Proprioceptive control of a hybrid legged-wheeled robot

In this work we describe an innovative proprioceptive control architecture for our hybrid legged-wheeled robot ASGUARD. This robot is able to cope with a variety of stairs, very rough terrain, and is able to move with the speed of two body-lengths per second on flat ground. An additional proprioceptive inclination feedback is used to make the same controller more robust in terms of stair-climbing capabilities. Contrary to existing approaches, we did not use a pre-defined walking pattern for stair climbing, but an adaptive approach based only on internal sensor information. The data we use in our architecture is based on proprioceptive information, like body inclination and external torques, which are acting on the driving motors. In this work we show how this adaptivity results in a versatile controller for hybrid legged-wheeled robots. For the locomotion control we use an adaptive model of motion pattern generators. In contrast to many other walking pattern based robots, we use the direct proprioceptive feedback in order to modify the internal control loop, thus adapting the compliance of each leg on-line. For different terrains and stairs we use a phase-adaptive pattern which is using directly the proprioceptive data from each leg. We show that our adaptive controller is able to improve the stair-climbing behaviour in terms of energy consumption and energy distribution.