Various bipedal robots have been developed. What has been lacking, however, are bipedal robots that can perform "non-stationary motion". "Non-stationary motion" means sitting down, standing up, moving limbs, and so on. We have constructed a legged robot that can perform these movements without using the ankle-torque. Though it is natural to use ankle-torque for non-stationary motions, the usable amount of torque is restricted. We have to consider this input constraint when we design a control scheme to use ankle-torque for non-stationary motions. Since it is difficult to design a control scheme to use ankle-torque for general non-stationary motion, we focus on standing upright (a special non-stationary motion). Concretely, we gave an invariant set where ankle-torque is effective, and designed a hybrid control scheme that alternates between nonlinear and LQ control schemes on the surface of the invariant set. We also discovered stability of the controlled system with constrained ankle-torque. Moreover, simulations demonstrated the effectiveness of the control scheme. We expect that it can also be applied to non-stationary motion in general.
[1]
Masayuki Fujita,et al.
Analysis of conditions for non-violation of constraints on linear discrete-time systems with exogenous inputs
,
1997,
Proceedings of the 36th IEEE Conference on Decision and Control.
[2]
K. T. Tan,et al.
Linear systems with state and control constraints: the theory and application of maximal output admissible sets
,
1991
.
[3]
P. R. Bélanger,et al.
Piecewise-linear LQ control for systems with input constraints
,
1994,
Autom..
[4]
Koichi Osuka,et al.
Motion Control of Legged Robot Emu via Nonlinear Gain Scheduling Control
,
1998
.
[5]
Franco Blanchini.
Feedback control for linear time-invariant systems with state and control bounds in the presence of disturbances
,
1990
.