Research On Variable Autonomy Level Mechanism Of Unmanned Ground Vehicle

With the continuous deepening and enrichment of related research on the combination of artificial intelligence technology and the unmanned systems,unmanned ground systems have been widely studied and applied in mission scenarios.In complex tasks,users are still required to provide varying degrees of assistance to achieve safety,controllability and maximum efficiency because of the incomplete intelligence of the unmanned ground system.This implies the importance of variable autonomous systems to the current unmanned systems.The current variable autonomous system still has the following problems:(1)The unmanned system currently lacks a universal variable autonomy level mechanism.It is difficult to "solve simple tasks independently,communicate and negotiate complex tasks,and remotely operate difficult tasks by users".(2)The existing human-computer interaction is usually dominated by the user’s experience.For the system itself,there is currently no scientific method that can infer the autonomy level of the human-machine system in real time and accurately according to the environmental situation,task conditions and user capabilities.In order to solve the above problems,this paper conducts research on the variable autonomy level mechanism of unmanned ground systems,and conducts simulations and actual experiments for the classic tasks of unmanned ground systems.The work and contributions are as follows:(1)Aiming at the variable autonomous requirements of human-machine systems,an "Adaptive-OODA-H" control system is proposed.This paper proposes a new control system that dynamically adds user’s will to the traditional OODA cycle,and elaborates the structure of the variable autonomous mechanism of the unmanned ground system and the variable autonomous level classification scheme.(2)In order to determine the human-machine system work distribution method according to the environmental situation,task situation and user capabilities,this paper proposes an autonomous level reasoning algorithm based on the situational awareness model.The situational awareness model firstly models the integrity of the ground unmanned system,and builds the target perception situation based on the task type;secondly,the user behavior capability model is introduced to establish the connection between the current situation,user capabilities and level of involvement.This paper establishes a situational utility value for each autonomous level,ranked the best by the TOPSIS method of mixed indicators through a multi-attribute decision-making level reasoning model.Finally,an simulation experiment is executed to verify the effectiveness and flexibility of the model.(3)To realize the autonomous hierarchical scheduling behavior according to the autonomous level,this paper proposes an 8-tuples state machine autonomous hierarchical scheduling strategy.The research presents the theoretical overview of the autonomous hierarchical scheduling strategy.Then,an 8-tuple autonomous hierarchical scheduling model is proposed,and the model is implemented based on the SMACH state machine.Finally,the functional verification of the autonomous hierarchical scheduling model is carried out,and the results indicate that the real-time performance and reliability of the scheduling behavior are within the expected acceptance.