Smartphone-based speed acquisition and differential evolution modelling for fitness running

摘要

There are two shortcomings in Brzostowski's method of modelling for fitness running. One is the low precision of speed acquisition by smartphones. The other is the limited search ability of modelling algorithms. As a result, it is difficult to obtain a highly precise model of fitness running. Aiming at the above two problems, a method for multiple sensors of smartphone and median filter for speed data acquisition in fitness running (MM4SA) is proposed. MM4SA can remove the impulse noise of three-axis acceleration signal, which is generated by the intermittent gesture changes of smartphones, and filter out the gravitational acceleration from the three-axis acceleration signal with the help of the orientation sensor. Secondly, the differential evolution modelling algorithm (DEMA) is applied to enlarge search space and find out a better model for fitness running. The experimental results show that the proposed method can obtain more accurate speed and better model for fitness running than Brzostowski's method.