基于自适应的增广状态-交互式多模型的机动目标跟踪算法

现有的增广状态-交互式多模型算法存在着依赖于量测噪声协方差矩阵这一先验信息的问题。当先验信息未知或不准确时，算法的跟踪性能将会下降。针对上述问题，该文提出一种自适应的变分贝叶斯增广状态-交互式多模型算法VB-AS-IMM。首先，针对增广状态的跳变马尔科夫系统，该文给出了联合估计增广状态和量测噪声协方差矩阵的变分贝叶斯推断概率模型。其次，通过理论推导证明了该概率模型是非共轭的。最后，通过引入一种“信息反馈+后处理”方案，提出联合后验密度的次优求解方法。所提算法能够在线估计未知的量测噪声协方差矩阵，具有更强的鲁棒性和适应性。仿真结果验证了算法的有效性。

Maneuvering Target Tracking Algorithm Based on the Adaptive Augmented State Interracting Multiple Model

The existing Augmented State-Interracting Multiple Model (AS-IMM) algorithm suffers from the problem that it relies on the prior information of the covariance matrix of the measurement noise. When the prior information is unavailable or inaccurate, the tracking performance of AS-IMM will be degraded. In order to overcome this problem, a novel adaptive Bayesian Variational Augmented State-Interracting Multiple Model (VB-AS-IMM) algorithm is proposed. Firstly, the variational Bayesian inference probabilistic model of the augmented state and the covariance matrix of the measurement noise for the jump Markovarian system is presented. Secondly, the probabilistic model is proven to be non-conjugated. Finally, by introducing a novel post processing method, the suboptimal solution to calculate the joint posterior distribution is proposed. The proposed algorithm can estimate the unknown covariance matrix of the measurement noise online, thus it is more robust and has higher adaptability. Simulation result verifies good performance of the proposed algorithm.