Multisensing Architecture for the Balance Losses During Gait via Physiologic Signals Recognition

In this paper, we propose an innovative multisensor architecture operating in the field of pre-impact fall detection (PIFD). The proposed architecture jointly analyzes cortical and muscular involvement when unexpected slippages occur during steady walking. The electrophysiological signals (EEG and EMG) are acquired through wearable and wireless acquisition devices interfaced with a central control unit. The control unit consists of a hybrid architecture that exploits both an STM32L4 microcontroller and a DSP-oriented Simulink model. The EMG computation block translates EMGs into binary signals, which are used to trigger the cortical analyses, to extract a score on the revealed muscular pattern and to distinguish “standard” muscular behaviors from anomalous ones (perturbation). A Simulink model evaluates the cortical responsiveness in five bands of interest and implements a logical-based network to detect near-falls or potential ones. The proposed architecture goal is to obtain a detection time conservatively below 550 ms, which represents a strict limit for the successful application of postural recovery strategies. The system has been tested on 6 healthy subjects and demonstrated to react in 370.62 ± 60.85 ms, while keeping competitive accuracy (96.21%).