Adaptive learning control for triggered switched systems based on unknown direction control gain function

This paper investigates event-triggered switched heuristic dynamic programming (HDP) control for switched systems based on an unknown direction control gain function (UDCGF). First, the Nussbaum gain function is introduced to describe the unknown control direction, which makes the control behaviour of switched systems more flexible. The proposed variable threshold event-triggering (VTET) condition can significantly reduce computation while maintaining a certain system performance. Then, a detailed Lyapunov stability analysis based on the new tight bound is given, which reflects the asymptotic stability of the switched systems under the proposed switched HDP control method. The system mode-matched action-critic neural networks are employed to approximate the optimal triggered control law and the optimal switching value function, respectively. Subsequently, the boundedness analysis of neural network (NN) weights and the convergence analysis of HDP algorithm are given. At last, a numerical example is provided to verify the effectiveness of the proposed method.