Proof of concept of a shoe based human activity monitor

This paper presents the proof of concept of a low power, low cost, wearable activity monitor. The functionality of the system is based on accurate stride detection from signals generated by two force sensing resistors integrated within a normal shoe. A novel algorithm is proposed that is able to differentiate between walking and non-walking activities with high accuracy. The performance of the proof of concept system was validated in five subjects who underwent five repetitions of three different speed walking tests, and five repetitions of five non-walking artefact generating tests. The system achieved a total sensitivity of 96% with 98% specificity and an overall accuracy of 94%.