Study On The Decision Making Method Of Multi-USV Variable Autonomy Level In Human/unmanned Cooperation

Unmanned surface vehicle(USV)is an unmanned intelligent water surface system,which has many advantages such as safety,high efficiency and autonomy.However,there are limitations for USV to perform tasks in complex environment and achieve full autonomy.Therefore,the autonomous level of surface unmanned aerial vehicle needs to be changed dynamically,so that the multi-USV can adjust the autonomous level dynamically,interact and coordinate with the manned system,and complete the task reliably with the effective intervention of the operator.In this thesis,the autonomous hierarchical decision-making and adjustment method of human/unmanned cooperative multi-USV is studied.In this thesis,the architecture model of human/unmanned cooperative multi-USV system is constructed,the trigger mechanism of multi-USV system’s autonomous level adjustment is designed,and the autonomous level adjustment method of multi-USV system based on group multi-attribute decision-making and the autonomous ability improvement method of multiUSV system based on human-computer hybrid Q-learning are proposed,so that the multi-USV system can perform tasks according to the task,environment and user In the process of cooperation,the intelligent level of unmanned system is improved by absorbing the experience of manned system.In this thesis,three typical multi-USV mission modes of formation navigation,area patrol and formation escort are taken as examples,and the simulation platform of human/unmanned multi-USV cooperative control is built by using Vega simulation software,which verifies the rationality of the architecture model of human/unmanned cooperative multi-USV system.At the same time,it verifies the effectiveness of the multi-USV system autonomous level adjustment method based on group multi-attribute decision-making.The autonomous ability of human/unmanned multi-USV cooperative multi-USV system is improved,which makes human/unmanned cooperative multi-USV system achieve the effective combination of human experience and system intelligence.