An amendment navigation algorithm is studied on shoe-mounted PNS(personal navigation system), which is composed of MEMS inertial sensors with low precision. Based on traditional SINS navigation algorithm, a ZUPT(zero velocity update) method is developed to improve navigation accuracy. By analyzing the acceleration's statistical characteristic on walker's foot, a so-called SFM/SV(specific force magnitude and slip variance) detection algorithm is designed for detecting stance phase between steps. Then through a Kalman filter which is designed for zero velocity update(ZUPT), the navigation errors including attitude error, velocity error and position error can be estimated. Therefore, the navigation precision can be improved by feeding these estimated errors back to original navigation system. In the end, two physical experiments are conducted using an integrated MEMS IMU, and the results indicate that the above algorithm is feasible for PNS. In addition, future research directions are pointed out.%研究了低精度鞋式个人惯性导航系统的导航修正算法.该系统由低精度MEMS惯性IMU单元组成,固联在步行者的鞋上.导航算法在传统捷联惯性导航算法基础上,引入了零速修正技术,根据人行走时脚部运动的加速度统计特性,设计了一种比力模值+滑动方差检测算法,用以检测行走过程中的静止时间段.然后通过设计的改良卡尔曼滤波器在静止时间段内滤波估计导航姿态、速度和位置的计算误差,通过反馈校正可以提高原系统的导航精度.最后通过两组MEMS实物实验验证了导航修正算法的有效性和可行性,并指出了进一步的研究方向.
展开▼
