An Improved Gait Detection Algorithm Based on Zero-Velocity Detection Method and its Application

Gait detection and recognition have proven to be valuable in various fields. Based on inertial sensors, wearable devices offer a suitable means for extracting gait information. This study focuses on designing a human gait detection algorithm for healthy subjects using inertial sensors. By placing a single sensor at the ankle, the algorithm estimates the body’s trajectory and extracts gait information through various data processing methods. A wearable device is designed to implement the proposed algorithm, which is then tested extensively. Experimental results demonstrate that the proposed algorithm achieves an average relative error, compared to a visual system serving as the gold standard, of 3.26% for travel distance, 0.02% for stride frequency, 3.26% for stride length, 3.26% for pace, 1.41% for stride time, 2.72% for stance time, and 2.37% for relevant stance. Furthermore, the algorithm and device prove to be suitable for different testers and various wearing methods (i.e., left or right ankle). When using two sensors, one on each ankle, additional gait information such as step time, single stance time, and symmetry can be extracted.