490 research outputs found
Multi-Channel Wireless Sensor Networks: Protocols, Design and Evaluation
Full text linkPervasive systems, which are described as networked embedded systems integrated with everyday environments, are considered to have the potential to change our daily lives by creating smart surroundings and by their ubiquity, just as the Internet. In the last decade, "Wireless Sensor Networks" have appeared as one of the real-world examples of pervasive systems by combining automated sensing, embedded computing and wireless networking into tiny embedded devices.A wireless sensor network typically comprises a large number of spatially distributed, tiny, battery-operated, embedded sensor devices that are networked to cooperatively collect, process, and deliver data about a phenomenon that is of interest to the users. Traditionally, wireless sensor networks have been used for monitoring applications based on low-rate data collection with low periods of operation. Current wireless sensor networks are considered to support more complex operations ranging from target tracking to health care which require efficient and timely collection of large amounts of data. Considering the low-bandwidth, low-power operation of the radios on the sensor devices, interference and contention over the wireless medium and the energy-efficiency requirements due to the battery-operated devices, fulfilling the mentioned data-collection requirements in complex applications becomes a challenging task.This thesis focuses on the efficient delivery of large amounts of data in bandwidth-limited wireless sensor networks by making use of the multi-channel capability of the sensor radios and by using optimal routing topologies. We start with experimenting the operation of the sensor radios to characterize the behavior of multi-channel communication. We propose a set of algorithms to increase the throughput and timely delivery of the data and analyze the bounds on the data collection capacity of the wireless sensor networks. The main contributions of the thesis are listed as follows: Contribution 1 - Characteristics, challenges and the use of multi-channel communication in wireless ad hoc networks and wireless sensor networks: We review the state of the art channel assignment protocols in wireless multi-hop networks, particularly in wireless ad hoc networks and wireless sensor networks. We classify the existing solutions according to the number of transceivers required per node and according to the dynamics of the channel assignment. Since the channel assignment methods designed for general wireless ad hoc networks may not be directly applicable to wireless sensor networks, we give brief comparisons of them and discuss the additional challenges and requirements for wireless sensor networks. Contribution 2 - Characterization of multi-channel interference: The assumption of perfectly orthogonal, interference-free channels, which is adopted in most of the multi-channel communication studies, may fail in practice. Radio signals are not limited to their allocated frequency band, but cause interference in adjacent bands as well —how much depends on the filtering characteristics of the transceivers. We conduct an extensive set of experiments, using NrF905 radio, to investigate the properties of multi-channel communication in wireless sensor networks. Based on these experiments, we explore an analytical model on the interference characteristics and by using the analytical model we discuss the impact of channel orthogonality on the network performance with extensive simulations. Contribution 3 - Design and implementation of a multi-channel MAC protocol for wireless sensor networks: We design a multi-channel MAC protocol, namely MCLMAC (Multi-Channel Lightweight Medium Access Control), which is a schedule based multi-channel MAC protocol that takes advantage of interference and collision free parallel transmissions over different channels. MC-LMAC is designed to provide high throughput and high delivery ratio during high-rate traffic whereas it also meets the traditional requirements of wireless sensor networks such as energy efficiency and scalability. Contribution 4 - Enhancing the rate of aggregated data collection: We consider enhancing the data collection rate of aggregated convergecast, which is one of the fundamental communication patterns in wireless sensor networks. We focus on the problem of finding the fastest rate of aggregated data collection with TDMA scheduling which is equivalent to minimizing the TDMA schedule length. We explore different techniques to address this question, such as transmission power control and multi-channel communication. We show that, once multiple frequencies are employed along with spatial-reuse TDMA, the aggregated data collection rate often becomes no longer interference-limited, but rather topology-limited. Accordingly, we show that the final step to enhance the rate of periodic aggregated data collection is to use an appropriate degree-constrained tree topology. Contribution 5 - Fast convergecast scheduling in wireless sensor networks: We focus on data delivery models where data cannot be aggregated and raw sensor readings need to be relayed towards the sink node. We study the minimum time to complete the delivery of the messages in a convergecast operation. Similar to the aggregated convergecast problem, we investigate the benefits of transmission power control and multiple channels to eliminate the effects of interference. Once the interference is completely eliminated, we show that with half-duplex single-transceiver radios, the achievable schedule length is lower-bounded by max(2nk −1, N), where nk is the maximum number of nodes on any subtree and N is the number of nodes in a network organized as a tree. We study a distributed time slot assignment algorithm to achieve this bound when a suitable routing scheme over a capacitated minimal spanning tree is employed
Sensor-Based Activity Recognition and Artificial Intelligence. 10th International Workshop, iWOAR 2025. Proceedings
No full textThis book constitutes the refereed proceedings of the 10th International Workshop on Sensor-Based Activity Recognition and Artificial Intelligence, iWOAR 2025, held in Enschede, The Netherlands, during September 18-19, 2025. The 19 full papers and 12 short papers presented in this volume were carefully reviewed and selected from 42 submissions. The accepted papers were presented in the following topical sections during the workshop presentations: Next-Gen Human Activity Recognition, AI Health Tech, Data Generation and Cleaning for Robust Human-Centric AI, Wearable Monitoring for Cognitive and Physiological States, and Advanced Sensing and Interaction for Human-Centred Systems
Quantifying Digital Biomarkers for Well-Being: Stress, Anxiety, Positive and Negative Affect via Wearable Devices and Their Time-Based Predictions
No full textWearable devices have become ubiquitous, collecting rich temporal data that offers valuable insights into human activities, health monitoring, and behavior analysis. Leveraging these data, researchers have developed innovative approaches to classify and predict time-based patterns and events in human life. Time-based techniques allow the capture of intricate temporal dependencies, which is the nature of the data coming from wearable devices. This paper focuses on predicting well-being factors, such as stress, anxiety, and positive and negative affect, on the Tesserae dataset collected from office workers. We examine the performance of different methodologies, including deep-learning architectures, LSTM, ensemble techniques, Random Forest (RF), and XGBoost, and compare their performances for time-based and non-time-based versions. In time-based versions, we investigate the effect of previous records of well-being factors on the upcoming ones. The overall results show that time-based LSTM performs the best among conventional (non-time-based) RF, XGBoost, and LSTM. The performance even increases when we consider a more extended previous period, in this case, 3 past-days rather than 1 past-day to predict the next day. Furthermore, we explore the corresponding biomarkers for each well-being factor using feature ranking. The obtained rankings are compatible with the psychological literature. In this work, we validated them based on device measurements rather than subjective survey responses
Multitask Learning for Mental Health: Depression, Anxiety, Stress (DAS) Using Wearables
No full textThis study investigates the prediction of mental well-being factors—depression, stress, and anxiety—using the NetHealth dataset from college students. The research addresses four key questions, exploring the impact of digital biomarkers on these factors, their alignment with conventional psychology literature, the time-based performance of applied methods, and potential enhancements through multitask learning. The findings reveal modality rankings aligned with psychology literature, validated against paper-based studies. Improved predictions are noted with temporal considerations, and further enhanced by multitasking. Mental health multitask prediction results show aligned baseline and multitask performances, with notable enhancements using temporal aspects, particularly with the random forest (RF) classifier. Multitask learning improves outcomes for depression and stress but not anxiety using RF and XGBoost
Federated Learning on Edge Sensing Devices: A Review
Full text linkThe ability to monitor ambient characteristics, interact with them, and
derive information about the surroundings has been made possible by the rapid
proliferation of edge sensing devices like IoT, mobile, and wearable devices
and their measuring capabilities with integrated sensors. Even though these
devices are small and have less capacity for data storage and processing, they
produce vast amounts of data. Some example application areas where sensor data
is collected and processed include healthcare, environmental (including air
quality and pollution levels), automotive, industrial, aerospace, and
agricultural applications. These enormous volumes of sensing data collected
from the edge devices are analyzed using a variety of Machine Learning (ML) and
Deep Learning (DL) approaches. However, analyzing them on the cloud or a server
presents challenges related to privacy, hardware, and connectivity limitations.
Federated Learning (FL) is emerging as a solution to these problems while
preserving privacy by jointly training a model without sharing raw data. In
this paper, we review the FL strategies from the perspective of edge sensing
devices to get over the limitations of conventional machine learning
techniques. We focus on the key FL principles, software frameworks, and
testbeds. We also explore the current sensor technologies, properties of the
sensing devices and sensing applications where FL is utilized. We conclude with
a discussion on open issues and future research directions on FL for further
studie
FedOpenHAR: Federated Multi-Task Transfer Learning for Sensor-Based Human Activity Recognition
Full text linkMotion sensors integrated into wearable and mobile devices provide valuable
information about the device users. Machine learning and, recently, deep
learning techniques have been used to characterize sensor data. Mostly, a
single task, such as recognition of activities, is targeted, and the data is
processed centrally at a server or in a cloud environment. However, the same
sensor data can be utilized for multiple tasks and distributed machine-learning
techniques can be used without the requirement of the transmission of data to a
centre. This paper explores Federated Transfer Learning in a Multi-Task manner
for both sensor-based human activity recognition and device position
identification tasks. The OpenHAR framework is used to train the models, which
contains ten smaller datasets. The aim is to obtain model(s) applicable for
both tasks in different datasets, which may include only some label types.
Multiple experiments are carried in the Flower federated learning environment
using the DeepConvLSTM architecture. Results are presented for federated and
centralized versions under different parameters and restrictions. By utilizing
transfer learning and training a task-specific and personalized federated
model, we obtained a similar accuracy with training each client individually
and higher accuracy than a fully centralized approach.Comment: Subimtted to Asian Conference in Machine Learning (ACML) 2023,
Pattern Recognition in Health Analysis Workshop, 7 pages, 3 figure
THE CHANGES IN THE SURFACE OF FLAT PRESSED WOOD-PLASTIC COMPOSITES EXPOSED TO ARTIFICIAL WEATHERING
No full textIn this study, the wood flour content's effect on the weathering performance of flat pressed WPC was investigated. The high density polyethylene was reinforced with four different wood flour content (10%, 30%, 50%, 70%). The weathering performance of WPC was determined by the 400 h of artificial weathering test. According to the results, the color change is inevitable as long as the wood flour is used as filler. Surprisingly, the highest color change was obtained from WPC containing 30% WF, contrary to 70% of wood flour. Similarly, the whiteness of the surface of WPC increased with exposure time. The photooxidation resulted in the chain scission and shorter molecules, which were observed by ATR-FTIR analysis. The changes in the intensity of characteristic polymer (2914 cm(-1) and 2846 cm(-1)) and wood peaks (1510 cm(-1) and 1027 cm(-1)) exhibited the photodegradation on WPCs' surface, which resulted in color change. Moreover, the light microscopy investigation showed surface degradation. The extensive weathering conditions caused surface cracks and surface roughness. The visual appearance of WPCs also demonstrated how to change the surface character of WPC during the 400 h of artificial weathering. In conclusion, the increase in the wood content increased the intensity of degradation
Complex Human Activity Recognition Using Smartphone and Wrist-Worn Motion Sensors
Full text linkThe position of on-body motion sensors plays an important role in human activity recognition. Most often, mobile phone sensors at the trouser pocket or an equivalent position are used for this purpose. However, this position is not suitable for recognizing activities that involve hand gestures, such as smoking, eating, drinking coffee and giving a talk. To recognize such activities, wrist-worn motion sensors are used. However, these two positions are mainly used in isolation. To use richer context information, we evaluate three motion sensors (accelerometer, gyroscope and linear acceleration sensor) at both wrist and pocket positions. Using three classifiers, we show that the combination of these two positions outperforms the wrist position alone, mainly at smaller segmentation windows. Another problem is that less-repetitive activities, such as smoking, eating, giving a talk and drinking coffee, cannot be recognized easily at smaller segmentation windows unlike repetitive activities, like walking, jogging and biking. For this purpose, we evaluate the effect of seven window sizes (2–30 s) on thirteen activities and show how increasing window size affects these various activities in different ways. We also propose various optimizations to further improve the recognition of these activities. For reproducibility, we make our dataset publicly available
FedOpenHAR:Federated Multitask Transfer Learning for Sensor-Based Human Activity Recognition
No full textWearable and mobile devices equipped with motion sensors offer important insights into user behavior. Machine learning and, more recently, deep learning techniques have been applied to analyze sensor data. Typically, the focus is on a single task, such as human activity recognition (HAR), and the data is processed centrally on a server or in the cloud. However, the same sensor data can be leveraged for multiple tasks, and distributed machine learning methods can be employed without the need for transmitting data to a central location. In this study, we introduce the FedOpenHAR framework, which explores federated transfer learning in a multitask setting for both sensor-based HAR and device position identification tasks. This approach utilizes transfer learning by training task-specific and personalized layers in a federated manner. The OpenHAR framework, which includes ten smaller datasets, is used for training the models. The main challenge is developing robust models that are applicable to both tasks across different datasets, which may contain only a subset of label types. Multiple experiments are conducted in the Flower federated learning environment using the DeepConvLSTM architecture. Results are presented for both federated and centralized training under various parameters and constraints. By employing transfer learning and training task-specific and personalized federated models, we achieve a higher accuracy (72.4%) compared to a fully centralized training approach (64.5%), and similar accuracy to a scenario where each client performs individual training in isolation (72.6%). However, the advantage of FedOpenHAR over individual training is that, when a new client joins with a new label type (representing a new task), it can begin training from the already existing common layer. Furthermore, if a new client wants to classify a new class in one of the existing tasks, FedOpenHAR allows training to begin directly from the task-specific layers.</p
A Synergic Effect of Water-Based Acrylic Resin with Boric Acid on Leachability
No full textIn this study, the Scots pine wood samples were impregnated (single treatment) with boric acid combined with two types of water-based acrylic resin (pure acrylic and semi-translucent acrylic emulsion) to limit the boron leaching and improve the decay resistance. The results showed dimensional stability in anti-swelling efficiency and water absorption improved in wood specimens treated with boric acid and acrylic types. While the leachability was over 90% for only 3% boric acid-impregnated wood (control), it was calculated at 36% for acrylic emulsions-impregnated wood. Although there were no weight losses for the unleached woods, it was up to 9% for leached woods impregnated with acrylic resin and emulsion. The 25% acrylic emulsion had no weight losses after the leaching test for Coniophora puteana and Trametes versicolor. The boric acid combined with acrylic resin can improve the leaching resistance with the synergic effect, enhancing resistance against biological threats
- …
