Research On Mobile Robot Path Planning For Target Search Task

Target search is one of the key functional modules for intelligent mobile robots to perform tasks.Its potential applications include factory inspection,deep-sea exploration,search and rescue in disasters,etc.Currently,the related researches do inadequate consideration about the uncertain factors in the process of search task execution.The existing search modes have lower efficiency in the environment with more uncertainties,and they take longer to complete search tasks.To solve the above problems,a probabilistic topological environment model for target search in indefinite environment is designed,and a hierarchical path optimization strategy based on the model is proposed.Two different target search task patterns are studied.The specific research contents are as follows:In view of the uncertainty factors of environment that can not be accurately expressed by the determined environment model,the probability topology model is designed,the distribution characteristics of the target in the environment are analyzed,and then the probability calculation model is constructed.According to the actual requirements of different tasks,the objective functions of the problems under two probability maps are constructed.To solve the path planning problem of the target search and retrieval task in uncertain environments,a motion planning strategy based on improved modified circle algorithm and improved Rapid-exploration Random Tree(RRT)algorithm are proposed.In the upper-level,the task matching probability map is constructed for sequence planning,and the original improved circle algorithm is optimized by the principle of weight optimality,which can significantly improves the global optimization performance of the amendment-circle algorithm.In the local path planning between lower-level observation points,target preference and collision detection mechanism are applied to classical RRT under feature maps,which can reduce the randomness of search process and the probability of falling into local trap.Simulation experiments and analysis show that the proposed strategy can significantly improve search efficiency,and it is suitable for target search retrieval tasks in uncertain environments.To solve no-retrieval tasks of robot target search,a two-level search strategy based on probability map is proposed,which takes the mathematical expectation of search time as the optimization goal.In no-retrieval tasks,the search can be stopped when the target is found,and the amendment-circle is not suitable for such tasks,so a improved ant colony algorithm is used in upper level sequence planning for optimizing the access order of observation points,and then,the improved RRT algorithm is employed for local path planning,and finally the optimal search path in the expected time is obtained.Simulation and analysis show that the proposed strategy can significantly reduce the desired search time,which is suitable for the target search task of autonomous robots with shorter expected time.