Accurate Trajectory Tracking for Multirotors with Real-Time Adaptation to System Changes Through DNN-MRFT

Abstract

High performance trajectory tracking for multirotor Unmanned Aerial Vehicles (UAVs) is a fast growing research area due to the increase in popularity and demand. In many applications, the multirotor UAV dynamics would change in-flight resulting in performance degradation, or even instability, such that the control system is required to adapt its parameters to the new dynamics. In this thesis, we investigated the application of Deep Neural Networks and the Modified Relay Feedback Test (DNN-MRFT) approach for real-time identification and tuning of controllers suitable for aggressive trajectory tracking in the presence of external wind. Achieving high performance trajectory tracking performance required adopting a feedback linearization technique along with additional feedforward terms. The thesis includes extensive investigation and analyzes of different position control configurations with the goal of maximizing the tracking performance in the presence of wind disturbance and system parameter changes. As a result of this study, we were able to systematically provide a trade-off between performance and robustness in the tuning. We also proved the effectiveness and applicability of our developed approach through a set of experiments where accurate trajectory tracking is maintained despite significant changes to the UAV aerodynamic characteristics and the application of external wind. We also demonstrated low discrepancy between simulation and experimental results which proves the potential of using the suggested approach for planning and fault detection tasks. The achieved tracking results on figure-eight trajectory is on par with the state-of-the-art in the field, while having the advantage of being real-time adaptive.

Date of Award	Jul 2021
Original language	American English