Event-Based Neural Networks Adaptive Control of Nonlinear Systems: A Fully Actuated System Approach

The event-triggered neural network (NN) adaptive control problems based on the fully actuated system (FAS) approach are studied for uncertain strict-feedback nonlinear systems. Firstly, event-based NN are utilized to approximate the unknown system nonlinearities, and the assumption that nonlinearities are known at all times is removed in the existing literature. Different from the backstepping design approach that the virtual control signals are non-differentiable at each triggering instant, this problem is avoided by utilizing the FAS approach. Then, an event-triggered NN adaptive controller that only receives states and using parameter estimations at each triggering instant is developed by using the FAS approach. To stabilize the control system, the adaptive parameters, the NN weights estimations, and Lyapunov solutions are used to design a novel adaptive event-triggering scheme (ETS), which can compensate the effect of triggering and save communication resources. It is proven that the ultimate boundedness of the system is guaranteed and the Zeno behavior can be eliminated. Finally, the effectiveness of the proposed method is illustrated by two simulation examples.