Research And Implementation Of Autonomous Navigation Of Food Delivery Robot Based On Laser SLAM

With the development of artificial intelligence and machine learning technology,service robots,especially intelligent food delivery robots,have gradually entered people’s daily life,and shopping malls,hotels,and guesthouses,especially due to the epidemic and other reasons,have a greater demand for food delivery service robots.When the robot is in a complex and changeable real environment,how to use the established map and how to effectively plan the best driving route has become a key factor in the evaluation of food delivery robots.Based on the ROS operating system and the hardware development platform of mobile robots,this paper constructs and optimizes the map,proposes an autonomous navigation algorithm for food delivery robots with improved A* algorithm,and applies it in complex scenarios such as self-built simulated restaurants,which can basically realize the autonomous navigation of robots The main research work of this paper is as follows:Firstly,the working scenarios and requirements of the food delivery robot are expounded,the kinematic model is established through the kinematic analysis of the food delivery robot,and the overall system scheme design of the robot is established.Secondly,the unscented Kalman filter algorithm(UKF)is used to effectively fuse the IMU and odometer data together to construct a raster map,and the observation information obtained by lidar scanning is also join to RBPF-SLAM(Rao Blackwellized Particle Filter-Simultaneous Localization and Mapping)to estimate the robot’s pose,thereby improving the accuracy and reliability of the map.Order to improve the pose estimation accuracy of the robot and the accuracy of the raster map,the KLD(Kullback-Leibler Divergence)algorithm is introduced in the resampling stage of RBPF,so as to better improve the pose estimation accuracy of the robot and draw the raster map more accurately.Thirdly,based on the shortcomings of the traditional A* algorithm,the algorithm structure of the A* algorithm is improved,so that when an obstacle is encountered in the process of searching for a path,the algorithm defines the node where the mobile robot is currently located as the parent node,instead of returning to the starting point to search for the path again.Aiming at the problems of A* algorithm in global path planning,such as unsmooth path twists and turns and many turning points,this paper proposes to design a new heuristic function to optimize path smoothing based on the combination of Manhattan distance and Euclidean distance,and based on the edge arc optimization principle in mathematics,a new path inflection segment optimization strategy is proposed to optimize the problem of more turning points,which can effectively reduce the system calculation cost and path planning time.The principle of TEB local path planning algorithm based on multi-objective optimization is analyzed in detail and system simulation is carried out.Finally,based on ROS(Robot Operation System),a simulation experimental platform is built and sensor data is collected and processed,SLAM and navigation modules are designed and the above algorithms are integrated into the software framework for use verification,and the SLAM and navigation algorithms of food delivery robots are experimentally verified in self-built experimental scenarios and complex laboratory scenarios.Experimental results show that the improved algorithm can accurately realize the autonomous navigation of food delivery robot based on laser SLAM.