Research On Path Planning Method Of Quadruped Robot Based On Elevation Map

Quadruped robots have become a hot research topic in both academia and industry due to their excellent traversability and ability to complete complex tasks.They have gradually moved from laboratory theoretical research towards practical applications in production and daily life,and are expected to play an irreplaceable role.Unlike wheeled robots that can only move in two-dimensional planes,quadruped robots have the ability to move vertically,such as climbing stairs and overcoming obstacles.However,due to the high complexity of three-dimensional environments,existing path planning techniques for quadruped robots suffer from poor path quality and long planning times.Therefore,it is necessary to study path planning techniques specifically for quadruped robots.This thesis proposes a path planning algorithm based on elevation grid maps that combines the flexible motion capabilities of quadruped robots.The algorithm consists of two parts:global path planning and local trajectory planning.Firstly,in the global path planning part,the path planning task in the elevation grid map is implemented based on the RRT-Connect algorithm by combining the motion characteristics of the quadruped robot;meanwhile,the algorithm is optimized in terms of path quality and planning efficiency:the quality of the generated path is optimized by introducing the consideration of the robot yaw angle and adding the joint judgment of the yaw angle in the cost function;the quality of the generated path is optimized by using the heuristic random state search,heuristic control input calculation and adaptive extension step length are used to achieve the improvement of planning efficiency.Finally,a fast and efficient path planning algorithm is achieved,and a safe and stable path is generated that is more consistent with the actual situation of quadruped robot motion.Secondly,in the local trajectory planning part,the control variables of the robot are solved using the TEB algorithm.By using the elevation grid map,the traditional TEB algorithm that relies on two-dimensional cost maps is extended to three-dimensional space,enabling the local trajectory planning algorithm to achieve functions such as climbing and overcoming obstacles.By using graph optimization,the robot’s pose state and the time interval between adjacent states are used as nodes,and global paths,obstacles,terrain,and robot velocity and acceleration constraints are used as edges to construct a hypergraph.Finally,the g2o optimization framework is used to solve the problem,achieving more comprehensive and flexible local trajectory planning for quadruped robots.Finally,this thesis validates the rationality and effectiveness of the proposed path planning method through simulation experiments and real robot platforms.The experimental results show that the proposed path planning method can efficiently generate smooth paths that satisfy the kinematic and dynamic constraints of the robot,and complete path planning tasks in actual scenarios.