基于预测信息的量测转换序贯滤波目标跟踪

Sequential filtering for target tracking based on the prediction position conditioned measurement conversion

针对多普勒雷达目标跟踪问题,提出了基于预测值量测转换的序贯滤波目标跟踪算法。该算法采用基于预测信息的量测转换方式解决量测与目标运动状态的非线性,其中位置量测转换采用乘性去偏、伪量测转换采用加性去偏,量测转换误差的统计特性基于预测值进行推导,并采用序贯滤波方式处理伪量测以实现目标跟踪。同时将该算法扩展到机动目标跟踪情况,综合利用了位置量测与伪量测滤波部分输出的概率组合作为该模型的更新概率,利用模型更新概率对各滤波器的状态估计结果进行加权求和获得最终估计。仿真结果表明该算法与传统的序贯滤波跟踪算法相比,具有更高的跟踪精度,其扩展方法可实现有效的机动目标跟踪。

A sequential filtering algorithm for target tracking in Doppler radar based on the prediction position conditioned conversion is presented. The nonlinear problem is solved by the measurement conversion and the sequential estimation is used to process the pseudo-measurement to achieve target tracking. The measurement conversion method is based on prediction information, which includes the unbiased position measurement conversion and the debiased pseudo-measurement conversion. At the same time, the algorithm is extended to maneuvering target tracking, the update probability is obtained from position measurement estimation and pseudo measurement estimation, and the final estimation is a combination of pseudo-measurement estimation and update probabilities. The simulation results show that compared with the conventional filtering algorithms, the proposed algorithm can realize target tracking effectively with preferable tracking precision and the extended method can achieve effective maneuvering target tracking.