Control of a humanoid robot walking with dynamic balance

In this paper, we proposed a simple algorithm for controlling humanoid walking. We take the body as control target and the legs as control tool. The passive dynamic of robot is used as much as possible, and the legs will execute necessary actions for adjusting body states through sensors' feedback. The hip actuators are used to avoid body from falling down including position and orientation which is the key for dynamic balance; the knee actuators are used to generate necessary spacing from ground when leg swinging; the ankle actuators are used to control walking velocity. Our control algorithm is test by simulation, the test shown that the robot can follow a given velocity and its maximum walking velocity is near to human, the robot won't fall down or deviate its destination when suffering a relative large disturbance. So our control algorithm is efficient and can be used to control humanoid robot walking.