Research On Map Representation And Navigation Method For Mobile Robot Environment Based On Multi-sensor Fusion

In the past few years,mobile robots have been used in an increasingly wide range of applications,ranging from home services to industrial healthcare.The shadow of mobile robot technology is everywhere in human life.Robots can efficiently complete designated tasks in different environments,achieve autonomous navigation,and cannot do without the construction of environmental maps and path planning.However,in complex environments,there are many limitations in using a single sensor to obtain environmental information to build a map,and traditional path planning algorithms have many shortcomings.Therefore,this article proposes an environmental map representation and navigation means for robots based on the fusion of Li DAR sensors and depth camera sensors to build more comprehensive and higher quality environmental maps and improve the path planning algorithm so that mobile robots can realize the function of autonomous navigation.The research work in this thesis consists of the following:(1)To address the limitation that the current 2D Li DAR sensor can only detect a flat object,this thesis proposes a method of fusing RGB-D cameras to construct a map.First,the data of the two sensors are collected and jointly calibrated.Then the grid map is selected as the environmental map of this thesis by comparing the map representation model.Finally,the data of the two sensors and the obtained odometer data are jointly input into the RTAB-MAP algorithm to output a fused map with more comprehensive environmental information.(2)Based on the multi-sensor fused environment map,this thesis proposes a hybrid path planning method that integrates global path planning and local path planning.The traditional A* algorithm has been improved by quantifying map information and incorporating obstacle probability into the evaluation function,reducing unnecessary node expansion.In addition,the child node selection strategy has been improved to enhance the safety of the robot.Furthermore,a bi-directional smoothing optimization algorithm has been designed to improve the smoothness of the path;finally,the improved A* algorithm was verified through simulation,which showed a reduction of 0.2 seconds in planning time compared to the traditional A* algorithm,a decrease of 57.1% in turning points,and a decrease of 83.8% in turning angles.The optimization effect was significant.(3)The traditional DWA algorithm is improved by introducing the global static path point planned by the improved A* algorithm as the path target point into the DWA algorithm,and adding a smoothing parameter to the evaluation function,improving the weight parameter in the evaluation function so that it can be automatically adjusted,Solving the problem of falling into local optima and also smoothed the path.Finally,simulation experiments were designed in both static and dynamic environments.The results show that the improved algorithm not only enhances the safety of the robot,but also optimizes the smoothness of the overall path,and solves the problem of falling into local optimal solutions.(4)A mobile robot experimental platform was constructed to design experiments for the multi-sensor fusion mapping module and navigation planning module in a laboratory environment.The sensors were jointly calibrated using calibration tools,and the real-time performance and completeness of the environment mapping method based on multisensor fusion were verified through experiments.The effectiveness and superiority of the improved local path planning algorithm were also validated through experiments.