| In recent years,the mobile robot industry has developed rapidly.Mobile robots in people’s lives has become more and more important,especially during the three-year epidemic period.In the field of mobile robots,autonomous navigation technology has always been the research hotspot and one of difficulties of mobile robots,and it is necessary to integrate robot autonomous positioning,path planning and motion control to realize the autonomous movement of robots.This thesis takes the independently built indoor mobile robot as the platform to conduct in-depth research on the key technologies in the autonomous navigation process of mobile robot,specifically as follows:(1)Independently designed and built the hardware platform of the mobile robot.The software system is built on the basis of the designed hardware platform,and the algorithms used in the key modules are studied.Firstly,two mapping algorithms commonly used in robot autonomous navigation,Gmapping and Cartographer,were compared in the mapping module,and the mapping experiments were carried out in a simulation environment.Secondly,in the odometer module,accumulative errors caused by non-uniform ground contact and inaccurate calculation are prone to occur in the wheeled odometer.It uses an extended Kalman filter algorithm to integrate information of the wheeled odometer and IMU(IMU).The cumulative error is corrected.Finally,the localization technology of automatic navigation is studied and the adaptive Monte Carlo localization algorithm is verified and analyzed in the simulation environment.(2)For global path planning,aiming at the security problems caused by the planned path being too close to obstacles in the planning process of A* algorithm,this thesis proposes A* algorithm based on security improvement,which adds obstacle evaluation factors into the eight-way search process of the child nodes of the traditional A*algorithm.Remove the obstacle child node and reduce the probability of the obstacle adjacent node being added to the final path.As for the problems of redundant nodes and too many corners in the traditional A* algorithm,this thesis reduces the number of redundant nodes by judging whether there are obstacles between the links of key points of the path,and finally smooths the path by using the segmented Bessel function.The path planning simulation is carried out with ROS simulation environment,and the results show that the improved A* algorithm can improve the security of the planned path and make the path smoother.(3)For local path planning,aiming at the jitter problem in the running process of DWA(Dynamic Window Approach)algorithm,this thesis proposes an adaptive directional Angle evaluation factor,which adaptively reduces the proportion of directional Angle in the selection of simulation trajectory to improve the smoothness of robot running.Simulation experiments with MATLAB are carried out to prove the effectiveness of the improvement.Finally,the above studies were verified and analyzed on the built hardware platform.First,the Cartographer algorithm was used to construct the real laboratory environment.Based on this,adaptive Monte Carlo positioning experiments and obstacle avoidance experiments were carried out,and the improved path planning algorithm was verified and analyzed.Experimental results show its feasibility on real hardware devices. |
PDF Full Text Request