Research On Real-time Hybrid Path Planning Algorithm In Dynamic Environment

Smart civil aviation is the main direction of current civil aviation construction and development,and the research and application of intelligent equipment is an important measure to build smart civil aviation.Based on the actual needs of civil aviation environment work,this research group has developed a set of intelligent vehicle general platform,which carries different equipment and can be applied to different civil aviation work scenarios to achieve fuel distribution,airport cruise,locomotive assistance,and other work.The core technology of the vehicle platform includes 3D scene restoration,"master slave relationship" following algorithm,and real-time path planning.In this thesis,a real-time hybrid path planning algorithm for vehicle platforms in dynamic environments is studied,taking locomotive maintenance applications as an example.Based on the dynamic characteristics of the locomotive maintenance environment and the particularity of equipment,this algorithm designs an improved hybrid path planning method based on A ~* algorithm and dynamic window method,achieving safe and reliable global path planning,tracking,and real-time obstacle avoidance,effectively solving the real-time mobility challenges of intelligent vehicle platforms.The algorithm can be appropriately revised to apply to other civil aviation work scenarios.In this thesis,we first adapt the traditional path planning algorithm to the environment of the maintenance site,analyze the shortcomings of the original algorithm,and make targeted improvements.The specific work is as follows: To address the problems of traditional A ~*global path planning algorithms in locomotive maintenance environments such as poor safety,multiple path inflection points,node redundancy,close proximity to obstacles,and large turning angles,this thesis improves the traditional A ~* algorithm,designs a safety factor embedded in the cost function of the A ~* algorithm,optimizes the selection of path key points,designs an adaptive adjustment step size algorithm,and smoothes paths based on Bezier curves,The above problems have been effectively solved.In addition,in order to solve the problems of traditional dynamic window algorithms such as hesitating in the face of U-shaped obstacles in unknown complex environments and being prone to falling into local optimization,this thesis improves and optimizes the basic structure of dynamic window obstacle avoidance and three evaluation sub functions of the objective function,which to some extent alleviates the interference of Ushaped obstacles to the algorithm.Based on the above two improved algorithms,this thesis ultimately proposes a real-time hybrid path planning algorithm for adaptive environments.By improving the cost function of the A ~* algorithm,it improves the security of the algorithm,reduces inflection points,and optimizes the running trajectory;The dynamic window method utilizes the prior information of the global path to enable the intelligent vehicle platform to reach each sub target point,and realtime avoid dynamic obstacles,completing path planning in the maintenance environment.This method not only solves the disadvantage that the A ~* algorithm cannot avoid dynamic obstacles,but also avoids the intelligent vehicle platform falling into local optimization.Finally,a simulation analysis of the hybrid algorithm based on Matlab shows that the proposed algorithm is effective.Finally,the intelligent vehicle platform equipped with ROS system is verified in a simulated locomotive environment,and the parameters of the hybrid path planning algorithm are modified using data to adapt to the actual environment.Experimental results show that the proposed algorithm can meet the requirements in both static and dynamic environments.