Servo control of unstable-wheeled system by using disturbance torque observer compensation and convex optimization

Presents the servo control for an unstable-wheeled system by using disturbance torque observer compensation and a controller designed by convex optimization. We consider that unstable behavior of the mono-wheeled system because of one supported point body is caused by the virtual disturbance torque. For this purpose, the paper firstly proposes a design method of the disturbance observer in feedback loop for such virtual disturbance torque. In the proposed disturbance observer, an ordinary 2nd-order low pass filter for noise attenuation is insufficient to stabilize the whole wheeled system, so a higher order filter is designed by the H_∞ loop shaping method from the view point of robust stability. Secondly, the servo controller for the closed loop system including such a disturbance observer is designed by the convex optimization algorithm in order to achieve multiple control specifications simultaneously. The validity of this control system is demonstrated through experiments