Optimization-Based Wheeled-Bipedal Robot-Graduation Project
- Version one, video
- Simplify the two-wheeled foot model to an inverted pendulum model and apply Newton’s Euler equation to obtain the state-space equations of the system.
- Control with Linear Quadratic Regulator (LQR)
- Used Virtual Model Control (VMC) to obtain a mapping of forces and moments in the workspace to moments in the joint space.
- Version two
- Neglect the robot leg dynamics and use Single Rigid Body Dynamics (SRBD) to obtain the state space equations of the system.
- Using convex model predictive control (Convex-MPC) to solve the optimal ground reaction force
- Using Whole Body Control (WBC) Based on Hierarchical Quadratic Optimization (HQP) to solve joint moments(unfinished)
- Motion state planning using finite state machine (FSM)
- Build the Webots simulation environment and the robot model on a Linux system based on the ROS platform.
