基于UKF的无人艇操纵响应模型的参数辨识

Parameter Identification of Maneuvering Response Model of Unmanned Surface Vessel Based on UKF

为解决无人艇(unmannedsurfacevessel,USV)的操纵性预报问题,提出一种基于无迹卡尔曼滤波(UnscentedKalmanFilter,UKF)的辨识方法用于获取无人艇二阶非线性操纵响应模型参数.首先在Matlab平台上进行20°Z形仿真操纵实验,基于差分法对辨识模型离散化,运用设计的UKF算法展开辨识,并分析其辨识过程中的收敛性,将辨识结果在实验室平台上进行10°,20°和30°半物理仿真正弦和Z形操纵性实验.结果表明,UKF是一种有效的辨识算法,辨识结果能有效的预报无人艇的操纵性预报.

In order to solve the maneuverability prediction problem of unmanned surface vessel (USV), an identification method based on Unscented Kalman Filter (UKF) was proposed to obtain the second-order nonlinear maneuvering response model parameters of unmanned craft. The 20° zigzag simulation manipulation experiment was carried out on the Matlab platform. The identification model was discretized based on the difference method. Then the designed UKF algorithm was used to identify and analyze the convergence in the identification process. Finally, the sine and zigzag maneuverability experiments of 10°, 20° and 30° semi-physical simulations were carried out based on the identification results. The results show that UKF is an effective identification algorithm, and the identification results can effectively predict the maneuverability prediction of unmanned surface vessel.