산업 응용 분야의 N 자유도 유압식 매니퓰레이터를 위한 고등 제어 전략

Abstract

This thesis presents advanced control strategies so as to address existing problems of an n-DOF serial hydraulic manipulator, subjected to matched and mismatched disturbances and faults identification for industrial applications. Due to the high power-to-weight ratio, the electro-hydraulic actuator and electro-hydraulic system (EHS) are reasonably employed to drive the manipulator link frame and enhance system performance. However, the EHS is a highly nonlinear system that cannot be neglected like electric actuators when developing control strategies. Besides, the influence of mismatched uncertainties from external impacts and matched uncertainty available existing in the EHS challenges the control development. Therefore, in order to address those problems, the proposed control scheme based on the adaptive backstepping sliding mode control (ABSMC) is designed to achieve system stability and robustness against perturbations. Based on the ABSMC technique, a disturbance estimator is also involved to estimate and compensate for the matched/mismatched disturbances and observe unmeasured variables. Besides, intelligent techniques are also used to satisfy the different working operations. Both comparative simulations and experiment between the proposed control scheme and other conventional algorithms are verified on a 3-DOF serial hydraulic manipulator, subjected to the payload variations in free motion, to evaluate the effectiveness of the proposed methodology. Subsequently, based on the obtained results, sensor fault and safe operation are then considered as the main objectives to develop the control strategy for the n-DOF serial hydraulic manipulator in constrained frameworks. During this process, position-based impedance control (PB-IC) is preferred as the key point to achieve force regulation and system stability under the influence of the sensor fault. This fault type is generated from the additive unknown faulty signals, which may be a constant offset or time-variant variable, to the sensors. Due to not obvious effect caused by the sensor faults, the system behavior is the same as the fault-free condition; however, the controller executes erroneous control action on the actuator. Therefore, a robust fault estimation based on extended state observer (ESO) is proposed to estimate the faulty signals to achieve the desired actuation. With the estimated signals, the active fault-tolerant control (FTC) based on the ABSMC technique is constructed. Furthermore, the behavior of the manipulator may be unstable or exhibit unexpected motion in a case of sudden contact-loss or environment step-change. Therefore, a modified PB-IC based on virtual energy tank and passivity control concepts is deployed to decouple the force control action when sudden contact-loss happens. The closed-loop system stability is theoretically proven by the Lyapunov theorem. Additionally, comparative simulations are given to verify the effectiveness of the proposed algorithm.