FallDeFi

Abstract

Falling or tripping among elderly people living on their own is recognized as a major public health worry that can even lead to death. Fall detection systems that alert caregivers, family members or neighbours can potentially save lives. In the past decade, an extensive amount of research has been carried out to develop fall detection systems based on a range of different detection approaches, i.e, wearable and non-wearable sensing and detection technologies. In this paper, we consider an emerging non-wearable fall detection approach based on WiFi Channel State Information (CSI). Previous CSI based fall detection solutions have considered only time domain approaches. Here, we take an altogether different direction, time-frequency analysis as used in radar fall detection. We use the conventional Short-Time Fourier Transform (STFT) to extract time-frequency features and a sequential forward selection algorithm to single out features that are resilient to environment changes while maintaining a higher fall detection rate. When our system is pre-trained, it has a 93% accuracy and compared to RTFall and CARM, this is a 12% and 15% improvement respectively. When the environment changes, our system still has an average accuracy close to 80% which is more than a 20% to 30% and 5% to 15% improvement respectively.

Human Activity Sensing with Wireless Signals: A Survey

“…HeadScan [43] extracts dominant frequency as the features. FallDeFi [92] proposes a frequencydomain feature called fractal dimension, which is robust to environment changes. Zeng et al [85] calculate the frequency domain energy as a feature.…”

“…WiSee [55] and FEMO [10] use the Doppler shifts in the classification of template matching. Many studies [43,70,72,73,80,90,92] extract spectral entropy from frequency streams as the classification features.…”

“…Template matching recognition is often training-free [10,19,20,39,55,74,91,93,94]. Training-once classification requires the valid features robust to the variations in the surrounding environment [12,13,32,37,40,42,44,45,[47][48][49]56,59,60,70,72,73,75,76,78,79,83,84,89,90,92,127,130]. Deep learning automatically extracts features, which often requires only one time of training [16,17,21,48,59,67,77,88,95,96,108].…”

“…SVM is widely used in various human activity detection and classification systems [12,13,32,37,40,42,44,45,[47][48][49]56,59,60,70,72,73,75,76,78,79,83,84,89,90,92,127,130]. SVM classifier are used to classify 10 digital finger gestures [78], eight human movement directions [59], fall detection [37,92], walking direction detection [70], fitness activity classification [56,79], human daily activities [13,40,42,72,75,76,84,90], and human gestures [12,32,73,75,83,89]. ...…”

“…The human motion detection can be further divided into fall detection [37,50,73,92], walking step detection [87], intrusion detection [48,49,54,98], and human activity detection [38,72,90,134]. For fall detection, it distinguishes between fall and non-fall activities.…”