重力辅助和模拟零速修正的手机航向修正方法

Heading correction method for smartphone based on gravity assisting and simulated zero-velocity updating

基于计步的传统航位推算的手机导航方法要求手机保持相对人体固定位置以保证航向的准确性,该要求严重影响了用户体验。针对行人的手机姿态改变和高精度定位的行人导航需求,提出了一种重力辅助和模拟零速修正的航向补偿方法。手机姿态发生改变时候的航向角度补偿可以采用手机重力计输出数据进行辅助判断;通常脚部捆绑式惯性导航定位中采用的航位推算技术无法应用于行人手持的手机,所以不具备零速修正算法的基本条件,为此提出了一种应用于行人手持手机的模拟零速修正算法,通过检测行人步态,采用卡尔曼滤波有效抑制了手机的航向发散。行人的综合行走实验结果表明,基于重力辅助和模拟零速修正的手机航向修正方法,能够自主判断并补偿由于手机使用方式改变造成的航向误差,在行走196m距离的情况下,行走误差仅有1.2%,有效提高了行人定位精度。

The traditional smartphone navigation algorithm based on step counting and pedestrian dead reckoning （PDR） requires the mobile phone to maintain a relatively fixed position to the human body to ensure the heading accuracy. This requirement seriously affects the user experience. In view of pedestrian navigation needs for mobile phone posture changing and high precision positioning, a heading compensation method with gravity assist （GA） and simulated zero-velocity update （S-ZUPT） is proposed to reduce the heading error. The changing of smartphone＇s attitude can be judged by using the data from the gravity sensor of the smart-phone. Although the ZUPT algorithm can be applied in the foot-mounted low-cost inertial sensors to restrain the error divergence, it cannot be used in the handheld smartphone for there is no zero-velocity status. An S-ZUPT algorithm for improving heading accuracy is proposed to improve the heading performance by detecting pedestrian walking step, which uses the Kalman Filtering algorithm to restrain the heading divergence. The walking experiment results show that the GA and S-ZUPT algorithms-based method can autonomously judge and compensate the heading error caused by smartphone attitude. The walking error is only 1.2% with the 196 m distance, which significantly improves the pedestrian positioning accuracy.