Research On Tracking Control Of High-speed Underactuated Unmanned Surface Vessels

The unmanned surface vessels(USV)is an unmanned special platform that connects air,ground,water and underwater,and plays a vital role in the national military and civilian aspects of future marine engineering.The tracking control,as an important guarantee for USV to complete various tasks,has very important scientific research value.The research focus of this thesis is to design the tracking controller of the underactuated USV based on non-linearity,model uncertainty,time-variability of external disturbances during high-speed navigation of underactuated USV,as well as parameter adaptability,controller robustness,tracking effect stability and navigation safety in practical engineering applications,and verify the applicability of the controller through simulation and actual USV test.The main research contents of this thesis include:1.Regarding the applicability of the conventional line-of-sight method under different speeds,different path combinations and different sea conditions,the dynamic-integral-line-ofsight(DILOS)method and the S-plane-angle-line-of-sight(SALOS)method are improved and designed.It can improve the adaptive ability of the guidance parameter and eliminate the influence of drift angle caused by disturbances,and can improve the tracking accuracy,completion degree and the robustness of the controller when the USV path is switched.2.For USV high-speed navigation,the parameter adjustment process of PID controller is affected by communication,safety factor and sea condition.Combined with fuzzy adaptive control,PID control parameters are adjusted online through fuzzy rules,which improves the robustness and adaptability of the USV path tracking process,and ensures the safety of navigation.3.For USV problems such as model parameter uncertainty and extermal disturbance uncertainty,Minimum-Learning-Parameter-based(MLP)Radial Basis Function Network(RBF)sliding mode adaptive model-free control mechanism is designed.Based on the robust sliding mode variable structure control,the unknown parameter of the system is approximated online by RBF-NN,and MLP is used to reduce the RBF＿NN on-line estimator,the computational time and complexity,improves the efficiency of high-speed UAV computing unit,reduces a lot of computing resources,reduces the chattering problem of sliding mode control by introducing hyperbolic tangent function,and improves the stability of the system.4.For USV problems such as model parameter uncertainty,extermal disturbance uncertainty,limited communication bandwidth and navigation safety,the prescribed performance control mechanism based on input quantification is designed.It uses a combination of MLP and RBF to estimate the system unknown items online.The HLP quantizer is used to convert continuous signals into discrete signals,reducing communication load.The tangent conversion function is used to limit the trajectory tracking error,which ensures the safety of navigation.