Exploring Accelerometer Sensors for Optimized Smart Shoe-Based Fall Detection

The need of immediate fall detection, particularly within the geriatric population, is emphasized by its pivotal role in curtailing consequences post-incident. In this paper we introduce an algorithm that harnesses the subject’s acceleration, angular velocity, and orientation (pitch and roll) to ascertain fall incidents. The evaluation is based on systematic extraction and comparison of features delineated over varied temporal intervals. The presented algorithm uses data extracted from a singular sensor, amalgamating a triaxial accelerometer and a gyroscope, embedded in footwear possessing integrated computational modules. The resilience of the algorithm and efficacy are empirically validated through rigorous assessments, involving five healthy participants simulating 160 fall events and 650 activities indicative of routine locomotion.