Integrating Features for Accelerometer-based Activity Recognition

Erdaş, Çağatay Berke; Atasoy, Işıl; Açıcı, Koray; Ogul, Hasan

doi:10.1016/j.procs.2016.09.070

Cited by 68 publications

(31 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular, good balance between accuracy and extraction cost is provided by basic time-domain features [17,37], which include: (i) maximum; (ii) median; (iii) minimum; (iv) mean; (v) and variance of the data sequence along the x, y and z axes. Such features, indeed, provide a characterisation of the central tendency of the data distribution (e.g., mean, median), as well as of its dispersion (e.g., maximum, minimum).…”

Section: Feature Extractionmentioning

confidence: 99%

From Bits of Data to Bits of Knowledge—An On-Board Classification Framework for Wearable Sensing Systems

Zalewski

Marchegiani

Elsts

et al. 2020

Sensors

View full text Add to dashboard Cite

Wearable systems constitute a promising solution to the emerging challenges of healthcare provision, feeding machine learning frameworks with necessary data. In practice, however, raw data collection is expensive in terms of energy, and therefore imposes a significant maintenance burden to the user, which in turn results in poor user experience, as well as significant data loss due to improper battery maintenance. In this paper, we propose a framework for on-board activity classification targeting severely energy-constrained wearable systems. The proposed framework leverages embedded classifiers to activate power-hungry sensing elements only when they are useful, and to distil the raw data into knowledge that is eventually transmitted over the air. We implement the proposed framework on a prototype wearable system and demonstrate that it can decrease the energy requirements by one order of magnitude, yielding high classification accuracy that is reduced by approximately 5%, as compared to a cloud-based reference system.

show abstract

Section: Feature Extractionmentioning

confidence: 99%

From Bits of Data to Bits of Knowledge—An On-Board Classification Framework for Wearable Sensing Systems

Zalewski

Marchegiani

Elsts

et al. 2020

Sensors

View full text Add to dashboard Cite

show abstract

“…Another widespread utility of acceleration data is fall detection, which is extremely important for elderly health monitoring. Numerous systems of this nature have been proposed, including [34,35]. In addition, Roy et al [36] present an infrastructure-assisted smart phone-based activity recognition system in multiinhabitant smart environment.…”

Section: Physical Activity Recognition With Wearablesmentioning

confidence: 99%

Wearable IoT enabled real-time health monitoring system

Wan

Al-awlaqi

Li³

et al. 2018

J Wireless Com Network

184

View full text Add to dashboard Cite

As the age profile of many societies continues to increase, in addition to the increasing population of people affected by chronic diseases, including diabetes, cardiovascular disease, obesity, and so on, supporting health, both mentally and physically, is of increasing importance if independent living is to be maintained. Sensing, remote health monitoring, and, ultimately, recognising activities of daily living have been an promising solution. From a technical perspective, the Internet of Things (IoT) is gaining a rapidly growing attention in many disciplines, especially in personalised healthcare. Meanwhile, body area sensor network (BASN) under the IoT framework has been widely applied for ubiquitous health monitoring, for example. ECG monitoring has been commonly adopted as vital approach for diagnosing heart disease. The main contribution of this paper include the following: firstly, this paper presents a novel system, the WISE (Wearable IoT-cloud-baSed hEalth monitoring system), for real-time personal health monitoring. WISE adopts the BASN (body area sensor network) framework in the support of real-time health monitoring. Several wearable sensors have been embedded, including the heartbeat, body temperature, and the blood pressure sensors. Secondly, the majority of existing wearable health monitoring systems requisite a smart phone as data processing, visualisation, and transmission gateway, which will indeed impact the normal daily use of the smart phone. Whilst in WISE, data gathered from the BASN are directly transmitted to the cloud, and a lightweight wearable LCD can be embedded as an alternative solution for quick view of the real-time data.

show abstract

“…In this study, we compared several supervised classification methods, which have been shown to be successful in various applications for our particular tasks. Considering their common use and reported success in several domains [29,30], we chose four among these algorithms and evaluated their classification performance for our tasks: k-Nearest Neighbors (kNN), AdaBoost, Decision Tree (DT) and Random Forest (RF).…”

Section: Classificationmentioning

confidence: 99%

HANDY: A Benchmark Dataset for Context-Awareness via Wrist-Worn Motion Sensors

Açıcı

Erdaş

Asuroglu

et al. 2018

Data

Self Cite

View full text Add to dashboard Cite

Being aware of a personal context is a promising task for various applications, such as biometry, human-computer interactions, telemonitoring, remote care, mobile marketing and security. The task can be formally defined as the classification of a person being considered into one of predefined labels, which may correspond to his/her identity, gender, physical properties, the activity that he/she performs or any other attribute related to the environment being involved. Here, we offer a solution to the problem with a set of multiple motion sensors worn on the wrist. We first provide an annotated and publicly accessible benchmark set for context-awareness through wrist-worn sensors, namely, accelerometers, magnetometers and gyroscopes. Second, we present an evaluation of recent computational methods for two relevant tasks: activity recognition and person identification from hand movements. Finally, we show that fusion of two motion sensors (i.e., accelerometers and magnetometers), leads to higher accuracy for both tasks, compared with the individual use of each sensor type.

show abstract

Integrating Features for Accelerometer-based Activity Recognition

Cited by 68 publications

References 9 publications

From Bits of Data to Bits of Knowledge—An On-Board Classification Framework for Wearable Sensing Systems

From Bits of Data to Bits of Knowledge—An On-Board Classification Framework for Wearable Sensing Systems

Wearable IoT enabled real-time health monitoring system

HANDY: A Benchmark Dataset for Context-Awareness via Wrist-Worn Motion Sensors

Contact Info

Product

Resources

About