Multi-Sensors Walking Speed Estimation Robust against the Dynamics of Handheld Devices

In this paper, we describe and evaluate the methods for walking speed estimation method. Influence factors of the different usage poses are taken into consideration. Multi-sensors data were used to detect the usage poses and estimate walking speed. Field tests were carried out to verify the proposed mobile device usage pose detection and walking speed estimation algorithms. Test result show that the proposed algorithms achieve a 97.19% precision and 96.96% of recall rate, and the performance of walking speed estimation algorithm was effective and achieved good accuracy and robustness, the mean of absolute error is 0.0804 m/s.