An adaptive composite control for a hydraulic actuator impedance system of legged robots

The hydraulic drive unit (HDU) is the actuator of a legged robot joint, whose control affects the robot's motion performance. Impedance control is one of the essential control methods for the legged robot, which can reduce the collision force in foot-ground contact and improve the robot's stability. However, hydraulic systems are strongly nonlinear and have time-varying parameters, particularly when the robot is in different environments. Therefore, the impedance control of the HDU should be adaptive to parameter and environmental changes. A novel adaptive control combining polynomial nonlinear extended state observer (PNLESO) and adaptive feedback linearization control (AFLC) for the HDU is designed in this study called the PNLESO-AFLC. The main aspects of this study are as follows: (i) A polynomial based error function is designed to form the PNLESO, which observes system disturbances and improves observer chattering; (ii) AFLC is designed to compensate for the PNLESO's observation error, improving adaptability to changing parameters and different working conditions. Finally, the effectiveness of the PNLESO−AFLC is verified by comparative experiments. Experimental results show that the PNLESO−AFLC has good adaptability to different working conditions and can significantly improve the impedance control performance of HDU.