WiN: Non-invasive Abnormal Activity Detection Leveraging Fine-Grained WiFi Signals

Zhu, Daqi; Pang, Na; Li, Gang; Rong, Wenjing; Fan, Zheming

doi:10.1109/trustcom.2016.0134

Cited by 5 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CARM system implemented in commercial Wi-Fi device [ 26 ] is based on CSI-speed model and CSI-activity model. For the detection of abnormal activities NotiFi was proposed in [ 27 ]. It automatically learns several movement categories for abnormal detection by creating multiple hierarchical Dirichlet processes.…”

Section: Related Workmentioning

confidence: 99%

UWB Radio-Based Motion Detection System for Assisted Living

Bregar

Hrovat

Mohorčič

2021

Sensors

View full text Add to dashboard Cite

Because of the ageing population, the demand for assisted living solutions that can help prolonging independent living of elderly at their homes with reduced interaction with caregivers is rapidly increasing. One of the most important indicators of the users’ well-being is their motion and mobility inside their homes, used either on its own or as contextual information for other more complex activities such as cooking, housekeeping or maintaining personal hygiene. In monitoring users’ mobility, radio frequency (RF) communication technologies have an advantage over optical motion detectors because of their penetrability through the obstacles, thus covering greater areas with fewer devices. However, as we show in this paper, RF links exhibit large variations depending on channel conditions in operating environment as well as the level and intensity of motion, limiting the performance of the fixed motion detection threshold determined on offline or batch measurement data. Thus, we propose a new algorithm with an online adaptive motion detection threshold that makes use of channel impulse response (CIR) information of the IEEE 802.15.4 ultra-wideband (UWB) radio, which comprises an easy-to-install robust motion detection system. The online adaptive motion detection (OAMD) algorithm uses a sliding window on the last 100 derivatives of power delay profile (PDP) differences and their statistics to set the threshold for motion detection. It takes into account the empirically confirmed observation that motion manifests itself in long-tail samples or outliers of PDP differences’ probability density function. The algorithm determines the online threshold by calculating the statistics on the derivatives of the 100 most recent PDP differences in a sliding window and scales them up in the suitable range for PDP differences with multiplication factors defined by a data-driven process using measurements from representative operating environments. The OAMD algorithm demonstrates great adaptability to various environmental conditions and exceptional performance compared to the offline batch algorithm. A motion detection solution incorporating the proposed highly reliable algorithm can complement and enhance various assisted living technologies to assess user’s well-being over long periods of time, detect critical events and issue warnings or alarms to caregivers.

show abstract

Section: Related Workmentioning

confidence: 99%

UWB Radio-Based Motion Detection System for Assisted Living

Bregar

Hrovat

Mohorčič

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…However, existing CSI-based schemes highly depend on the observation of the repetitions with reproducible features, while violent behaviours can be random and irregular. NotiFi [42] proposed a non-parameter training scheme for abnormal activity detection by using Dirichlet process. Yet it requires the user's continuous walking and cannot handle the presence of multiple users.…”

Section: Related Workmentioning

confidence: 99%

Enabling Noninvasive Physical Assault Monitoring in Smart School with Commercial Wi-Fi Devices

Zhou

Xing

et al. 2019

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Monitoring physical assault is critical for the prevention of juvenile delinquency and promotion of school harmony. A large portion of assault events, particularly school violence among teenagers, usually happen at indoor secluded places. Pioneering approaches employ always-on-body sensors or cameras in the limited surveillance area, which are privacy-invasive and cannot provide ubiquitous assault monitoring. In this paper, we present Wi-Dog, a noninvasive physical assault monitoring scheme that enables privacy-preserving monitoring in ubiquitous circumstances. Wi-Dog is based on widely deployed commodity Wi-Fi infrastructures. The key intuition is that Wi-Fi signals are easily distorted by human motions, and motion-induced signals could convey informative characteristics, such as intensity, regularity, and continuity. Specifically, to explicitly reveal the substantive properties of physical assault, we innovatively propose a set of signal processing methods for informative components extraction by selecting sensitive antenna pairs and subcarriers. Then a novel signal-complexity-based segmentation method is developed as a location-independent indicator to monitor targeted movement transitions. Finally, holistic analysis is employed based on domain knowledge, and we distinguish the violence process from both local and global perspective using time-frequency features. We implement Wi-Dog on commercial Wi-Fi devices and evaluate it in real indoor environments. Experimental results demonstrate the effectiveness of Wi-Dog which consistently outperforms the advanced abnormal detection methods with a higher true detection rate of 94% and a lower false alarm rate of 8%.

show abstract