摘要
标准的无损卡尔曼滤波算法(UKF)是一种高效的线性化航迹追踪算法,但UKF控制策略中的AUV系统采样时间间隔一般都被设置为常数,这可能会影响导航预测结果的误差精度。文章在简要阐述标准UKF算法原理的基础上,通过两种判断与反馈机制,调整UKF算法每一步的采样时间间隔t,从而实现对系统采样时间间隔的自适应变化。由此提出了基于马氏距离的优化无损卡尔曼滤波算法(MUKF)与基于灰色关联度的优化无损卡尔曼滤波算法(GUKF),以实现对AUV航行轨迹的精确预测与控制。对比两种优化后的UKF算法的仿真实验结果,进一步验证了之前所提出的假设。基于AUV航迹追踪技术的两种优化UKF算法与标准的UKF算法相比,具有更高的航迹预测误差精度和鲁棒性。
The standard unscented Kalman filter(UKF) algorithm is an efficient linear tracking algorithm. However, the AUV system sampling time interval in the UKF control strategy is generally set to a constant, which may affect the error precision of the simulation results. Based on the principle of the standard UKF control algorithm, the paper adjusts the sampling time interval t of the UKF algorithm by two decision and feedback mechanism, so as to realize the adaptive change of the system time interval. According to this, the optimized unscented Kalman filter algorithm based on Mahalanobis distance(MUKF) and the optimized unscented Kalman filter algorithm based on gray relational degree(GUKF) are proposed in order to achieve accurate trajectory prediction and control for AUV navigation. After comparing the simulation results of the two optimized UKF algorithms, the assumptions made before are further verified. Compared with the standard UKF algorithm, two optimized UKF algorithms based on AUV track technology have higher trajectory prediction error precision and robustness.
出处
《船舶工程》
CSCD
北大核心
2018年第S1期206-211,229,共7页
Ship Engineering
