基于转换测量卡尔曼滤波算法的目标跟踪研究

在去偏转换测量中,先以目标的真实位置为条件求出真实的偏差和协方差阵,然后以传感器的测量为条件对真实偏差和协方差求取期望,这种方法求得的转换测量的偏差和协方差阵会引入附加的误差,因而不够精确。针对这种情况,直接以传感器的测量为条件求取转换测量的偏差和协方差阵,这种方法不仅精确而且简单。仿真结果表明算法的有效性。

Converted Measurements Kalman Filter Algorithm for Target Tracking

In the method of debiased converted measurements, the error statistics were computed conditioned on true target position at first, and then averaged conditioned on observations. The error statistics computed by this method were, in general, not precise enough. To avoid this problem, the error statistics were averaged directly conditioned on the observations. This method was simpler, but also more accurate. Results of simulation prove the effectiveness of the new algorithm.