DCM based architectures for humanoid robot walking on inclined planes

The walking control of a Humanoid robot is still a challenging task due to the non-linear dynamics and the high number of degrees of freedom. A Humanoid robot is equipped with legs that give the robot the ability to avoid obstacles that the wheels cannot overcome and the adaptation to human environments can be achieved straight forwards. Walking trajectories planner, online walking control and sensor measurements are the pieces that make up the mosaic. Under the idea provide before, the thesis has the aim to make a Humanoid robot ables to walk on inclined planes using novel control architectures based on Divergent Component of Motion (DCM). First, an essential literature review has been conducted to create the basis of the work. After, the specific problem of walking on inclined plane was investigated, understanding the main features which allow the robot to perform the task. Then, properly changes were carried out to the DCM control architectures such as the trajectories planner for the feet and the DCM and Walking controllers. The DCM architectures, designed with the planner and the controllers, have the same planner but different controller strategies, position and torque control. In the thesis is shown the differences between them in walking on an inclined plane. In the end, the DCM architectures have been tested in simulation and in the experiment which validated the approach