A pedestrian movement direction recognition method based on inertial sensors

Abstract

Most of the phone has been configured IMU currently, includes an acceleration sensor and an electronic compass. In this paper, using the pedestrian dead reckoning algorithm based accelerometer and electronic compass composition IMU to assist indoor positioning. Detection walking step length and state. Currently detect direction of motion requires the user wear or hand in IMU with a fixed posture. To address this issue, using the terminal gesture recognition algorithm in the third quarter. When pedestrian walking on the same direction, the angle walking forward and walking backward was the same. To address this issue, using pedestrian movement direction detection method based on the differential cross-correlation of the acceleration in the fourth quarter. Based on the above, the article conducted experiments in Section V and the result shows that the proposed method can effectively detect pedestrian movement forward and backward state, the average accuracy of the detection results is 85.67 %.