An adaptive feedforward control method for under-actuated bipedal walking on the compliant Ground

Motivated by the potential use of humanoid robot in real environment, an adaptive feedforward control strategy is developed to stabilize the underactuated bipedal walking on the compliant ground. First, the robot-ground coupling dynamic system is modelled as a rigid kinematical chain coupled with a spring-damper system. Then by observing the human’s gait, we find the walking speed has a direct effect upon the walking stability. In consideration of the highly complicated impact of real road surface on direct walking speed control, through analysis on laws governing the robot’s walking speed and the centre-of-mass (CM) motion, we establish a parameterized equivalent rod-ground coupling model based on the robot’s actual state, and identify the mapping relation between its walking speed and the CM forward moving distance of a full walking cycle (xf) adaptively. Finally, the robot’s walking is stabilized through a feedforward control over xf . The availability and adaptability of this method were validated through simulations: specific to one initial gait and three compliant conditions with different damping parameters, the walking was stabilized and the rate of stable convergence improved; specific to three initial gaits and three compliant conditions with stochastic varying damping parameters, the walking was still stabilized and its performance also improved.