相关噪声下的机动目标跟踪SMFEKF-IMM算法

SMFEKF-IMM algorithm for maneuvering target tracking with correlated noise

一种基于Kalman和扩展Kalman滤波器的相互作用多模型(IMM)方法可以减小模型的不确定性,但无法消除由于噪声相关引起的状态偏差的弱点。为了提高目标状态估计的精度,把IMM和一种带多重渐消因子的扩展Kalman滤波器(SMFEKF)相结合,提出了一种具有相关噪声的混合随机模型的机动目标跟踪方法。这种方法引入了一个多重渐消因子,当输出残差发生变化时,动态调节增益和系统噪声水平,使输出残差近似正交,从而抑制了相关噪声的影响,适应目标的状态变化。理论分析和仿真实验表明了这种算法的有效性和可行性。

The interacting multiple m odel (IMM) estimator based on the Kalman and extended Kalman filters can reduce the model uncertainty, but can not eliminate the state bias. The accuracy of tar get state estimates was improved with a new tracking algorithm for maneuvering t argets with correlated noise which combines IMM with the suboptimal multiple fad ing extended Kalman filter (SMFEKF). This algorithm makes the residual error app roximately orthogonal by introducing a suboptimal fading factor matrix. This app roach reduces the effect of the correlated noise so that the maneuvering target can be fit by dynamically adjusting the gain and noise levels based on the chang ing of the residual error. Simulation results demonstrate the effectiveness of the proposed approach.