基于粒子滤波的多信息融合室内定位算法研究

The research on multiple information fusion indoor positioning algorithm based on particle filter

针对现有室内定位技术精度低、实现复杂度高等问题,提出一种基于粒子滤波的多信息融合室内定位算法。在传统的行人航迹推算(PDR)以及地磁匹配等室内定位算法的基础上,通过粒子滤波动态地纠正行人步长和方向角,从而有效地减小了定位误差。通过PDR算法获得行人的步频、步长、方向等信息,由地磁匹配算法得到行人所在位置对应的地磁值,最后利用粒子滤波对以上信息进行融合处理从而得到粒子的权重,由粒子权重对步长和方向角不断地修正。实验结果表明,该算法可以实时动态地补偿PDR的定位误差,能够获得较高的定位精度。

Aiming at the problem of low accurcy and high implementation complexity for existing indoor location methods, a multiple information fusion indoor positioning algorithm based on particle filter is presented. On the basis of traditional indoor location algorithm of Pedestrian Dead Reckoning （PDR）and geomagnetic matching, we use the particle filter to dynamically correct the pedes- trians＇ step length and attitude angle, so as to effectively reduce the positioning error. Step length and attitude information are both obtained through the PDR algorithm, then geomagnetic matching algorithm is used to obtain geomagnetic corresponding to pedestrians position. The particle filter is used to get the weight of the particle through fusing above information, then the step length and at- titude angle are corrected continuously by particle weight. The experimental results show that the algorithm can real-time dynamicly compensate PDR＇s positioning error, and the system can realize accurate indoor positioning.