Research On The Key Technologies Of Mobile Robot Navigation System

With the increasingly complex working environment of mobile robots,the past mobile robot navigation technologies have some shortcomings,such as slow response and wrong decision making.Therefore,new methods and solutions are needed for environmental perception and motion planning of mobile robots,which are key technologies of navigation.Robots use sensors to detect the surrounding and expect to reach the destination smoothly with the aid of sensors.However,in unknown environments,robots may encounter various obstacles such as bumpy obstacles and pedestrians,while the information provided by the in-vehicle sensors is limited,that cannot be obstructed,which results in it is unable to determine the size and shape of the obstruction.Even in dynamic environments,there are many path planning algorithms,some of which can be applied to unknown environments.However,these algorithms are complex and that robots often cannot quickly make a reasonable way to avoid when encounter obstructions and pedestrians.Another problem is that,in completely known environments,because mobile robots have far more than two navigation points in practice,the robot cannot guarantee that it is the best route that in many navigation points.In view of the above problems,this paper starts from a key technical issue of mobile robot navigation to implement a behavior-based navigation system for mobile robots in unknown circumstances and a multi-point navigation system based on genetic algorithms in known maps.In the unknown environment,behavioral libraries are set up to meet the navigation needs,including go-to-goal,following wall,avoiding obstacles in unsafe distance,and avoiding pedestrians;In known map,the combination of genetic algorithms,A~*algorithm and DWA algorithm is used to obtain the optimal route for multi-point navigation.The main work of this paper is as follows:(1)According to the environment information of laser radar scanning and wheel encoder value,the Monte Carlo algorithm is used to get the robot accurate pose estimation;(2)In the unknown environment,according to the environment information of laser scanning,the environment around the robot is analyzed to select the appropriate behavior to run;(3)Genetic algorithm is used to plan the motion sequence of multiple navigation points.the A~*algorithm is used to globally plan the path between two navigation points.the DWA algorithm is used to plan the path of local planning that analyze the area that robots cannot pass;(4)Design the robot motion model for the unknown navigation system and the navigation system in the known map.Based on the above method,this paper experimentally simulated the path planning effect of robot navigation in unknown environment and the multi-point navigation effect of robot navigation in the existing map several times.The later uses the self-developed mobile robot platform in real environment many times.The experimental results verify the effectiveness of the robot navigation system based on behavior and the multi-point navigation system based on genetic algorithm in the known environment.The research on the mobile robot navigation technology has certain theoretical significance and applied value.