Seminar Series

Unmanned aerial vehicles – Online parameter identification of an adaptive minimal model for flight control

Speaker

Ph.D. Candidate Daniel Morris

Institute

Department of Mechanical Engineering, University of Canterbury

Time & Place

Tue, 11 May 2021 11:20:00 NZST in E13, Engineering Core

All are welcome

Abstract

Unmanned aerial vehicles (UAVs) have seen a significant increase in use by both the civilian and private sectors over the last decade. They are often used in remote sensing applications, where speed and cost are of concern. However most modern UAVs use closed-loop flight controllers that do not include modelling of the UAV’s flight dynamics. This limits their use to calm, open-air conditions.
There are a number of applications for UAVs that do not exhibit such limited conditions. These include structural inspection, urban search and rescue and precision agriculture, where a UAV may experience wind and weather effects, turbulence and near-field effects from proximity to structures, and complex payloads. By considering the known dynamics of UAV airframes, we can formulate model-aware flight controllers that are capable of operating in these more challenging conditions.
As real-time computation of the fluid dynamics experienced by a UAV are not feasible, we instead use a simplified model of the UAV’s flight dynamics. To compensate for the limitations of this minimal model, we use parameter estimation to determine the modelled and un-modelled dynamics in flight. This approach also lets us model external disturbances caused by complex operating conditions.

Daniel Morris

Supervisor: Associate Prof Chris Pretty