| The development of society and progress of science and technology have spawned the generation and application of the service robots.With the improvement of people's life and aggravation of social aging problems,the mobile service robots become more and more important in recent years.The robot relies on the sensors carried by itself to perceive the surrounding environmental information and decides the next action and moves from the starting point to the destination,which is the hot problems of the robot field.Hence,the study of this issue has dual meaning of theory and reality for the development of service robots.The main content of this paper is as follows.(1)The hardware control system,sensor systems,software architecture,kinematics model and sensor errors of the mobile service robot are analyzed.(2)The two-dimensional and three-dimensional modeling problems are studied and the random sample consensus and iterative closest points are used to perform coarse registration and fine registration to process the data obtained by the depth sensor.A three-dimensional incremental mapping method based on octree structure is proposed to improve the accuracy of the map and reduce the storage memory of the data.(3)This paper adopts the strategy of localization and motion planning for mobile robots based on the known map.Probabilistic localization methods are analyzed and the particle filter has advantages over the A* algorithm through comparison.In the end,the improvement and realization of the motion planning method based on particle filter are given.(4)The experimental platform has been established and the experimental scheme has been designed.The correctness and validity of this method are verified by simulations and practical experiments. |
PDF Full Text Request