Neural-based event-triggered observer design for adaptive sliding mode control of nonlinear networked control systems

This study investigates the application of sliding mode control within the context of networked control systems subject to both internal and external disturbances, employing an event-triggered mechanism that leverages neural networks and incorporates an adaptive control strategy. The networked control system is first modelled, and an event-triggered communication strategy based on neural networks is proposed, allowing the observer to selectively receive the latest sampled data. Next, a sliding mode observer is devised to track the sliding motion and error system, demonstrating system stability and robustness through linear matrix inequalities. In order to pledge the attainment of the sliding surface within a prescribed period, a dynamically adaptive sliding mode controller driven by event-based triggering is devised, proving the positivity of the lower bound of event-triggered intervals. Finally, simulations using a single-link mechanical arm model validate the superiority and effectiveness of the recommended approach.