Humanoid Robot Balancing

Abstract

There have been many researches about humanoid robot motion control, for example, walking pattern generation (Huang et al, 2001) (Kajita et al, 2002, 2003), walking control (Choi et al, 2006) (Grizzle et al, 2003) (Hirai et al, 1998) (Kajita et al, 2001) (Kim & Oh, 2004), (Lohmeier et al, 2004) (Park, 2001) (Takanishi et al, 1990) (Westervelt et al, 2003), running control (Nagasaki et al, 2004, 2005), balancing control (Kajita et al, 2001) and whole body coordination (Choi et al, 2007) (Kajita et al, 2003) (Sentis & Khatib, 2005) (Goswami & Kallem, 2004) (Harada et al, 2003) (Sugihara & Nakamura, 2002). Especially, the whole body coordination algorithm with good performance becomes a core part in the development of humanoid robot because it is able to offer the enhanced stability and flexibility to the humanoid motion planning. In this chapter, we explain the kinematic resolution method of CoM Jacobian with embedded motion which was suggested in (Choi et al, 2007), actually, which offers the ability of balancing to humanoid robot. For example, if humanoid robot stretches two arms forward, then the position of CoM(center of mass) of humanoid robot moves forward and its ZMP(zero moment point) swings back and forth. In this case, the proposed kinematic resolution method of CoM Jacobian with embedded motion offers the joint configurations of supporting limb(s) calculated automatically to maintain the position of CoM fixed at one point. Also, a design of balancing controller with good performance becomes another important part in development of humanoid robot. In balancing control, the ZMP control is the most important factor in implementing stable bipedal robot motions. If the ZMP is located in the region of supporting sole, then the robot will not fall down during motions. In order to compensate the error between the desired and actual ZMP, various ZMP control methods have been suggested; for example, direct/indirect ZMP control methods (Choi et al, 2004) (Kajita et al, 2003) and the impedance control (Park, 2001). Despite many references to bipedal balancing control methods, research on the stability of bipedal balancing controllers is still lacking. The exponential stability of periodic walking motion was partially proved for a planar bipedal robot in (Grizzle et al, 2003) (Westervelt et al, 2003). Also, the ISS (disturbance input-to-state stability) of the indirect ZMP controller was proved for the simplified bipedal robot model in (Choi et al, 2004, 2007). In this chapter, we will explain the balancing control method and its ISS proof which were suggested in (Choi et al, 2007). The O pe n A cc es s D at ab as e w w w .ite ch on lin e. co m