Optimization And Research Of Robot Path Planning Method Under Temporal Task Constraints

The widespread application of mobile robots makes the path planning become the research focus.Path planning with temporal task constraints is one of the key technologies in path planning research.The theory of linear temporal logic(LTL)provides a theoretical basis for describing the temporal tasks of robots,and the LTL formula describes the complex temporal tasks normatively.At present,there are certain solutions to the path planning problem of temporal task constraints,but the complexity of this method is high,and it has some limitations in practical applications.Multi-robot cooperation to complete assigned tasks is also a research direction for researchers.Therefore,it is of great significance to optimize the path planning approach which satisfies constraints of complex temporal tasks,and to research the path planning of multi-robots.The paper does the following research on the above issues:(1)The constrained path planning approach to temporal tasks in a single robot static environment is studied.The existing algorithms have a good effect on solving the path planning problem of temporal tasks,but with the increase of task points,the algorithm response speed becomes very slow.Therefore,aiming at the problem of high time complexity of the existing algorithm,three kinds of constraint criteria are proposed to optimize the feasible network topology,reduce the state points that need to be searched,and improve the speed of the solution process.In order to improve the response speed of the algorithm,a method of small root heap is proposed to optimize the bottom path search algorithm.(2)A single robot path planning approach based on Markov decision process(MDP)is studied in dynamic environment.Aiming at the shortcomings of the existing algorithms,a non-deterministic automaton is proposed to replace the deterministic automaton,which solves the situation where the deterministic automaton cannot parse some LTL formulas,and builds a finite weighted switching system based on a dynamic environment model to build a feasible network topology graph.Moreover,the proposed three constraint criteria are able to reduce the state of a feasible network topology graph,and then construct a Markov decision process based on the remaining state points.The state points are used to construct Markov decision process,and the value iteration algorithm with dynamic programming property is proposed to replace the traditional strategy iteration algorithm,which can reduce the difficulty of solving Markov decision process and improve the response speed of the algorithm.(3)The path planning approach to multi robots based on game theory and LTL is studied.In order to solve the shortcoming of path planning for multi robots,the game theory in economics is used to solve the problem of path planning for multi robots that satisfies the complex temporal tasks.The multi robots are regarded as players in the game,and the tasks that the robots need to complete are divided into sub-robots.The Nash equilibrium is solved by the double-drawing method.The value(distance)of each task is calculated by LTL combined with a path search algorithm.The robots are able to judge the benefits and losses(traveling distance)of itself,and decide the tasks that each robot needs to complete.While cooperating to complete the task,ensure that the value obtained is the highest(the shortest traveling distance).Numerical experiments and simulations are provided to illustrate the validity of the proposed path planning approach.The results show that it has a good effect on the path planning problem of single robot and multi-robots.The proposed approach is able to solve the path planning problems of complex temporal tasks and the path planning problems of robots in static and dynamic environments.Moreover,the proposed approach promotes the application of robots in practical environments.