Learning Quadrotor Dynamics Using Neural Network for Flight Control

TL;DR

This paper explores using neural networks to learn quadrotor dynamics from specific trajectories and then generalize control to different trajectories, demonstrating effective control of yaw and position in experiments.

Contribution

It introduces a neural network-based dynamics model trained on separate translational and rotational trajectories to control complex quadrotor motions.

Findings

Neural network model successfully generalizes to new trajectories.

Effective control of yaw and position demonstrated in experiments.

Model outperforms traditional methods in trajectory generalization.

Abstract

Traditional learning approaches proposed for controlling quadrotors or helicopters have focused on improving performance for specific trajectories by iteratively improving upon a nominal controller, for example learning from demonstrations, iterative learning, and reinforcement learning. In these schemes, however, it is not clear how the information gathered from the training trajectories can be used to synthesize controllers for more general trajectories. Recently, the efficacy of deep learning in inferring helicopter dynamics has been shown. Motivated by the generalization capability of deep learning, this paper investigates whether a neural network based dynamics model can be employed to synthesize control for trajectories different than those used for training. To test this, we learn a quadrotor dynamics model using only translational and only rotational training trajectories, each of which can be controlled independently, and then use it to simultaneously control the yaw and position of a quadrotor, which is non-trivial because of nonlinear couplings between the two motions. We validate our approach in experiments on a quadrotor testbed.