基于多传感器的行人航向推算算法研究

Multiple sensors based algorithm for pedestrian heading estimation

摘要:

针对室内行人航位推算(PDR)系统中单一地以陀螺仪作为航向估计会出现误差累积和航向角偏移的问题，提出一种由方向传感器和陀螺仪组合的航向角校正算法。首先，利用卡尔曼滤波（Kalman Filter KF）分别消除方向传感器的信号干扰和陀螺仪的动态漂移误差；然后，通过陀螺仪和方向传感器测量的航向角差值是否超过阈值来判断是否存在硬磁场干扰；最后，根据硬磁场干扰情况对航向角估计值进行相应的角度补偿得到新的航向角估计值。实验表明，该算法的航向推算性能优于启发式漂移消除算法和增强式启发式漂移消除算法。

Abstract:

Because the heading estimation in the indoor pedestrian dead reckoning(PDR) system has the problems of cumulative and course angle drifting when gyroscope is used, an course angel calibrating algorithm combined direction sensor with gyroscope is proposed.Firstly, Kalman filter is adopted to eliminate signal interference of direction sensor and dynamic drifting error of gyroscope.Then, the difference value of course angel between direction sensor and gyroscope is calculated to judge whether there existing hard magnetic interference.Finally, the new course angel is got by angel compensation according to the hard magnetic interference degree.The experimental results show that the proposed algorithm has got better performance than Heuristic Drift Elimination algorithm and Advanced Heuristic Drift Elimination algorithm in term of pedestrian dead reckoning