Path Planning And Autonomous Navigation For Unmanned Vehicles Considering Multiple Scenarios

With the continuous development of science and technology,unmanned vehicles are widely used in military fields,including material transportation,hazardous operations,and special missions,etc.However,there are still many problems with their autonomous navigation due to the environmental factors and the structural constraints of the unmanned vehicles themselves.Relying on the National Natural Science Foundation of China,this project aims to solve the key problems of unmanned vehicle path planning,such as poor feasibility,low safety,poor scene adaptability,and insufficient adaptability of autonomous navigation in some scenes.The main content and research work are as follows.For the internal configuration composition and external structure of the twowheel differential unmanned vehicle system,analyze the ROS system and navigation framework.Design the overall scheme of the navigation software and hardware system,and adopt a layered approach to divide the system into three layers.Establish coordinate system model,map model,and sensor model respectively.On this basis,the kinematics model of the two-wheel differential unmanned vehicle is established to provide a modeling basis for the follow-up research content.Based on this model,a dynamic variable sampling area RRT(Dynamic variable sampling area RRT,DVSA-RRT)path planning algorithm is proposed for the traditional RRT(rapidly exploring random tree)algorithm node search randomness and poor scene adaptation.Divide the sampling area according to the sampling area formula,and then use the collision detection with a reserved safety distance,the probability target bias strategy and the multi-level step expansion for global path planning.Finally,the initial path is optimized by using the reverse optimization considering the maximum rotation angle constraint and the B-spline curve,so as to improve the node search efficiency and path feasibility.A Pose Auxiliary Point TEB(PAP-TEB)algorithm is proposed based on the global path planning results for the problem of low passage rate of unmanned vehicles in narrow spaces.By combining the fixed-point planning algorithm to improve the passage rate in narrow space environment,the output speed of autonomous navigation and fixed-point navigation is optimized by using velocity interpolation algorithm and 7-segment S-type velocity planning algorithm.Combining the research results of global path planning and local path planning for unmanned vehicles,we build a navigation system for two-wheeled differential speed unmanned vehicles.Analyze the experimental platform and the cartographer algorithm used in the mapping and positioning,and design the navigation framework.The feasibility and effectiveness of DVSA-RRT algorithm and PAP-TEB algorithm are verified by simulation.Compared with similar algorithms to prove the effectiveness of the two improved algorithms.Finally,based on the two-wheel differential unmanned vehicle autonomous navigation platform,the real vehicle verification is carried out in different environments.The autonomous navigation of the unmanned vehicle is realized,and the feasibility and robustness of the navigation system are verified.