改进预测协方差门限法的采样周期自适应选择算法

Adaptive Sampling Period Selection Algorithm Based on an Improved Prediction Covariance Threshold Method

传感器采样周期是影响目标跟踪的一个重要参数。现有自适应采样周期策略中,一些算法运算量比较大,计算效率低,不具有一般性。为此提出了一种改进的预测协方差门限法。该算法改进传统采样周期的全遍历寻优策略。最后与几种自适应采样周期算法与固定采样周期算法通过交互式多模型(IMM)滤波算法进行仿真比较。仿真结果表明该算法在目标跟踪过程中能满足跟踪需求,具有较少的计算量,较高的运行效率,比固定采样周期算法更能节约资源。

Sensor sampling period is an important parameter in affecting target tracking. Some sampling period selection algorithms have a large amount of computation time and are lack of efficiency and universality. An improved prediction covariance threshold method is proposed in this paper. Instead of using full enumeration optimization, the algorithm gives an improved optimization strategy. Finally it is compared with several adaptive sampling period algorithms and the fixed sampling period algorithm with interactive multiple model （IMM） filter algorithm. Comparison results show that the improved algorithm can satisfy the tracking requirement in the target tracking process with less computational time and higher computational efficiency than that of the adaptive sampling period algorithms. It also can conserve more sensor resources than that of the fixed sampling period algorithm.