Aggressive Online Control of a Quadrotor via Deep Network Representations of Optimality Principles

TL;DR

This paper introduces G&CNets, a deep neural network approach that enables aggressive, real-time quadrotor control by imitating optimal control solutions, significantly reducing computation time and improving maneuver speed.

Contribution

The work presents a novel deep learning method for on-board optimal control of quadrotors, bridging the reality gap and outperforming traditional differential-flatness-based methods.

Findings

G&CNets achieve faster trajectory execution in experiments.

The method effectively closes the reality gap between simulation and real-world control.

Deep networks can imitate optimal control to enable aggressive quadrotor maneuvers.

Abstract

Optimal control holds great potential to improve a variety of robotic applications. The application of optimal control on-board limited platforms has been severely hindered by the large computational requirements of current state of the art implementations. In this work, we make use of a deep neural network to directly map the robot states to control actions. The network is trained offline to imitate the optimal control computed by a time consuming direct nonlinear method. A mixture of time optimality and power optimality is considered with a continuation parameter used to select the predominance of each objective. We apply our networks (termed G\&CNets) to aggressive quadrotor control, first in simulation and then in the real world. We give insight into the factors that influence the `reality gap' between the quadrotor model used by the offline optimal control method and the real quadrotor. Furthermore, we explain how we set up the model and the control structure on-board of the real quadrotor to successfully close this gap and perform time-optimal maneuvers in the real world. Finally, G\&CNet's performance is compared to state-of-the-art differential-flatness-based optimal control methods. We show, in the experiments, that G\&CNets lead to significantly faster trajectory execution due to, in part, the less restrictive nature of the allowed state-to-input mappings.