A Framework for Activity Monitoring and Fall Detection Based on the Characteristics of Indoor Channels

Abstract

A robust fall detection system is essential to support the independent living of elderlies. In this context, we develop a machine learning framework for fall detection and daily living activity recognition. Using acceleration data from public databases, we test the performance of two algorithms to classify seven different activities including falls and activities of daily living. We extract new features from the acceleration signal and demonstrate their effect on improving the accuracy and the precision of the classifier. Our analysis reveals that the quadratic support vector machine classifier achieves an overall accuracy of 93.2% and outperforms the artificial neural network algorithm.

Estimation of the Velocity of a Walking Person in Non-Stationary Indoor Environments from the Received RF Signal

“…Here, the scenario consists of a room with length A, width B, and height H centered at the origin O and equipped with a distributed 3×3 MIMO system consisting of three transmit antennas (A T 1 , A T 2 , and A T 3 ) and three receive antennas (A R 1 , A R 2 , and A R 3 ). The propagation phenomenon taking place in the indoor area is modelled using the three-dimensional indoor multipath propagation scenario introduced in [26]. A simplified geometrical description of the propagation scenario is illustrated in Fig.…”

“…The trajectory of the moving point scatterer S M is described by its TV speed v(t) = | v(t)|, the TV VAOM β v (t), and the TV HAOM α v (t). The TV positions x(t), y(t), and z(t) along the x-, yand z-axis of the moving person S M can be deduced from the velocity v(t) according to [26,Eqs. (5)− 7].…”

“…Under the assumption of perfect channel state information (CSI) at the receiver side, the complex channel gain µ ij (t) between the jth transmit antenna and ith receive antenna (i, j = 1, 2, 3) can be expressed as [26] µ…”

“…Mij mij =1 c mij cos(θ mij )). According to [26], by approximating the Doppler frequencies f ij (t) by N piecewise linear functions, the spectrogram S ij (f, t) of µ ij (t) can be expressed as…”

“…Here, the room is assumed to be equipped with a distributed 3×3 multiple-input multiple-output (MIMO) system. In the present work, the TV speed, TV vertical angle-of-motion (VAOM), and TV horizontal angle-of-motion (HAOM) are optimized by fitting the spectrograms of the model in [26] to the spectrograms of the received signals. Although the primarily objective of the proposed algorithm is to estimate the TV velocity, this task requires the determination of all channel parameters, i.e., gains, phases, TV azimuth (elevation) angles-of-departure (AAODs (EAODs)), and TV azimuth (elevation) angles-of-arrival (AAOAs (EAOAs)).…”