Research On Optimal Path Planning Of Mobile Robots

Optimizing the path of a mobile robot is a very challenging subject in the field of robot research,and it has a certain role in promoting the future development of navigation system.The path planning of multiple mobile robots in dynamic environment needs to find optimal paths from the starting point to the target point for each mobile robot in a working space with dynamic and static obstacles,and ensure that the mobile robots and obstacles do not touch.At present,most works focus on the global path planning problems of which the environmental information is known and there are no dynamic obstacles.The path planning of multiple mobile robots in dynamic environment is still a difficult problem to be solved.In this paper,with unknown background,unscented Kalman Filter neural network was used to establish a dynamic environment model.Then the AP clustering algorithm was used to realize the automatic partitioning of the environment.Finally,combined with local collision avoidance and improved dynamic programming algorithm to solve to solve the path planning.The main work of this paper includes:Firstly,it summarizes the advantages of multi-robot system in the implementation of path planning tasks.Then the simulation model of multi-robot system was presented,including the physical model,the motion model,the sensor model and the communication model.The coordinate transformation formula is given,which can be used to determine the node information of the robot during the moving process.In order to improve the positioning accuracy of three-dimensional space,this paper presents a three-dimensional space intersection centroid algorithm.The algorithm has a high spatial positioning accuracy,so the position of the mobile robot can be better determined.In the context of known application background,the dynamic programming mathematical model of mobile robot is established according to the given scenario.With regard to path planning success rate and average time spent on path planning,compared with Dijkstra algorithm and Floyd algorithm,it is found that the dynamic programming algorithm combined with the road weight coefficient has high efficiency and the advantages of reducing the unstable state of the mobile robot in the course of action,it also can effectively suppress the algorithm into a local optimum.This algorithm can realize the fast search of the optimal path from the starting point to the target point,the main robot can reach the target point in an optimal way according to the coordinate position from the slave robot.In the context of unknown application background,this paper presents a dynamic evolutionary modeling method based on unscented Kalman Filter neural network(UKFNN)and its evaluation index.The effectiveness of the model and the evaluation index were verified by solving the practical problems.The results showed that theaccuracy of dynamic UKFNN modeling is 22.16% higher than that of static back propagation neural network(BPNN).According to the data of continuous sections from Chongqing Urban garden station to Renhe station,this paper puts forward different segmentation strategies,and uses the AP algorithm to realize the automatic division of the region.Then,the path planning of the robot is carried out according to the information collected by the continuous section,the optimal collision free path is found.In this method,Despite the lack of cognitive memory of the scene,but the robot can obtain the coordinate position information in the process,making use of the principle of optimization and feedback principle in different sections,Combined with Dynamic programming algorithm of road weight coefficient.The calculation of local planning is greatly reduced,Moreover,the convergence of the algorithm results to the optimal solution is guaranteed.The results of this research can make decisions according to the change of scene,especially in the process of the operation contains many of uncertain moving obstacles,it is easier to get the optimal path.It can be widely used in the field of remote control robot,target detection robot,service perception robot and vehicle navigation.