Tracking Control for Robot Manipulator Based on Deterministic Learning and Event-Triggered Mechanism

Abstract

In this paper, a control scheme is presented for robotic manipulator by combining deterministic learning and event- triggered mechanism. Notice that the traditional adaptive neural control needs to re-adapt neural network (NN) weights even for performing the same control task, thereby result in the limited learning ability. To enhance the neural leaning ability, a dynamic learning controller is firstly constructed based on the radial basis function (RBF) NN, which has the ability to learn the knowledge of the unknown closed-loop system dynamics. By reusing the learned knowledge, the dynamic learning-based event-triggered control scheme is put forward to achieve a tradeoff between the network resource and the tracking performance. Specially, an easy-to-implemented event-triggered condition is designed by the Lyapunov technique due to the use of experience knowledge. The proposed scheme ensures that the tracking error converges to a small neighborhood of the origin, all the signals in the closed- loop system are bounded and meanwhile the communication resources are greatly reduced. A comparison simulation example is conducted to demonstrate the effectiveness and advantage of the proposed control scheme.