基于修订矩阵的最小二乘法多传感器融合障碍物位置估计

摘要

在障碍物位置检测时,由于传感器测量误差等不确定因素的存在,希望通过多传感器融合降低不确定性使得位置估计值更为可靠.提出了一种基于修订矩阵的最小二乘法多传感器融合估计新方法.根据红外传感器得出的先验信息建立修订矩阵,再由最小二乘法的原理得出线性可信度加权方程.分别对加入和不加入修订矩阵进行仿真和"未来之星"移动机器人实验,结果可知最小二乘法大大降低了不确定性,而加入修订矩阵使得估计更加接近真实值.%In obstacle position detection,because of sensor error existence, the multisensor fusion are used to reduce the uncertainty so that the estimation result is reliable. A new method based on correction matrix of the least square method multisensor fusion is proposed. A correction matrix is set up by the prior information of IR ( infrared ray) sensor. The linear credibility weighted equation are obtained, according to the least square method principle. Simulation and experiment in mobile robot“ FutureStar”are done with and without correction matrix, and it shows that the least square method can greatly reduce uncertainty and the estimation with correction matrix is more close to the true value.