Research On Motion Planning Of Hexapod Robot Under The Influence Of Multidimensional Terrain Features

Complex outdoor environments pose certain challenges to the perception and planning of robots due to the unpredictable and non-structured terrain features,such as rugged and steep terrain geometries,and the coupled changes of physical terrain features such as soft and slippery surfaces.The soft sand and slippery mud in complex environments can cause the robot to sink or tip over,and the physical characteristics of such terrains cannot be described by traditional geometric models.Therefore,this dissertation studies the multidimensional feature modeling of terrain and the motion planning of hexapod robots based on the terrain model,aiming to improve the safety and stability of hexapod robots passing through complex terrain in the field.First,the terrain modeling part proposes a multidimensional terrain feature representation model for hexapod robots.An elevation map containing terrain geometric information is constructed,along with a semantic map based on semantic segmentation that contains terrain category information.This allows for a comprehensive analysis of the terrain’s multidimensional geometric and physical characteristics.Next,the motion planning part takes into account the impact of multidimensional terrain features on the traversal of the hexapod robot and constructs a traversal cost map of the terrain.The traditional A* algorithm is improved based on the cost of terrain crossing to improve path safety in complex terrain.Additionally,the footing cost map of the terrain is constructed,and a free gait planning method based on the footing cost is proposed to improve the robot’s trafficability and stability in wild terrain.Overall,this approach provides a comprehensive and effective method for motion planning in complex and varied terrain conditions.Finally,In the experimental verification section,a hexapod robot model and a complex terrain environment in the field were built using ROS and Gazebo.Using this setup,experiments were carried out on map construction,path planning,and gait planning.The results of the experiments demonstrated the effectiveness of the map construction method.The improved A* algorithm,based on traversal cost,was found to have better path security compared to the traditional A* algorithm.Additionally,the free gait planning algorithm was found to have better trafficability and stability compared to the traditional free gait algorithm.Finally,a real machine platform was built to further verify the effectiveness of the map construction,path planning,and gait planning methods presented in this dissertation.