Balance Control Research For NAO Humanoid Robot On Standing State

The biped walking robot is a kind of intelligent robot which uses human as a template. Since the biped robot can adapt to the human normal living environment, it has many advantages than other kinds of robots. But how to keep the balance of the biped robot is a very complicated problem, because the robot system is nonlinear, high degree of freedom, coupling joints, and only small support area can be provided. The walking plan and the gait control are the main study aspects in the past, but in this paper we mainly study the standing robot and this paper focuses on the strategies to realize balance control of Nao robot.In this paper the Nao robot is the study object. On the basis of the predecessors’research results, the ankle joint strategy, hip joint strategy and step strategy are applied to realize the robot balance control according to the size of the external force successfully. In the ankle strategy, this paper derived the stable region using liner inverted pendulum model and realized the balance sustain and state recovery by controlling the torque of ankle joint. In the hip strategy, this paper derived the stable region using the inverted pendulum with a flywheel model. And in the realizing process, the balance sustained and state recovered successfully according the RNS (Reaction Null Space) theory. Furthermore, the leg lift strategy realized the robot balance control by combining the ankle strategy and hip strategy. In the step strategy, we calculated the step starting condition using single inverted model and ankle joint torque control. In the realizing process, the Capture point is calculated by combining single inverted pendulum model, double support inverted pendulum model and inverted pendulum with a flywheel model, then the robot taken a step by inverse kinematics method. At last, this paper realized the robot walking plan according the Capture point theory. These three balance control strategies mimic the situation that human deal with the external forces, and in theory the robot can resist any size of external disturbance force through the combination of these three control strategies.The creative works in this paper are applying the three balance strategies to the Nao robot successfully by reasonable stiffness control; analyzing the balance areas of the three different strategies which including the selfish balance areas; proposing some improvements which specific to the Nao robot characteristics during the balance control process. Finally, when exposed to different external forces, the balance control strategies are successfully verified by experiments on the Nao robot.