An Event-Triggered Control Approach with Single-Critic Architecture for a Class of Nonlinear Systems

Efficient control scheme in many large-scale or long-term industrial production systems is always the core research issue for engineers and scholars. Optimal control design, which can stabilize the system dynamic, and optimize the performance index simultaneously, provides an architecture to control the systems with the minimized energy consumption. This paper proposes a novel intelligent control scheme with an event-triggered single-critic architecture, which converts the optimal nonlinear control problem into solving the well-known Hamilton-Jacobi-Bellman (HJB) equation in systems. Compared with existing methods, the event-triggered mechanism is utilized to overcome the crucial and strict requirements of the system states in real time, and the single-critic architecture simplifies the conventional structure by eliminating the actor neural network. Both the stability and convergence of the proposed algorithm have been completely proved, where the cases that event is not triggered or is triggered are considered. Finally, the effectiveness of the novel event based single-critic control approach has been demonstrated by the simulation results.