A High-Gain Approach To Event-Triggered Controller Design

In this thesis,a high-gain observer approach is applied to event-triggered control for a class of nonlinear continuous-time systems.An event-triggered high-gain observer(HGO)is introduced by designing an event-triggering condition,and the performance of the proposed observer is investigated.Comparative experiments are carried out on a DC torque motor system to validate the theoretic results.Considering that packet dropout randomly occurs when transmitting signals,a compensation mechanism is proposed to reduce the influence of the packet dropout.Simulations are carried out to validate the theoretic analysis.The main contributions in this thesis are summarized as follows:1)An event-triggered HGO is introduced,and an event-triggering mechanism which only relies on the measurable signals of the controlled system is designed for HGO.With the proposed event-triggering mechanism,the performance of the event-triggered HGO is investigated.The upper bounds of the observation error and the system states in steady-state are provided,and their relationships with the parameters in the event-triggering condition are developed.2)Comparative experiments are carried out on a DC torque motor platform to validate the obtained theoretic results.The experimental results show that for different parameter choices in the event-triggering condition,satisfactory control performances(in comparison with time-triggered HGO)can be achieved at reduced average sampling rate.3)The problem that packet dropout may randomly occur when transmitting signals is investigated.The mathematic model of packet dropout is established,and a compensation mechanism is proposed to decrease the influence of the packet dropout.The effectiveness of the proposed mechansim is validated by simulation results.