Mobile Robot Path Planning Based On Continuous Dynamic Movement Primitives

As an important research field of robotics,path planning has always been one of the main research contents of motion planning.The common mobile robot path planning methods mostly establish the objective function for the motion constraints and environmental constraints of the robot ontology,and use the mathematical optimization method to search and solve the optimal path.The calculation process is often very complicated,and each task needs to be recalculated.Based on the idea of demonstration,learning and generalization,this thesis combines the actual motion environment and task scene to complete the path planning of the robot: firstly,obtain relatively optimal robot motion samples and use the algorithm to learn,and then generalize the motion state of the new task target,at the same time avoiding static or dynamic obstacles that may be encountered.In this thesis,the path planning of mobile robot based on continuous dynamic movement primitives is researched as follows:Firstly,the research frontier of the traditional path planning algorithm is summarized.Based on the idea of demonstration learning,the kernel of this thesis,the dynamic movement primitives,is introduced.And the basic principle of the method,generalization,dimension promotion,improvement and obstacle avoidance are introduced.On this basis,the robot's planning task is completed by establishing the movement primitive library in the global static environment,and the continuous dynamic movement primitives are proposed to complete the local trajectory modification of the robot under dynamic scenes and tasks.Then,the relevant algorithms in this thesis are verified by simulation experiments.The experiments on the basic characteristics of dynamic movement primitives,the matching generalization of movement primitives library,and the response mechanism of continuous dynamic movement primitives to dynamic scenes are proved.Simulation experiments verify the feasibility and effectiveness of the proposed algorithms.Finally,the mobile robot path planning and its learning platform are built.The software and hardware components and system principles of the designed system are expounded.The cross-platform dynamic movement primitives path planning and learning experiment of the mobile robot is completed.The reliability and practicability of the system are proved.