Research On SLAM Mapping And Path Planning Of Indoor Mobile Car With Laser And Vision Fusion

Simultaneous Localization and Mapping(SLAM)technology and real-time path planning are the keys to realize intelligent mobile vehicle at present.Based on this,this paper studies SLAM mapping and path planning of mobile vehicle in large indoor scene.Combining low-cost radar with visual technology can reduce system development cost,and improve the accuracy and completeness of SLAM system.Furthermore,the route planning algorithm is studied in this paper,which is combined with the improved algorithm,so as to enhance the vehicle’s path planning ability.The main work of this paper is as follows:Firstly,the coordinate system and kinematic model of the dolly are established.The main situation and calibration principle of the sensors used in this paper are studied,and the simulation environment of ROS(Robot Operating System)and mobile car is modeled and the sensor function is simulated.Multi-sensor data fusion algorithms and methods were introduced,and multi-sensor data fusion schemes and frameworks were established.Based on the comparison of four typical laser SLAM algorithms,this paper chooses the Cartographer algorithm as the SLAM algorithm,and presents a new method of laser and visual fusion.After converting visual information to pseudo laser information,the extended Kalmanfilter algorithm was adopted to combine the two data.Then,the image was constructed by fusing laser information.Finally,the simulation results indicate that the laser SLAM algorithm is able to construct more abundant map environment.Secondly,the path planning algorithm is researched.JPS(Jump Point Search algorithm)and improved DWA(Dynamic Window Approach)algorithm are used as the path planning algorithm in this paper.The simulation results verify that the fusion of JPS algorithm and improved DWA algorithm improves the overall efficiency of path planning and obstacle avoidance of mobile vehicle and makes it run more smoothly.Finally,based on the ROS platform,the mobile car system experiment platform was built,and the SLAM diagram construction and path planning obstacle avoidance experiments of mobile car were carried out respectively under the real indoor scene.The integrity and richness of SLAM mapping based on laser and vision fusion were successfully verified,and the fusion of DWA algorithm and JPS algorithm was improved,as well as the reliability,effectiveness and stability of the whole dolly path planning system.