基于LBES-SVM的室内UWB定位研究

On the Indoor UWB Positioning Based on LBES-SVM

在超宽带(UWB)室内定位系统中,非视距(NLOS)误差是影响定位结果准确性的主要因素。针对室内非视距环境下UWB信号受干扰较为严重,无法精确定位的问题,提出一种优化支持向量机(SVM)模型修正最小二乘法(LS)的定位算法。该算法首先利用LS得到初次定位坐标,再利用改进秃鹰算法(LBES)优化SVM模型,最后利用优化后的SVM模型对初次定位坐标进行误差修正,从而减小初次定位的坐标误差,提高定位精度。实验与测试结果表明,利用LBES-SVM模型修正LS的定位坐标后,三轴的误差均有减小,效果最佳的Z轴上误差从14.3cm减少至4.5cm,欧式空间平均距离由16.0cm减小至6.7cm,说明修正后的定位坐标更接近于真实坐标,即提高了原有的定位精度。

In ultra-wideband(UWB)indoor positioning system,non-line-of-sight(NLOS)error is the main factor affecting the accuracy of positioning results.Aiming at the problem that UWB signal in indoor non-line-of-sight environment is seriously interfered with and cannot be accurately located,an optimized support vector machine(SVM)model modified least square(LS)localization algorithm is proposed.Firstly,LS is used to obtain the initial positioning coordinates,and then the improved Condor algorithm(LBES)is used to optimize the SVM model.Finally,the optimized SVM model is used to correct the initial positioning coordinates,so as to reduce the coordinate errors of the initial positioning and improve the positioning accuracy.The experimental and test results show that the errors of the three axes are reduced after the LBES-SVM model is used to modify the positioning coordinates of LS.The error on the Z axis with the best effect is reduced from 14.3cm to 4.5cm,and the average distance in the Euclidean space is reduced from 16.0cm to 6.7cm,indicating that the corrected positioning coordinates are closer to the real coordinates,the original positioning accuracy is improved.