Adaptive Dynamic Programming-Based Model-Free Optimal Trajectory Tracking Control Of An Unmanned Surface Vehicle With State And Input Constraints

Unmanned surface vehicles(USVs)will be widely used in future Marine defense and Marine resources development due to their high autonomy and intelligence.However,in view of the current technical level,there are obvious bottlenecks in the motion control performance of USV in the aspects of model dependence,control energy consumption optimality,unknown dynamics,etc.States constraints,unknown system dynamic and unknown environment disturbance are considered in this thesis.For above problems,the optimal control strategy based on adaptive dynamic programming is put forward.Unmanned surface vehicle can track desired trajectory by using the designed strategy with unknown system dynamic and state and input constraints.The main contributions are as follows:Firstly,an optimal control strategy based on adaptive dynamic programming is proposed to solve the optimal tracking control problem of USV with unknown system dynamics.The neural network approximation technology is used to quickly identify and compensate the unknown system dynamics to realize the modelless tracking control of USV.Then,under the framework of backstepping technology,the adaptive dynamic programming algorithm is applied to design the optimal control strategy and the optimal trajectory tracking control of USV is realized.Strict theoretical derivation proves the stability of the system and the system errors are semi-global uniformly ultimately bounded.A large number of simulations and comparative analyses show that the proposed optimal control strategy can track the reference trajectory more accurately with unknown system dynamics compared with the traditional backstepping control strategy.Secondly,an optimal control strategy combining barrier Lyapunov function and adaptive dynamic programming technology is proposed to solve the optimal control problem of trajectory tracking for model-free USV with constrainted states.The logarithmic barrier Lyapunov function is adopted to ensure that the tracking error is limited in the specified range,so as to solve the problem of the pose and velocity limitation of USV in the trajectory tracking process.At the same time,the optimal control strategy is designed by combining backstepping technology and policy iterative algorithm to realize the optimal trajectory tracking control of USV.Strict theoretical derivation proves the stability of the system and the convergence of errors.Simulation results verify that the proposed control strategy can track the reference trajectory more accurately compared with the adaptive neural network algorithm and keep the pose and velocity within the specified range during the tracking process.Finally,a control strategy combining sliding mode technique,barrier Lyapunov function and adaptive dynamic programming is proposed to solve the optimal control problem of trajectory tracking for model-free USV with constrainted state and dead-zone input.To be specific,sliding mode technique is used to compensate the external disturbance to improve the robustness of the system.Then,barrier Lyapunov function is used to limit the state of USV and the optimal control strategy is designed under the actorr-critic structure.So the optimal trajectory tracking of USV can be achieved while the pose and velocity are kept within the specified range.Theoretical derivation proves the stability of the system and the error can converge to zero.Simulation results show that the designed control strategy can track the reference trajectory more accurately with unknown disturbance and dead zone input compared with the adaptive fuzzy optimal control and the pose and velocity of USV are kept within the preset range during the tracking process.