本文对多目标情况下的多传感器系统误差估计问题进行了研究,提出了反馈式多目标多传感器系统误差融合估计算法。算法首先对EX算法进行了修正,无需计算各目标状态估计及其增益伪逆,直接利用各传感器量测数据来构建伪量测,通过滤波获得局部传感器组合系统误差估计;其次,算法通过构建状态空间转移矩阵实现了全局估计到局部组合估计关系描述,并从多目标多传感器两个层面对系统误差估计进行更新,即对多目标信息进行递归融合以有效利用空间分布的多目标信息,对多传感器组合估计信息进行反馈融合获得全局估计。蒙特卡洛仿真结果说明该算法能够进行多传感器系统误差的快速精确融合估计,相比EX算法在实时性与估计精度方面均具有较大的优越性。
The problem of multi-sensor systematic error estimation under the condition of multi-targets is researched in this paper,and a feedback based multi-target multi-sensor systematic error fusion estimation algorithm is presented.Firstly,the EX algorithm is modified by directly using the sensor measurements to form the pseudo-measurement without calculating the target state estimation and pseudo-inverse of gain.Thus,the systematic error of the combination of sensors can be obtained through filtering.Secondly,by describing the relationship between systematic estimation and local combination estimation through constructing a state space transfer matrix,the algorithm is used to update the systematic error estimation at two different levels,namely,space distributed multi-target information to be used to achieve the estimation fusion recursively and the fusion with feedback information of multi-sensor combined estimation to be used to obtain global estimation.The Monte-Carlo simulation result shows that the algorithm can estimate the systematic error of sensors quickly and exactly,and have a advantage over EX algorithm in the requirements for real time and estimation accuracy.