ARX-Based Model Identification in Hovering Flight of Micro Air Vehicle

This paper focuses on investigating the problem of the longitudinal (pitch) and lateral (roll) attitude dynamic model for a Micro Air Vehicle with small fixed wing in the hovering mode. A small flight control system with multi-channel data logger units and data acquisition software, which consists of a ground subsystem and an onboard subsystem, is designed to control the aircraft flying and to collect the flying attitude information. Some hovering flight experiments were performed to pick up the inputs and outputs of the aircraft, and then to identify the models directly from these measurements using time series analysis method. Two fourth-order ARX models were finally obtained from a lot of candidate models to describe the attitude dynamics of the longitudinal (pitch) and lateral (roll) control channels, respectively.