Research On Mobile Robot Path Planning Based On Improved Hybrid A~* Algorithm

With the development of electronic information technology,mobile robots have been applied in many fields such as industrial manufacturing,medical education,logistics and transportation,autonomous driving,and disaster relief.Path planning,as an important research direction in the field of mobile robot,affects the performance of mobile robots'autonomous navigation.As many traditional path planning algorithms have a defect that the planning paths are not smooth enough,the algorithms are not effective in practical applications.The Hybrid A~*algorithm considers the constraint on the minimum turning radius of the mobile robot,and it can restrict the path curvature during path planning.So the path planned by Hybrid A~*algorithm is more suitable for mobile robots to run in practice.In terms of the minimum turning radius constraint of mobile robots,this paper proposes a path planning problem with curvature constraints.In order to solve the problems,the paper firstly studies the Hybrid A~*algorithm,and elaborate the principle of the algorithm from four aspects:the node property,the node expansion method,the heuristic function and the algorithm flowchart.It shows that the Hybrid A~*algorithm can plan a path that meets the curvature constraints,while the planned path still has two defects.One is that the distance between the path and the obstacle is too close,which may cause the mobile robot to collide with the obstacle.The second is that the path steering state changes rapidly,which will cause a large sudden change in planning speed.In this case,the control module is required to have high performance to track the corresponding speed change,otherwise the actual plan would deviate from the planned path.In this paper,two factors are proposed to overcome the aforementioned defects of the Hybrid A~*algorithm.(1)Aiming at elongating the distance between the path and the obstacle,this paper addresses a piecewise model of the distance penalty factorc_d.The piecewise model divides the map into the collision occurrence area,the collision risk area,and the collision risk-free area.The distance penalty factor is modeled in different areas according to application requirements.In addition,a magnification factork_d and a risk distribution factoraare set in the model to adjust the overall amplitude and distribution of the distance penalty factor,respectively.Finally,the distance penalty factorc_d is superimposed on the path cost function g(n)to reduce the priority of the nodes closer to the obstacle and make the path away from the obstacle.(2)Aiming at overcoming the defect of large changes in path steering state,this paper proposes a model of steering penalty factorc_t.The model first needs to add a steering sign t_i to the node property,and then uses the absolute differential steering sign(35)t_i between the child node and the parent node to characterize the degree of path steering state change.In addition,a magnification factork_t is set in the model to adjust the overall magnitude of the steering penalty factor.Finally,the steering penalty factorc_t is superimposed on the path cost function g(n)to reduce the steering state change of the path and the required performance of the control module.Moreover,this paper uses a lot of simulation and mobile robot platform experiments to verify the effectiveness of the improved Hybrid A~*algorithm.