Research On Mobile Robot Map Building And Path Planning Technology In Indoor Environment

Today,the development of mobile robots has occupied an increasing proportion in the fields of life,entertainment,industry,and military.The rise of artificial intelligence,big data and cloud computing has also contributed to the development of mobile robots.Nowadays,mature sweeping robots,guided robots,outer space landing robots,robotic pipelines in industrial automation,etc.,the penetration of robots in all profession has reached an unimaginable level.Mobile robot research mainly solves path pla nning problems,map building problems and real-time positioning navigation problems.The research on single mobile robots has gradually become saturated,and the research of multi-mobile robots has become more and more popular.Compared with the research of single mobile robots,multi-agent based multi-mobile robots are more valuable in the study of large scenes and complex environments.Firstly,the path planning problems of single mobile robot are studied in this thesis,then we focus on the path planning problems of multiple mobile robots,which is based on Multi-Agent System.At last,the paper studies multiple robots cooperate to build map,which is based on Multi-Agent System.The following three research results are presented in this thesis,which are the mobile robot path planning based on the experience learning fish swarm algorithm with detection operator,the real-time navigation of multi-mobile robot based on the expansion logic experience fish swarm algorithm and the map building of multi-agent unknown environment based on edge gradient algorithm,respectively.Simulation experiments verify the reliability,convergence and rapidity of the above three algorithms.The mobile robot path planning based on the experience learning fish swarm algorithm with detection operator,which is mainly used in the process of studying a single mobile robot.The group intelligence algorithm found in the process of applying the artificial fish swarm algorithm to the grid map is always having the problem of suboptimal solution affect the best solution and the lengthy problem of the path.By designing the empirical learning mechanism,the fish population has ability to learn the wrong experience,and finally solves the problem of the optimal solution of the suboptimal solution.By designing the detection operator,the problem of the path length in the planning process of the fish swarm algorithm is solved.The expansive logic based fish swarm algorithm designed in this thesis is mainly to study the theoretical algorithm while considering the application level,and extends the research results of a single mobile robot to the field of multi-mobile robots.When the map is too large,the mobile robot is often assumed to be a particle when studying the algorithm.However,in practical applications,mobile robots must have enough space to adjust their position and posture.The shortest path planned by theoretical algorithms often leads to the implementation of mobile robots being too extreme.It is impossible to guarantee that mobile robots can turn around at any time and the ability to adjust posture.By designing the expansion logic algorithm,this problem can be effectively solved,and a scoring mechanism is introduced to score the path so that the optimal path is not missed.The edge gradient algorithm designed in the thesis is mainly to solve the problem of fast map construction of multi-mobile robot based on multi-agent system.In larger scenarios,the efficiency of a single mobile robot to build a map is limited,but the efficiency is multiplied by building a map with multiple mobile robots.Therefore,with reference to the function and characteristics of the hippocampus in the brain,an edge gradient algorithm is designed to realize the ability of multiple mobile robots to collect map information around obstacles.To complete the splicing of map information efficiently and quickly,a virtual coordinate system is designed.Finally,a kind of experiment platform is developed,the reliability and convergence of the experience learning artificial fish swarm algorithm and the expansion logic algorithm based on the test operator experience are verified.