Research on Autonomous Navigation of Two Wheel-leg Robot in Indoor Dynamic Environment

Abstract

The object of this paper is the autonomous navigation control system of the two wheel-leg robot in indoor dynamic environment. To give full play to the kinematic ability of the two wheel-leg robot, the terrain adaptation ability of the leg structure and the control ability of the upper layer autonomous navigation system, the multi-mode control method of the robot and the path planning and autonomous navigation in indoor environment are studied in this paper. Firstly, based on the simulation physics platform, the kinematic transfer relationship of distributed model was proposed, the kinematic model of the wheel-leg subsystem and the two wheel-leg system was established, and the full-drive dynamic model with the wheel-leg-torso interaction force as the end output force was established. Then, a distributed control framework with torso postures as task space was proposed to plan the torso force and joint torque of the body in layers. The proposed walking and jumping motion planner of the two wheel-leg system and the multi-mode state machine are integrated with the control frame to realize the wheeled walking and jumping smooth control of the sagittal plane wheel-leg coordination. Finally, the ROS software framework based RGB-D visual sensor SLAM local environment awareness system is used to convert the point cloud map into an octree map, and then the oblique projection principle is converted into a two-dimensional raster map. A * algorithm and TEB algorithm are respectively used for global and local path planning. Finally, the autonomous navigation ability of the two wheel-leg robot in indoor dynamic environment is realized.