基于学生t分布的变分贝叶斯UKF算法在无人船对准中的应用

Application of Student’s T Distribution Based Variational Bayesian UKF Algorithm In Unmanned Ship Alignment

针对无人船在恶劣海况下对准时会受到复杂未知干扰,从而导致量测噪声时变,且因存在野值而表现出厚尾特性的问题,提出了一种自适应鲁棒的基于学生t分布的变分贝叶斯UKF(St-VB UKF)算法,将量测噪声建模成学生t分布来描述噪声的厚尾特性,并通过变分贝叶斯方法实时自适应地估计量测噪声的统计特性。仿真结果表明,提出的算法在受到短时和瞬时的强干扰影响时,对准精度和速度均优于传统的UKF算法和Sage-Husa自适应UKF算法;且当量测中存在部分野值而呈现厚尾特性时,提出的算法可以有效抑制野值的影响,对准精度保持不变,提升了算法的鲁棒性。

An adaptive robust Student’s t distribution based variational Bayesian UKF(St-VB UKF)algorithm is proposed to deal with the heavy-tailed measurement noise and its time-varying statistical characteristics when unmanned ships alignment suffers from complicated unknown interference under rough sea conditions.The measurement noise is modeled as a student’s t distribution to describe the heavy-tail characteristics and the statistical characteristics of the measurement noise is adaptively estimated by using the variational Bayesian method in real time.The simulation results show that proposed algorithm has better align speed and accuracy than the conventional UKF algorithm and the Sage-Husa adaptive UKF algorithm usder the condition of short-term and instantaneous strong interference.Furthermore,under outlier corrupted heavy-tailed measurement noise condition,the proposed algorithm can effectively suppress the influence of outliers,the alignment accuracy remains unchanged,and the robustness of the algorithm is improved.