate features. This work focuses on the spatiotemporal prediction of nitrogen dioxide concentration using Bidirectional Convolutional LSTM integrated with the exploration of nitrogen dioxide and associated features, as well as the implementation of feature selection methods. The Root Mean Square Error and the Mean Absolute Error were used to evaluate the proposed approach.},
keywords = {air quality prediction, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Nitrogen dioxide is one of the most hazardous pollutants identified by the World Health Organisation. Predicting and reducing pollutants is becoming a very urgent task and many methods have been used to predict their concentration, such as physical or machine learning models. In addition to choosing the right model, it is also critical to choose the appropriate features. This work focuses on the spatiotemporal prediction of nitrogen dioxide concentration using Bidirectional Convolutional LSTM integrated with the exploration of nitrogen dioxide and associated features, as well as the implementation of feature selection methods. The Root Mean Square Error and the Mean Absolute Error were used to evaluate the proposed approach.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Bidirectional convolutional LSTM for the prediction of nitrogen dioxide in the city of Madrid Journal Article
In: PLOS ONE, vol. 17, no. 6, pp. e0269295, 2022, ISSN: 932-6203.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning
@article{Iskandaryan2022b,
title = {Bidirectional convolutional LSTM for the prediction of nitrogen dioxide in the city of Madrid},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.1371/journal.pone.0269295},
issn = {932-6203},
year = {2022},
date = {2022-05-01},
journal = {PLOS ONE},
volume = {17},
number = {6},
pages = {e0269295},
abstract = {Nitrogen dioxide is one of the pollutants with the most significant health effects. Advanced information on its concentration in the air can help to monitor and control further consequences more effectively, while also making it easier to apply preventive and mitigating measures. Machine learning technologies with available methods and capabilities, combined with the geospatial dimension, can perform predictive analyses with higher accuracy and, as a result, can serve as a supportive tool for productive management. One of the most advanced machine learning algorithms, Bidirectional convolutional LSTM, is being used in ongoing work to predict the concentration of nitrogen dioxide. The model has been validated to perform more accurate spatiotemporal analysis based on the integration of temporal and geospatial factors. The analysis was carried out according to two scenarios developed on the basis of selected features using data from the city of Madrid for the periods January-June 2019 and January-June 2020. Evaluation of the model’s performance was conducted using the Root Mean Square Error and the Mean Absolute Error which emphasises the superiority of the proposed model over the reference models. In addition, the significance of a feature selection technique providing improved accuracy was underlined. In terms of execution time, due to the complexity of the Bidirectional convolutional LSTM architecture, convergence and generalisation of the data took longer, resulting in the superiority of the reference models.},
keywords = {air quality prediction, machine learning},
pubstate = {published},
tppubtype = {article}
}
Nitrogen dioxide is one of the pollutants with the most significant health effects. Advanced information on its concentration in the air can help to monitor and control further consequences more effectively, while also making it easier to apply preventive and mitigating measures. Machine learning technologies with available methods and capabilities, combined with the geospatial dimension, can perform predictive analyses with higher accuracy and, as a result, can serve as a supportive tool for productive management. One of the most advanced machine learning algorithms, Bidirectional convolutional LSTM, is being used in ongoing work to predict the concentration of nitrogen dioxide. The model has been validated to perform more accurate spatiotemporal analysis based on the integration of temporal and geospatial factors. The analysis was carried out according to two scenarios developed on the basis of selected features using data from the city of Madrid for the periods January-June 2019 and January-June 2020. Evaluation of the model’s performance was conducted using the Root Mean Square Error and the Mean Absolute Error which emphasises the superiority of the proposed model over the reference models. In addition, the significance of a feature selection technique providing improved accuracy was underlined. In terms of execution time, due to the complexity of the Bidirectional convolutional LSTM architecture, convergence and generalisation of the data took longer, resulting in the superiority of the reference models.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Application of deep learning and machine learning in air quality modeling Book Chapter
In: Marques, Gonçalo; Ighalo, Joshua (Ed.): pp. 11-23, Elsevier, 2022, ISBN: 9780323855976.
Links | BibTeX | Tags: air quality prediction, deep learning, machine learning
@inbook{Iskandaryan2022a,
title = {Application of deep learning and machine learning in air quality modeling},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
editor = {Gonçalo Marques and Joshua Ighalo },
doi = {https://doi.org/10.1016/B978-0-323-85597-6.00018-5},
isbn = {9780323855976},
year = {2022},
date = {2022-03-30},
pages = {11-23},
publisher = {Elsevier},
keywords = {air quality prediction, deep learning, machine learning},
pubstate = {published},
tppubtype = {inbook}
}
2021
Bellavista-Parent, Vladimir; Torres-Sospedra, Joaquín; Perez-Navarro, Antoni
New trends in indoor positioning based on WiFi and machine learning: A systematic review Proceedings Article
In: Proceedings of the Eleventh International Conference on Indoor Positioning and Indoor Navigation, IEEE, 2021.
Abstract | Links | BibTeX | Tags: Indoor positioning, machine learning
@inproceedings{Bellavista2021a,
title = {New trends in indoor positioning based on WiFi and machine learning: A systematic review},
author = {Vladimir Bellavista-Parent and Joaquín Torres-Sospedra and Antoni Perez-Navarro},
doi = {https://doi.org/10.1109/IPIN51156.2021.9662521},
year = {2021},
date = {2021-12-15},
booktitle = {Proceedings of the Eleventh International Conference on Indoor Positioning and Indoor Navigation},
publisher = {IEEE},
abstract = {Currently there is no standard indoor positioning system, similar to outdoor GPS. However, WiFi signals have been used in a large number of proposals to achieve the above positioning, many of which use machine learning to do so. But what are the most commonly used techniques in machine learning? What accuracy do they achieve? Where have they been tested? This article presents a systematic review of works between 2019 and 2021 that use WiFi as the signal for positioning and machine learning models to estimate indoor position. 64 papers have been identified as relevant, which have been systematically analyzed for a better understanding of the current situation in different aspects. The results show that indoor positioning based on WiFi trends use neural network-based models, evaluated in empirical experiments. Despite this, many works still conduct an assessment in small areas, which can influence the goodness of the results presented.},
keywords = {Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Currently there is no standard indoor positioning system, similar to outdoor GPS. However, WiFi signals have been used in a large number of proposals to achieve the above positioning, many of which use machine learning to do so. But what are the most commonly used techniques in machine learning? What accuracy do they achieve? Where have they been tested? This article presents a systematic review of works between 2019 and 2021 that use WiFi as the signal for positioning and machine learning models to estimate indoor position. 64 papers have been identified as relevant, which have been systematically analyzed for a better understanding of the current situation in different aspects. The results show that indoor positioning based on WiFi trends use neural network-based models, evaluated in empirical experiments. Despite this, many works still conduct an assessment in small areas, which can influence the goodness of the results presented.
Rodriguez-Martinez, Cristina; Torres-Sospedra, Joaquín
Revisiting the Analysis of Hyperparameters in k-NN for Wi-Fi and BLE Fingerprinting: Current Status and General Results Proceedings Article
In: Proceedings of the Eleventh International Conference on Indoor Positioning and Indoor Navigation, IEEE, 2021.
Abstract | Links | BibTeX | Tags: machine learning, Wi-Fi fingerprint
@inproceedings{Rodriguez2021a,
title = {Revisiting the Analysis of Hyperparameters in k-NN for Wi-Fi and BLE Fingerprinting: Current Status and General Results},
author = {Cristina Rodriguez-Martinez and Joaquín Torres-Sospedra},
doi = {https://doi.org/10.1109/IPIN51156.2021.9662542},
year = {2021},
date = {2021-12-15},
booktitle = {Proceedings of the Eleventh International Conference on Indoor Positioning and Indoor Navigation},
publisher = {IEEE},
abstract = {Wi-Fi Fingerprinting is a very popular technique in the field of indoor positioning, since the release of Microsoft RADAR system back in 2000. Since that milestone, the vast majority of studies and improvements in this field keep using the same base algorithm, an adaptation of the k-NN algorithm to treat geospatial data (e.g., x/y or lat/lon). One of the most relevant drawbacks of k-NN algorithm resides in its initial design, focused on resolving general classification problems. Wi-Fi fingerprinting technique is based on the measurement of the signal strength emitted by close and available Wi-Fi networks. However, the nature of signal propagation is not linear, and it is impacted by the fixed and dynamic obstacles present in the environment. This work consists in the study of k-NN algorithm parameters, k value, distance metric and data representation, to improve the efficiency of this prediction model. The evaluation will be conducted over several different heterogeneous databases and propose a model to automatically set the value of k.},
keywords = {machine learning, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {inproceedings}
}
Wi-Fi Fingerprinting is a very popular technique in the field of indoor positioning, since the release of Microsoft RADAR system back in 2000. Since that milestone, the vast majority of studies and improvements in this field keep using the same base algorithm, an adaptation of the k-NN algorithm to treat geospatial data (e.g., x/y or lat/lon). One of the most relevant drawbacks of k-NN algorithm resides in its initial design, focused on resolving general classification problems. Wi-Fi fingerprinting technique is based on the measurement of the signal strength emitted by close and available Wi-Fi networks. However, the nature of signal propagation is not linear, and it is impacted by the fixed and dynamic obstacles present in the environment. This work consists in the study of k-NN algorithm parameters, k value, distance metric and data representation, to improve the efficiency of this prediction model. The evaluation will be conducted over several different heterogeneous databases and propose a model to automatically set the value of k.
Sánchez-Pozo, Nadia N; Trilles-Oliver, Sergio; Solé-Ribalta, Albert; Lorente-Leyva, Leandro L.; Mayorca-Torres, Dagoberto; Peluffo-Ordóñez, Diego H
Algorithms Air Quality Estimation: A Comparative Study of Stochastic and Heuristic Predictive Proceedings Article
In: Hybrid Artificial Intelligent Systems (International Conference on Hybrid Artificial Intelligence Systems), pp. 293-304, Springer, Cham, 2021, ISBN: 978-3-030-86271-8.
Abstract | Links | BibTeX | Tags: air quality prediction, machine learning
@inproceedings{SanchezPozo2021a,
title = {Algorithms Air Quality Estimation: A Comparative Study of Stochastic and Heuristic Predictive},
author = {Nadia N Sánchez-Pozo and Sergio Trilles-Oliver and Albert Solé-Ribalta and Leandro L. Lorente-Leyva and Dagoberto Mayorca-Torres and Diego H Peluffo-Ordóñez},
doi = {https://doi.org/10.1007/978-3-030-86271-8_25},
isbn = {978-3-030-86271-8},
year = {2021},
date = {2021-09-15},
booktitle = {Hybrid Artificial Intelligent Systems (International Conference on Hybrid Artificial Intelligence Systems)},
pages = {293-304},
publisher = {Springer, Cham},
abstract = {This paper presents a comparative analysis of predictive models applied to air quality estimation. Currently, among other global issues, there is a high concern about air pollution, for this reason, there are several air quality indicators, with carbon monoxide (CO), sulfur dioxide (SO2), nitrogen dioxide (NO2) and ozone (O3) being the main ones. When the concentration level of an indicator exceeds an established air quality safety threshold, it is considered harmful to human health, therefore, in cities like London, there are monitoring systems for air pollutants. This study aims to compare the efficiency of stochastic and heuristic predictive models for forecasting ozone (O3) concentration to estimate London's air quality by analyzing an open dataset retrieved from the London Datastore portal. Models based on data analysis have been widely used in air quality forecasting. This paper develops four predictive models (autoregressive integrated moving average - ARIMA, support vector regression - SVR, neural networks (specifically, long-short term memory - LSTM) and Facebook Prophet). Experimentally, ARIMA models and LSTM are proved to reach the highest accuracy in predicting the concentration of air pollutants among the considered models. As a result, the comparative analysis of the loss function (root-mean-square error) reveled that ARIMA and LSTM are the most suitable, accomplishing a low error rate of 0.18 and 0.20, respectively.},
keywords = {air quality prediction, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper presents a comparative analysis of predictive models applied to air quality estimation. Currently, among other global issues, there is a high concern about air pollution, for this reason, there are several air quality indicators, with carbon monoxide (CO), sulfur dioxide (SO2), nitrogen dioxide (NO2) and ozone (O3) being the main ones. When the concentration level of an indicator exceeds an established air quality safety threshold, it is considered harmful to human health, therefore, in cities like London, there are monitoring systems for air pollutants. This study aims to compare the efficiency of stochastic and heuristic predictive models for forecasting ozone (O3) concentration to estimate London's air quality by analyzing an open dataset retrieved from the London Datastore portal. Models based on data analysis have been widely used in air quality forecasting. This paper develops four predictive models (autoregressive integrated moving average - ARIMA, support vector regression - SVR, neural networks (specifically, long-short term memory - LSTM) and Facebook Prophet). Experimentally, ARIMA models and LSTM are proved to reach the highest accuracy in predicting the concentration of air pollutants among the considered models. As a result, the comparative analysis of the loss function (root-mean-square error) reveled that ARIMA and LSTM are the most suitable, accomplishing a low error rate of 0.18 and 0.20, respectively.
Chukhno, Nadezhda; Chukhno, Olga; Pizzi, Sara; Molinaro, Antonella; Iera, Antonio; Araniti, Giuseppe
Unsupervised Learning for D2D-Assisted Multicast Scheduling in mmWave Networks Proceedings Article
In: 2021 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, pp. 1-6, IEEE, 2021, ISBN: 978-1-6654-4909-0.
Abstract | Links | BibTeX | Tags: A-wear, machine learning, wearables
@inproceedings{Chukhno2021b,
title = {Unsupervised Learning for D2D-Assisted Multicast Scheduling in mmWave Networks},
author = {Nadezhda Chukhno and Olga Chukhno and Sara Pizzi and Antonella Molinaro and Antonio Iera and Giuseppe Araniti},
doi = {10.1109/BMSB53066.2021.9547189},
isbn = {978-1-6654-4909-0},
year = {2021},
date = {2021-08-08},
booktitle = {2021 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting},
pages = {1-6},
publisher = {IEEE},
abstract = {The combination of multicast and directional mmWave communication paves the way for solving spectrum crunch problems, increasing spectrum efficiency, ensuring reliability, and reducing access point load. Furthermore, multi-hop relaying is considered as one of the key interest areas in future 5G+ systems to achieve enhanced system performance. Based on this approach, users located close to the base station may serve as relays towards cell-edge users in their proximity by using more robust device-to-device (D2D) links, which is essential, e.g., to reduce the power consumption for wearable devices. In this paper, we account for the limitations and capabilities of directional mmWave multicast systems by proposing a low-complexity heuristic solution that leverages an unsupervised machine learning algorithm for multicast group formation and by exploiting the D2D technology to deal with the blockage problem.},
keywords = {A-wear, machine learning, wearables},
pubstate = {published},
tppubtype = {inproceedings}
}
The combination of multicast and directional mmWave communication paves the way for solving spectrum crunch problems, increasing spectrum efficiency, ensuring reliability, and reducing access point load. Furthermore, multi-hop relaying is considered as one of the key interest areas in future 5G+ systems to achieve enhanced system performance. Based on this approach, users located close to the base station may serve as relays towards cell-edge users in their proximity by using more robust device-to-device (D2D) links, which is essential, e.g., to reduce the power consumption for wearable devices. In this paper, we account for the limitations and capabilities of directional mmWave multicast systems by proposing a low-complexity heuristic solution that leverages an unsupervised machine learning algorithm for multicast group formation and by exploiting the D2D technology to deal with the blockage problem.
Pascacio-de-los-Santos, Pavel; Casteleyn, Sven; Torres-Sospedra, Joaquín
Smartphone Distance Estimation Based on RSSI-Fuzzy Classification Approach Proceedings Article
In: Proceedings of the 2021 International Conference on Localization and GNSS (ICL-GNSS), pp. 1-6, IEEE, 2021, ISBN: 978-1-7281-9645-9.
Abstract | Links | BibTeX | Tags: Indoor positioning, machine learning
@inproceedings{Pascacio2021b,
title = {Smartphone Distance Estimation Based on RSSI-Fuzzy Classification Approach},
author = {Pavel Pascacio-de-los-Santos and Sven Casteleyn and Joaquín Torres-Sospedra},
doi = {https://doi.org/10.1109/ICL-GNSS51451.2021.9452226},
isbn = {978-1-7281-9645-9},
year = {2021},
date = {2021-06-01},
booktitle = {Proceedings of the 2021 International Conference on Localization and GNSS (ICL-GNSS)},
pages = {1-6},
publisher = {IEEE},
abstract = {Positioning people indoors has known an exponential growth in the last few years, especially thanks to Bluetooth Low Energy (BLE) technology and the Received Signal Strength Indicator (RSSI) technique. This approach is available in wearable devices, is easy to implement and has energy consumption advantages. However, the relative distance calculation is inaccurate, as the strength of BLE signals significantly fluctuates in indoor environments. Typical coping mechanisms, such as path-loss propagation models, require mathematical modeling and time-consuming calibration, that depend on the environment. In this paper, we propose a novel distance estimator based on RSSI-fuzzy classification of the BLE signals. Fuzzy-logic improves the robustness and accuracy of RSSI-based estimators, does not require an explicit propagation model and is easy and intuitive to (graphically) tune (using basic statistical analysis). The estimator's suitability and the feasibility to provide an easy implementation were experimentally demonstrated in two scenarios with real-world data.},
keywords = {Indoor positioning, machine learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Positioning people indoors has known an exponential growth in the last few years, especially thanks to Bluetooth Low Energy (BLE) technology and the Received Signal Strength Indicator (RSSI) technique. This approach is available in wearable devices, is easy to implement and has energy consumption advantages. However, the relative distance calculation is inaccurate, as the strength of BLE signals significantly fluctuates in indoor environments. Typical coping mechanisms, such as path-loss propagation models, require mathematical modeling and time-consuming calibration, that depend on the environment. In this paper, we propose a novel distance estimator based on RSSI-fuzzy classification of the BLE signals. Fuzzy-logic improves the robustness and accuracy of RSSI-based estimators, does not require an explicit propagation model and is easy and intuitive to (graphically) tune (using basic statistical analysis). The estimator's suitability and the feasibility to provide an easy implementation were experimentally demonstrated in two scenarios with real-world data.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Features Exploration from Datasets Vision in Air Quality Prediction Domain Journal Article
In: Atmosphere, vol. 12, no. 3, pp. 312, 2021, ISSN: 2073-4433.
Abstract | Links | BibTeX | Tags: air quality prediction, data, machine learning
@article{Iskandaryan2021a,
title = {Features Exploration from Datasets Vision in Air Quality Prediction Domain},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.3390/atmos12030312},
issn = {2073-4433},
year = {2021},
date = {2021-02-28},
journal = {Atmosphere},
volume = {12},
number = {3},
pages = {312},
abstract = {Air pollution and its consequences are negatively impacting on the world population and the environment, which converts the monitoring and forecasting air quality techniques as essential tools to combat this problem. To predict air quality with maximum accuracy, along with the implemented models and the quantity of the data, it is crucial also to consider the dataset types. This study selected a set of research works in the field of air quality prediction and is concentrated on the exploration of the datasets utilised in them. The most significant findings of this research work are: (1) meteorological datasets were used in 94.6% of the papers leaving behind the rest of the datasets with a big difference, which is complemented with others, such as temporal data, spatial data, and so on; (2) the usage of various datasets combinations has been commenced since 2009; and (3) the utilisation of open data have been started since 2012, 32.3% of the studies used open data, and 63.4% of the studies did not provide the data.},
keywords = {air quality prediction, data, machine learning},
pubstate = {published},
tppubtype = {article}
}
Air pollution and its consequences are negatively impacting on the world population and the environment, which converts the monitoring and forecasting air quality techniques as essential tools to combat this problem. To predict air quality with maximum accuracy, along with the implemented models and the quantity of the data, it is crucial also to consider the dataset types. This study selected a set of research works in the field of air quality prediction and is concentrated on the exploration of the datasets utilised in them. The most significant findings of this research work are: (1) meteorological datasets were used in 94.6% of the papers leaving behind the rest of the datasets with a big difference, which is complemented with others, such as temporal data, spatial data, and so on; (2) the usage of various datasets combinations has been commenced since 2009; and (3) the utilisation of open data have been started since 2012, 32.3% of the studies used open data, and 63.4% of the studies did not provide the data.
2020
Klus, Roman; Klus, Lucie; Solomitckii, Dmitrii; Talvitie, Jukka; Valkama, Mikko
Deep Learning-Based Cell-Level and Beam-Level Mobility Management System Journal Article
In: Sensors, vol. 20, no. 24, pp. 7124, 2020, ISSN: 1424-8220.
Abstract | Links | BibTeX | Tags: A-wear, machine learning, urban mobility
@article{Klus2020a,
title = {Deep Learning-Based Cell-Level and Beam-Level Mobility Management System},
author = {Roman Klus and Lucie Klus and Dmitrii Solomitckii and Jukka Talvitie and Mikko Valkama},
doi = {https://doi.org/10.3390/s20247124},
issn = {1424-8220},
year = {2020},
date = {2020-12-11},
journal = {Sensors},
volume = {20},
number = {24},
pages = {7124},
abstract = {The deployment with beamforming-capable base stations in 5G New Radio (NR) requires an efficient mobility management system to reliably operate with minimum effort and interruption. In this work, we propose two artificial neural network models to optimize the cell-level and beam-level mobility management. Both models consist of convolutional, as well as dense, layer blocks. Based on current and past received power measurements, as well as positioning information, they choose the optimum serving cell and serving beam, respectively. The obtained results show that the proposed cell-level mobility model is able to sustain a strong serving cell and reduce the number of handovers by up to 94.4% compared to the benchmark solution when the uncertainty (representing shadowing, interference, etc.) is introduced to the received signal strength measurements. The proposed beam-level mobility management model is able to proactively choose and sustain the strongest serving beam, even when high uncertainty is introduced to the measurements.},
keywords = {A-wear, machine learning, urban mobility},
pubstate = {published},
tppubtype = {article}
}
The deployment with beamforming-capable base stations in 5G New Radio (NR) requires an efficient mobility management system to reliably operate with minimum effort and interruption. In this work, we propose two artificial neural network models to optimize the cell-level and beam-level mobility management. Both models consist of convolutional, as well as dense, layer blocks. Based on current and past received power measurements, as well as positioning information, they choose the optimum serving cell and serving beam, respectively. The obtained results show that the proposed cell-level mobility model is able to sustain a strong serving cell and reduce the number of handovers by up to 94.4% compared to the benchmark solution when the uncertainty (representing shadowing, interference, etc.) is introduced to the received signal strength measurements. The proposed beam-level mobility management model is able to proactively choose and sustain the strongest serving beam, even when high uncertainty is introduced to the measurements.
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Trilles-Oliver, Sergio
Air quality prediction in smart cities using machine learning technologies based on sensor data: A Review Journal Article
In: Applied sciences, vol. 10, no. 7, pp. 2401, 2020.
Abstract | Links | BibTeX | Tags: Air quality sensors, machine learning, Sensors, Smart Cities
@article{Iskandaryan2020,
title = {Air quality prediction in smart cities using machine learning technologies based on sensor data: A Review},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Sergio Trilles-Oliver},
doi = {https://doi.org/10.3390/app10072401 },
year = {2020},
date = {2020-04-15},
journal = {Applied sciences},
volume = {10},
number = {7},
pages = {2401},
abstract = {The influence of machine learning technologies is rapidly increasing and penetrating almost in every field, and air pollution prediction is not being excluded from those fields. This paper covers the revision of the studies related to air pollution prediction using machine learning algorithms based on sensor data in the context of smart cities. Using the most popular databases and executing the corresponding filtration, the most relevant papers were selected. After thorough reviewing those papers, the main features were extracted, which served as a base to link and compare them to each other. As a result, we can conclude that: (1) instead of using simple machine learning techniques, currently, the authors apply advanced and sophisticated techniques, (2) China was the leading country in terms of a case study, (3) Particulate matter with diameter equal to 2.5 micrometers was the main prediction target, (4) in 41% of the publications the authors carried out the prediction for the next day, (5) 66% of the studies used data had an hourly rate, (6) 49% of the papers used open data and since 2016 it had a tendency to increase, and (7) for efficient air quality prediction it is important to consider the external factors such as weather conditions, spatial characteristics, and temporal features},
keywords = {Air quality sensors, machine learning, Sensors, Smart Cities},
pubstate = {published},
tppubtype = {article}
}
The influence of machine learning technologies is rapidly increasing and penetrating almost in every field, and air pollution prediction is not being excluded from those fields. This paper covers the revision of the studies related to air pollution prediction using machine learning algorithms based on sensor data in the context of smart cities. Using the most popular databases and executing the corresponding filtration, the most relevant papers were selected. After thorough reviewing those papers, the main features were extracted, which served as a base to link and compare them to each other. As a result, we can conclude that: (1) instead of using simple machine learning techniques, currently, the authors apply advanced and sophisticated techniques, (2) China was the leading country in terms of a case study, (3) Particulate matter with diameter equal to 2.5 micrometers was the main prediction target, (4) in 41% of the publications the authors carried out the prediction for the next day, (5) 66% of the studies used data had an hourly rate, (6) 49% of the papers used open data and since 2016 it had a tendency to increase, and (7) for efficient air quality prediction it is important to consider the external factors such as weather conditions, spatial characteristics, and temporal features
Petkov, Mihail
Universidade Nova De Lisboa, Lisboa, 2020.
Abstract | Links | BibTeX | Tags: Geographic Information Systems (GIS), machine learning, Mastergeotech
@mastersthesis{Petkov2020,
title = {Evaluation of spatial data’s impact in mid-term room rent price through application of spatial econometrics and machine learning. Case study: Lisbon},
author = {Mihail Petkov},
editor = {Roberto Henriques and Joel Ferreira da Silva and Carlos Granell-Canut (Supervisors)
},
url = {http://hdl.handle.net/10362/93716},
year = {2020},
date = {2020-02-28},
address = {Lisboa},
school = {Universidade Nova De Lisboa},
abstract = {Household preferences is a topic whose relevance can be found to dominate the applied economics, but whereas urban economies view cities as production centers, this thesis aims to give importance to the role of consumption. Provision to PoIs might give explanation to what individuals value as an important asset for improvement of their quality of life in a chosen city. As such, understanding short-term rentals and real estate prices have induced various research to seek proof of impacting factors, but analysis of mid-term rent has faced the challenge of being an overlooked category. This thesis consists of an integrated three-steps approach to analyze spatial data’s impact over the mid-term room rent, choosing Lisbon as its case study. The proposed methodology constitutes use of traditional spatial econometric models and SVR, encompassing a large set of proxies for amenities that might be recognized to hold a possible impact over rent prices. The analytical frameworks’ first step is to create a suitable HPM model that captures the data well, so significant variables can be detected and analyzed as a discrete dataset. The second step applies subsets of the dataset in the creation of SVR models, in hopes of identifying the SVs influencing price variances. Finally, SOM clusters are chosen to address whether more natural order of data division exists. Results confirm the impact of proximity to various categories of amenities, but the enrichment of models with the proposed proxies of spatial data failed to corroborate attainment of model with a higher accuracy. (Nüst et al., 2018) provides a self-assessment of the reproducibility of research, and according to the criteria given, this dissertation is evaluated as: 0, 2, 1, 2, 2 (input data, preprocessing, methods, computational environment, results).},
keywords = {Geographic Information Systems (GIS), machine learning, Mastergeotech},
pubstate = {published},
tppubtype = {mastersthesis}
}
Household preferences is a topic whose relevance can be found to dominate the applied economics, but whereas urban economies view cities as production centers, this thesis aims to give importance to the role of consumption. Provision to PoIs might give explanation to what individuals value as an important asset for improvement of their quality of life in a chosen city. As such, understanding short-term rentals and real estate prices have induced various research to seek proof of impacting factors, but analysis of mid-term rent has faced the challenge of being an overlooked category. This thesis consists of an integrated three-steps approach to analyze spatial data’s impact over the mid-term room rent, choosing Lisbon as its case study. The proposed methodology constitutes use of traditional spatial econometric models and SVR, encompassing a large set of proxies for amenities that might be recognized to hold a possible impact over rent prices. The analytical frameworks’ first step is to create a suitable HPM model that captures the data well, so significant variables can be detected and analyzed as a discrete dataset. The second step applies subsets of the dataset in the creation of SVR models, in hopes of identifying the SVs influencing price variances. Finally, SOM clusters are chosen to address whether more natural order of data division exists. Results confirm the impact of proximity to various categories of amenities, but the enrichment of models with the proposed proxies of spatial data failed to corroborate attainment of model with a higher accuracy. (Nüst et al., 2018) provides a self-assessment of the reproducibility of research, and according to the criteria given, this dissertation is evaluated as: 0, 2, 1, 2, 2 (input data, preprocessing, methods, computational environment, results).
Iskandaryan, Ditsuhi; Ramos-Romero, Francisco; Palinggi, Dany Asarias; Trilles-Oliver, Sergio
The effect of weather in soccer results: An approach using machine learning techniques Journal Article
In: Applied sciences, vol. 10, no. 19, pp. 6750, 2020.
Links | BibTeX | Tags: machine learning
@article{Iskandaryan2020b,
title = {The effect of weather in soccer results: An approach using machine learning techniques},
author = {Ditsuhi Iskandaryan and Francisco Ramos-Romero and Dany Asarias Palinggi and Sergio Trilles-Oliver},
doi = {https://doi.org/10.3390/app10196750},
year = {2020},
date = {2020-01-31},
journal = {Applied sciences},
volume = {10},
number = {19},
pages = {6750},
keywords = {machine learning},
pubstate = {published},
tppubtype = {article}
}
2019
Rojo, Jordi; Mendoza-Silva, Germán Martín; Cidral, Gabriel Ristow; Laiapea, Jorma; Parrello, Gerardo; Simó, Arnau; Stupin, Laura; Minican, Deniz; Farrés, María; Corvalán, Carmen; Unger, Florian; López, Sara Marín-; Soteras, Ignacio; Bravo, Daniel Castejón; Torres-Sospedra, Joaquín
Machine Learning applied to Wi-Fi fingerprinting: The experiences of the Ubiqum Challenge Proceedings Article
In: Proceedings of the Tenth International Conference on Indoor Positioning and Indoor Navigation, 30 Sept. – 3 Oct. 2019, Pisa, Italy. , IEEE, 2019, ISBN: 978-1-7281-1788-1 .
Abstract | BibTeX | Tags: machine learning, Wi-Fi fingerprint
@inproceedings{Rojo2019,
title = {Machine Learning applied to Wi-Fi fingerprinting: The experiences of the Ubiqum Challenge },
author = {Jordi Rojo and Germán Martín Mendoza-Silva and Gabriel Ristow Cidral and Jorma Laiapea and Gerardo Parrello and Arnau Simó and Laura Stupin and Deniz Minican and María Farrés and Carmen Corvalán and
Florian Unger and Sara Marín- López and Ignacio Soteras and Daniel Castejón Bravo and Joaquín Torres-Sospedra},
isbn = { 978-1-7281-1788-1 },
year = {2019},
date = {2019-12-01},
booktitle = {Proceedings of the Tenth International Conference on Indoor Positioning and Indoor Navigation, 30 Sept. – 3 Oct. 2019, Pisa, Italy. },
publisher = {IEEE},
abstract = {Wi-Fi Fingerprinting is widely adopted for smartphone-based indoor positioning systems due to the availability of already deployed infrastructure for communications. The UJIIndoorLoc database contains Wi-Fi data for indoor positioning in a large environment covering three multi-tier buildings collected with multiple devices. Since the evaluation set is private, the indoor positioning systems of developers and researchers can still be evaluated under the same evaluation conditions than the participants of the 2015 EvAAL-ETRI competition. This paper shows the results and the experiences of such kind of external evaluation based on a competition provided by the the students of the “Data Analytics and Machine Learning” program of the Ubiqum data academy, who applied machine learning models they learnt during the program. The results show that state-ofart Machine Learning methods provide good positioning results, but expertise on the problem is still needed},
keywords = {machine learning, Wi-Fi fingerprint},
pubstate = {published},
tppubtype = {inproceedings}
}
Wi-Fi Fingerprinting is widely adopted for smartphone-based indoor positioning systems due to the availability of already deployed infrastructure for communications. The UJIIndoorLoc database contains Wi-Fi data for indoor positioning in a large environment covering three multi-tier buildings collected with multiple devices. Since the evaluation set is private, the indoor positioning systems of developers and researchers can still be evaluated under the same evaluation conditions than the participants of the 2015 EvAAL-ETRI competition. This paper shows the results and the experiences of such kind of external evaluation based on a competition provided by the the students of the “Data Analytics and Machine Learning” program of the Ubiqum data academy, who applied machine learning models they learnt during the program. The results show that state-ofart Machine Learning methods provide good positioning results, but expertise on the problem is still needed
Palinggi, Denny Asarias
Predicting Soccer Outcome with Machine Learning Based on Weather Condition Masters Thesis
INIT, Castellón, 2019.
BibTeX | Tags: machine learning, Mastergeotech, sports
@mastersthesis{Palinggi2019,
title = {Predicting Soccer Outcome with Machine Learning Based on Weather Condition},
author = {Denny Asarias Palinggi},
editor = {Francisco Ramos-Romero and Roberto Henriques and Roberto André Pereira and Jorge Mateu-Mahiques},
year = {2019},
date = {2019-03-04},
address = {Castellón},
school = {INIT},
keywords = {machine learning, Mastergeotech, sports},
pubstate = {published},
tppubtype = {mastersthesis}
}
2018
Belmonte-Fernández, Óscar; Montoliu, Raúl; Torres-Sospedra, Joaquín; Sansano-Sansano, Emilio; Chia-Aguilar, Daniel
A radiosity-based method to avoid calibration for indoor positioning systems Journal Article
In: Expert Systems with Applications, vol. 105, pp. 89 - 101, 2018, ISSN: 0957-4174.
Links | BibTeX | Tags: Classification algorithm, Indoor positioning, machine learning, Radiosity
@article{BELMONTEFERNANDEZ201889,
title = {A radiosity-based method to avoid calibration for indoor positioning systems},
author = {Óscar Belmonte-Fernández and Raúl Montoliu and Joaquín Torres-Sospedra and Emilio Sansano-Sansano and Daniel Chia-Aguilar},
url = {http://www.sciencedirect.com/science/article/pii/S0957417418302112},
doi = {https://doi.org/10.1016/j.eswa.2018.03.054},
issn = {0957-4174},
year = {2018},
date = {2018-09-01},
journal = {Expert Systems with Applications},
volume = {105},
pages = {89 - 101},
keywords = {Classification algorithm, Indoor positioning, machine learning, Radiosity},
pubstate = {published},
tppubtype = {article}
}
Twanabasu, Bikesh
Sentiment Analysis in Geo Social Streams by using Machine Learning Techniques Masters Thesis
Departamento de Lenguajes y Sistemas Informáticos, Castellón, 2018.
BibTeX | Tags: machine learning, Mastergeotech, sentiment analysis
@mastersthesis{Twanabasu2018,
title = {Sentiment Analysis in Geo Social Streams by using Machine Learning Techniques},
author = {Bikesh Twanabasu},
editor = {Francisco Ramos-Romero and Óscar Belmonte-Fernández and Roberto Henriques (supervisor)},
year = {2018},
date = {2018-03-02},
address = {Castellón},
school = {Departamento de Lenguajes y Sistemas Informáticos},
keywords = {machine learning, Mastergeotech, sentiment analysis},
pubstate = {published},
tppubtype = {mastersthesis}
}
2016
Calia, Andrea
Viability assessment of WPS 2.0 services as communication standard for expensive web-based machine learning analysis. A case of study: Indoor Location Masters Thesis
Universitat Jaume I, 2016.
Abstract | BibTeX | Tags: Indoor positioning, machine learning, Mastergeotech, WPS
@mastersthesis{Calia2016,
title = {Viability assessment of WPS 2.0 services as communication standard for expensive web-based machine learning analysis. A case of study: Indoor Location},
author = { Andrea Calia},
editor = {Óscar Belmonte Fernández (supervisor) and Marco Painho (co-supervisor) and Raúl Montoliu Colás (co-supervisor)},
year = {2016},
date = {2016-03-07},
school = {Universitat Jaume I},
abstract = {Communication between client and server is a key factor in the modern age. Nowadays, telecommunications are at the base of every system and Software that is available. The way Software communicates can determine the efficacy of it. In the GIS world, a server is often used for offloading expensive tasks such as geospatial operations or statistical analysis. This technique improves the performance of the Software systems and makes them able to scale based on the demand on real time. For making the communication between client and server more efficient, interoperable and standard, the OGC released the standard WPS. WPS defines abstract operations that are able to describe a client server communication for remote process executions. This thesis focuses on the asynchronous execution feature introduced in the version 2.0 of WPS. The main goal is to study how asynchronous process execution can benefit a client both in performance and availability. The result are promising and it is demonstrated that WPS is a solid standard for describing web services operations. Based on the obtained results, future studies can extend the standard in order to make it more general and suitable for more situations.},
keywords = {Indoor positioning, machine learning, Mastergeotech, WPS},
pubstate = {published},
tppubtype = {mastersthesis}
}
Communication between client and server is a key factor in the modern age. Nowadays, telecommunications are at the base of every system and Software that is available. The way Software communicates can determine the efficacy of it. In the GIS world, a server is often used for offloading expensive tasks such as geospatial operations or statistical analysis. This technique improves the performance of the Software systems and makes them able to scale based on the demand on real time. For making the communication between client and server more efficient, interoperable and standard, the OGC released the standard WPS. WPS defines abstract operations that are able to describe a client server communication for remote process executions. This thesis focuses on the asynchronous execution feature introduced in the version 2.0 of WPS. The main goal is to study how asynchronous process execution can benefit a client both in performance and availability. The result are promising and it is demonstrated that WPS is a solid standard for describing web services operations. Based on the obtained results, future studies can extend the standard in order to make it more general and suitable for more situations.