Developing real-time system identification for UAVs

Abstract

This work is based on adapting suitable techniques in real-time aircraft system identification for the use in rigid body dynamic modelling of Unmanned Aerial Vehicles (UAVs). In general due to their reduced size UAVs lend themselves well to the recovery of flight data in full scale atmospheric trials. Timely recovery of the true model parameters is necessary to minimise flight tests for rapid prototype development. Using the Cranfield University Jetstream-31 Flying Laboratory (G-NFLA) as a test bed an investigation into suitable parameter update methods applicable for UAV purposes is being performed; and builds on previous work which considered the constraints pertinent to small UAVs, such as the absence of air flow vanes. This paper outlines an approach to achieve post-manoeuvre parameter estimation by applying the equation-error method in the frequency domain, and an example using flight data for the Jestream-31 flying laboratory is presented.