Wireless Sensing for Human Activity: A Survey

Liu, Jian; Liu, Hongbo; Chen, Yingying; Wang, Yan; Wang, Chen

doi:10.1109/comst.2019.2934489

Cited by 215 publications

(96 citation statements)

References 119 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Wireless- [59] and IR- [60] based systems rely on anomalous activity detection. They utilize the signatures for a fall that are not immediately obvious to the naked eye [61]. Large amounts of data are generally required to train a model to detect falls.…”

Section: Poses Captured By Caploc For Fall Detectionmentioning

confidence: 99%

CapLoc: Capacitive Sensing Floor for Device-Free Localization and Fall Detection

et al. 2020

View full text Add to dashboard Cite

Passive indoor positioning, also known as Device-Free Localization (DFL), has applications such as occupancy sensing, human-computer interaction, fall detection, and many other location-based services in smart buildings. Vision-, infrared-, wireless-based DFL solutions have been widely explored in recent years. They are characterized by respective strengths and weaknesses in terms of the desired accuracy, feasibility in various real-world scenarios, etc. Passive positioning by tracking the footsteps on the floor has been put forward as one of the promising options. This article introduces CapLoc, a floor-based DFL solution that can localize a subject in real-time using capacitive sensing. Experimental results with three individuals walking 39 paths on the CapLoc show that it can detect and localize a single target's footsteps accurately with a median localization error of 0.026 m. The potential for fall detection is also shown with the outlines of various poses of the subject lying upon the floor.

show abstract

Section: Poses Captured By Caploc For Fall Detectionmentioning

confidence: 99%

CapLoc: Capacitive Sensing Floor for Device-Free Localization and Fall Detection

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Non-contact methods for HAR have also been studied recently [ 32 , 33 , 34 , 35 , 36 , 37 ]. These approaches use ambient Wi-Fi signals or radars to track user activities.…”

Section: Related Researchmentioning

confidence: 99%

“…These approaches use ambient Wi-Fi signals or radars to track user activities. The techniques based on Wi-Fi signals use the channel state information from Wi-Fi signals to infer human activities [ 34 , 35 , 37 ]. In particular, Taylor et al [ 37 ] use radio signals from a USRP radio system to identify standing up and sitting down.…”

Section: Related Researchmentioning

confidence: 99%

w-HAR: An Activity Recognition Dataset and Framework Using Low-Power Wearable Devices

Bhat

Tran

Shill³

et al. 2020

Sensors

View full text Add to dashboard Cite

Human activity recognition (HAR) is growing in popularity due to its wide-ranging applications in patient rehabilitation and movement disorders. HAR approaches typically start with collecting sensor data for the activities under consideration and then develop algorithms using the dataset. As such, the success of algorithms for HAR depends on the availability and quality of datasets. Most of the existing work on HAR uses data from inertial sensors on wearable devices or smartphones to design HAR algorithms. However, inertial sensors exhibit high noise that makes it difficult to segment the data and classify the activities. Furthermore, existing approaches typically do not make their data available publicly, which makes it difficult or impossible to obtain comparisons of HAR approaches. To address these issues, we present wearable HAR (w-HAR) which contains labeled data of seven activities from 22 users. Our dataset’s unique aspect is the integration of data from inertial and wearable stretch sensors, thus providing two modalities of activity information. The wearable stretch sensor data allows us to create variable-length segment data and ensure that each segment contains a single activity. We also provide a HAR framework to use w-HAR to classify the activities. To this end, we first perform a design space exploration to choose a neural network architecture for activity classification. Then, we use two online learning algorithms to adapt the classifier to users whose data are not included at design time. Experiments on the w-HAR dataset show that our framework achieves 95% accuracy while the online learning algorithms improve the accuracy by as much as 40%.

show abstract

“…Human behaviors have also been employed as credentials. For example, prior works have shown it is possible to identify individuals based on hand gestures [21,22], voice commands [5,41], as well as finger inputs made on touchscreens [29,31,32], solid surfaces [23], and wearable devices [3]. These are less personally identifiable, and thus pose a smaller risk to user privacy, but can be challenging to associate with a given identity due to natural inconsistencies users exhibit when asked to reproduce these characteristics Proposals have been made to measure credentials in a passive, low-effort manner, which we consider closely related to EchoLock.…”

Section: Related Workmentioning

confidence: 99%

EchoLock: Towards Low-effort Mobile User Identification Leveraging Structure-borne Echos

Yang

Wang

Chen

et al. 2020

Proceedings of the 15th ACM Asia Conference on Computer and Communications Security

Self Cite

View full text Add to dashboard Cite

Many existing identification approaches require active user input, specialized sensing hardware, or personally identifiable information such as fingerprints or face scans. In this paper, we propose EchoLock, a low-effort identification scheme that validates the user by sensing hand geometry via commodity microphones and speakers. EchoLock can serve as a complementary verification method for high-end devices or as a stand-alone user identification scheme for lower-end devices without using privacy-sensitive features. In addition to security applications, our system can also personalize user interactions with smart devices, such as automatically adapting settings or preferences when different people are holding smart remotes. To this end, we study the impact of hands on structureborne sound propagation in mobile devices and develop a user identification scheme that can measure, quantify, and exploit distinct sound reflections in order to differentiate distinct identities. Particularly, we propose a non-intrusive hand sensing technique to derive unique acoustic features in both time and frequency domain, which can effectively capture the physiological and behavioral traits of a user's hand (e.g., hand contours, finger sizes, holding strengths, and holding styles). Furthermore, learning-based algorithms are developed to robustly identify the user under various environments and conditions. We conduct extensive experiments with 20 participants, gathering 80,000 hand geometry samples using different hardware setups across 160 key use case scenarios. Our results show that EchoLock is capable of identifying users with over 94% accuracy, without requiring any active user input. CCS CONCEPTS • Security and privacy → Authentication.

show abstract

Wireless Sensing for Human Activity: A Survey

Cited by 215 publications

References 119 publications

CapLoc: Capacitive Sensing Floor for Device-Free Localization and Fall Detection

CapLoc: Capacitive Sensing Floor for Device-Free Localization and Fall Detection

w-HAR: An Activity Recognition Dataset and Framework Using Low-Power Wearable Devices

EchoLock: Towards Low-effort Mobile User Identification Leveraging Structure-borne Echos

Contact Info

Product

Resources

About