基于粒子滤波的RSSI测距优化的牛顿定位算法

摘要

针对以基于RSSI测距为基础的无线传感网络定位算法存在定位精度低的问题,提出了基于粒子滤波的RSSI测距优化的牛顿定位算法P F-RSSI-NL.在RSSI测距方面,采用粒子滤波对RSSI值进行预处理,降低测距误差;在定位计算方面,运用牛顿法估计未知节点的位置.先用最小二乘法估计牛顿迭代算法的初始值,再用牛顿法对未知节点估计值进行迭代修正.仿真结果表明,与传统的基于统计均值RSSI测距相比,基于粒子滤波的RSSI优化的测距误差降低0.6 m.与同类的定位算法相比,归一化平均定位误差下降36%.%Newton Localization algorithm based on Particle filter RSSI ( PF-RSSI-NL) was proposed to solve the common problems in RSSI-ranging localization algorithm in wireless sensor network, namely low localization accuracy. In terms of RSSI ranging, RSSI ( Received Signal Strength Index) value was preprocessed by particle filter, and thus reducing ranging error. In terms of localization, the position of unknown node was estimated by Newton algorithm. The initial value of Newton algorithm was computed by least-squares, and iterative refinement of unknown node coordinate was done by Newton algorithm. Simulation re-sults show that the ranging error based on RSSI particle filter reduces 0.6 m compared with ranging error based on RSSI statistical mean. Compared with similar localization algorithm, the normalization localization error of PF-RSSI-NL reduce 36%.