基于期望系统噪声模型的自适应机动目标跟踪

Adaptive Maneuvering Target Tracking Based on Expected System Noise Model

对于机动目标跟踪问题,由于目标机动能力的增强,使建立的目标模型与目标的实际运动失配。为解决这个问题,需建立大量模型来逼近真实模式。但这使计算量增大,而且性能不一定能提高。本文提出基于期望系统噪声模型的自适应交互式多模型(IMM)算法。该算法自适应调整部分系统噪声模型,使之接近符合实际的系统噪声模型。对目标机动运动的Monte-Carlo仿真结果表明,本算法对机动目标的跟踪精度比标准IMM算法有较大改进,且计算量适中。

For maneuvering target tracking, the increase in the maneuvering capability of targets makes the established models mismatch the true modes of maneuvering targets. Lots of models need to be established to approach the true mode in order to solve it. But this brings on dramatic increase in calculation, and cannot always improve the performance of the system. An interacting multiple model （IMM） adaptive filtering algorithm based on expected system noise model was presented. In this approach, a part of the system noise models are intended to adaptively match the unknown true mode. Comparison between the new algorithm and the standard IMM algorithm is evaluated though Monte-Carlo simulation. The results show that this algorithm improves the tracking accuracy for maneuvering targets and involves only moderate amount of the computation.