Dynamic Event-Triggered Hᵢ/H∞ Optimization Approach to Fault Detection for Unmanned Aerial Vehicles

Abstract

This paper dealt with the problem of dynamic event-triggered fault detection (FD) for unmanned aerial vehicles (UAVs) in the Hi/H∞ optimization criterion. In order to make use of limited communication resources, a dynamic event-triggered mechanism (ETM) was considered to determine whether system information will be transmitted to the ground station. Note that the transmitted data failing to satisfy a predefined triggered condition were discarded directly. As such, FD performance was affected by not only disturbances and faults, but also dynamic event-triggered transmission errors, namely the errors between the data of non-event time and the current actual system. To solve this problem, a new dynamic event-triggered Hi/H∞ optimization approach was presented, whose contribution lied in the complete decoupling of residuals and event transmission errors under dynamic ETM, which can avoid continuous communication and Zeno phenomenon. Firstly, a dynamic ETM was presented to determine communication time. Then, a variable sampling period model of UAV was established based on the dynamic ETM. On this basis, a dynamic event-triggered residual generator was constructed. In the Hi/H∞ optimization criterion, Riccati recursion was used to compute the optimum solution of a dynamic event-triggered FD filter. Moreover, a new residual evaluation method was proposed, whose effectiveness and feasibility were finally verified by the altitude control system of a UAV.