Research On Hierarchical Obstacle Avoidance For Unmanned Surface Vehicle In Complicated Marine Environments

Unmanned surface vehicle(US V)is a kind of important autonomous marine robots that have been studied and gradually applied into practice.However,the autonomous navigation of USVs,especially the issues of obstacle avoidance in complicated marine environments,is still a fundamental problem.In order to realize the safe navigation of USV in the complicated marine environment,the research is executed from four aspects:global trajectory planning module,far-field obstacle avoidance module,near-field obstacle avoidance module,adaptive mechanism of near-field obstacle avoidance,and the work is as follows:Firstly,the problem of global trajectory planning for USV is considered.In the research,the gried-based world model is constructed on the base of electronic chart,and the multi-objective constrained optimization model is proposed by the analysis of global trajectory planning problem.Focus on the constraints of global trajectory planning of USV,a distance function and two-penalty function is proposed.By introducing the Pareto strength and the minimum gap model into the research,a multi-objective genetic algorithm based global trajectory planning algorithm is proposed for USV.At last,the validity of the algorithm is demonstrated in the experiment.Secondly,accoding the obstacle avoidance characteristics of USV in the complicatd marine environment,the far-field obstacle avoidance module is added between global trajectory planning module and near-field obstacle avoidance module.In the study,far-field obstacle avoidance region,collision region and International Collision Regulations of USV are defined.On the base of sparse A*and feasible direction angle set,a daynamic local trajectory replanning algorithm is proposed for the far-field obstacle avoidance problem of USV.Meanwhile,the feasible direction angle set which is calculated according to the constraints from International Collision Regulations and collision conflict of obstacles.In the simulation,various senarios are designed to verify the performance of the algorithm.The vaidity of the algorithm is demonstrated in the experiment by designing different types of scenarios.Thirdly,at present,various local obstacle avoidance algorithms have been proposed for USVs.However,the majority of algorithm results have only been verified by simulations,with only a few demonstrated by low-speed USVs(= 5knots)in the sea.Therefore,the issues of obstacle avoidance for high-speed USVs in complicated marine environments have to be studied.This paper presents a novel local reactive obstacle avoidance algorithm for high-speed USVs.In this algorithm,the direction steady-state model and translational velocity model are proposed for the base normal motion and base motion control characteristics of high-speed vessels.During navigation,USV motion depends on the guidance angle and guidance translational velocity.The validity of the proposed algorithm is verified by experiments in simulations and sea trials.Finally,during the navigation in the complicated sea-state marine environment,USV will be affected by sea wind,currents and other disturbance factors.However,the disturbances from outer enviorments are ignored by the near-field obstacle avoidance module.In the study,an adaptive avoidance decision process model is proposed for USV to solve the problem of obstacle avoidance in complicated sea-state marine environments.By analyzing the disturbance factors from complicated sea-state marine environments,the model is constructed on the base of Sarsa on-policy reinforcement learning algorithm.By setting the GLIE(greedy in the limit and infinite exploration)as the action exploration,the convergence of the adaptive avoidance decision process has been proved.The convergence shows that the action can converge to the optimal action strategy with the probability value of one.The proved result demonstrates that the performance of obstacle avoidance of USV in the complicated sea-state marine environment can be enhanced under the action of on-policy reinforcement learning algorithm.According to the convergence result,a novel adaptive obstacle avoidance algorithm based on the Sarsa on-policy reinforcement learning algorithm is proposed for USVs.The proposed algorithm is composed of local avoidance module and adaptive learning module,which are organized by a "divide and conquer" strategy-based architecture.The course angle compensation strategy is proposed to offset the disturbance from sea wind and currents.In the design of payoff value function of the learning strategy,the course deviation angle and its tendency are introduced into action rewards and penalty policies.The validity of the proposed algorithm is verified by comparative experiments in simulations...