基于多传感器零速修正的行人导航系统研究

Research on pedestrian navigation system based on multi-sensor zero speed correction

针对行人导航定位系统中惯性导航解算误差累积问题,提出基于多传感器零速修正的行人三维导航定位方法。根据行人足部运动特征,采用支持向量机决策方法对行人运动和静止阶段进行检测分类。在静止阶段,根据多传感器测量数据对导航解算速度、角速度、方向和高度进行误差修正,采用卡尔曼滤波递推方法进行数据融合和滤波估计,实现对方向、位置和速度的跟踪。行人行走实验表明,设计的行人导航定位系统能够有效地跟踪行人行走轨迹,水平方向各楼层平均误差为1. 82%,高度方向平均误差为2. 53%。

Aiming at the accumulation problem of inertial navigation solver error in pedestrian navigation and positioning system,a pedestrian three-dimensional navigation and positioning method based on multi-sensor zero speed correction is proposed. According to the movement characteristics of pedestrian foot,the support vector machine decision-making method is used to detect and classify pedestrian movement and stationary stage. In the stationary stage,according to the multi-sensor data,the navigation solution speed,angular velocity,direction and height can be corrected,and the Kalman filter recursive method is used for data fusion and filter estimation to track the direction,position and speed. Pedestrian walking experiments show that the pedestrian navigation system can effectively track the walking trajectory of pedestrians. The average error of each floor in the horizontal direction is 1. 82%,and the average error in the height direction is 2. 53%.