Online Identification and Prediction of USV Maneuverability Parameters Based on Multi-innovation Recursive Least Squares Algorithm with a Forgetting Factor

The parameter accuracy of the ship motion model has an important influence on the maneuverability prediction accuracy of the high-speed unmanned surface vehicle (USV). In order to solve the problem of data saturation and lack of reasonable use of historical data by recursive least squares method (RLS), the multi-innovation recursive least squares algorithm with a forgetting factor (FF-MRLS) is proposed to identify the parameters of second-order nonlinear maneuvering response model and achieve accurate online prediction of maneuvering motion of USV according to the identification results. The zigzag test is carried out based on the second-order response model parameters of an USV under complex interference. The FF-MRLS is designed is designed to identify the model parameters, and the maneuvering motion is predicted online based on the identified parameters. The experimental results show that the maneuvering motion of USV under complex interference can be accurately predicted within a certain time range by identified parameters obtained by FF-MRLS, and the identification results can meet the requirements of maneuvering prediction accuracy.