| Mobile robots have attracted attention from all walks of life because of their flexibility and various forms,which have been widely used in various fields.For intelligent mobile robot,navigation function is essential,meanwhile path planning is a vital part of navigation system,which affects the overall navigation effect of the robot,and determines the efficiency of the navigation process to a large extent.During navigation,the robot needs to deal with dynamic uncertain obstacles in addition to the stationary obstacle environment.However,the existing planning methods still have some shortcomings in this aspect.Therefore,this paper proposes an improvement on the existing path planning algorithm in the static and dynamic level,and proposes solutions to the local extreme value and the obstacle avoidance of moving obstacles from the global and local perspectives respectively.On this basis,a robot simulation platform was built based on ROS,and Turtlebot3 Burger is used to conduct experiments on the improved planning algorithm,and the validity of the algorithm is verified.Firstly,at the global path planning level,for the unreachable problem of the target point in the traditional artificial potential field method,the method of increasing the gravitational coefficient is adopted to reduce the repulsive potential field gradually when the path approaches the end point.For the local minimum problem,an adaptive sub-goal method is proposed.By introducing the concept of obstacle potential field function,the robot can choose the appropriate position to escape to the sub-goal when it is trapped in the extreme value,and finally re-plan to the end point.Considering that the parameters in the artificial potential field could not adapt to the environment,the particle swarm optimization algorithm was introduced,and the path length and smoothness were taken as fitness functions to solve the local extremum problem,and a smoother and shorter global path was obtained.Secondly,at the local path planning level,the local extremum problem in the dynamic window method is solved by adding the following global path part into the cost function.To solve the problem that the dynamic window method can not avoid moving obstacles in time,the fuzzy control idea is introduced,and the strategy of dynamic adjustment and real-time planning of risk assessment coefficient for moving obstacles is established.A dynamic window fuzzy controller is designed according to the relative velocity relationship between the moving obstacle and the robot to avoid the collision of the robot or the situation that the robot cannot continue planning.Thirdly,in order to verify the effectiveness of the planning algorithm,Gazebo was used to build a simulation test environment under the ROS robot operating system.A simulation map was constructed by SLAM,and the Turtlebot3 Burger model was used for the simulation experiments in both static and dynamic environments.Based on the original global programming A* algorithm,the direction extension improves the path smoothness and reduces the redundant trajectory.In the planning process,Gazebo obstacle is temporarily added to simulate the dynamic environment,and the robot successfully re-plan the obstacle avoidance trajectory under the action of the dynamic window method through the sensor data.Finally,dynamic and static environments are designed in the laboratory respectively,and Turtlebot3 Burger is used to verify the proposed improved algorithm.The improved A*algorithm is tested by the workspace coverage mechanism,and final experiment results indicate that the path quality of the improved algorithm is better than the original one.In the dynamic environment,the dynamic window method can realize the real-time obstacle avoidance in different scenes and different tasks,which verifies the reliability of the proposed algorithm.In conclusion,the study of mobile robot planning in this paper is of reference significance for autonomous navigation of robots. |
PDF Full Text Request