基于非视距鉴别的超宽带室内定位算法

UWB indoor positioning algorithm based on NLOS identification

针对室内定位精度受非视距(NLOS)环境影响严重的问题,结合超宽带(UWB)信号特征,提出了一种基于非视距鉴别的信号到达时间(TOA)定位算法。首先,对视距和非视距环境下的UWB信号特征采样,构建随机森林判别模型鉴别出NLOS基站,将NLOS下的TOA测量值和计算值的平均值作为该基站的TOA值;然后对所有基站的TOA值赋予不同的权重,采用加权最小二乘法估计出移动目标的位置;最后用卡尔曼滤波算法优化定位结果,进一步提高定位精度。仿真实验表明,该算法的非视距鉴别率达98. 6%;在不同的非视距误差下,定位精度相比最小二乘法提高了52%~55%,具有稳定的定位性能。

Aiming at the problem that Non-Line Of Sight （NLOS） environment has serious impact on indoor positioning, combining with Uhra-Wide Band （UWB） signal characteristics, a Time Of Arrival （TOA） algorithm based on NLOS identification was proposed. Firstly, UWB signal features in Line Of Sight （LOS） and NLOS environments were sampled to identify NLOS by random forest algorithm, then the average value of TOA measurement and the calculated value under NLOS were used as the TOA value of the base station. Secondly, the TOA values of all the base stationswere set different weights, and the position of moving target was estimated by the weighted least square method. Finally, the Kalman fiher algorithm was used to optimize the positioning resuhs. The positioning simulation experiments show that, the NLOS identification rate of the proposed algorithm is 98.6% ; compared with the least square algorithm, the positioning accuracy is improved by 52% - 55% under different NLOS errors and the proposed algorithm has a stable positioning performance.