KNODE-MPC: A Knowledge-based Data-driven Predictive Control Framework for Aerial Robots

TL;DR

This paper introduces KNODE-MPC, a hybrid modeling framework combining physics-based models and neural networks for improved predictive control of aerial robots, demonstrating significant accuracy and performance gains.

Contribution

The work presents a novel hybrid modeling approach using knowledge-based neural ODEs integrated into MPC for quadrotor control, outperforming existing Gaussian Process models.

Findings

Hybrid model outperforms Gaussian Process in prediction accuracy.

KNODE-MPC achieves over 60% improvement in simulation.

Over 21% enhancement in real-world trajectory tracking.

Abstract

In this work, we consider the problem of deriving and incorporating accurate dynamic models for model predictive control (MPC) with an application to quadrotor control. MPC relies on precise dynamic models to achieve the desired closed-loop performance. However, the presence of uncertainties in complex systems and the environments they operate in poses a challenge in obtaining sufficiently accurate representations of the system dynamics. In this work, we make use of a deep learning tool, knowledge-based neural ordinary differential equations (KNODE), to augment a model obtained from first principles. The resulting hybrid model encompasses both a nominal first-principle model and a neural network learnt from simulated or real-world experimental data. Using a quadrotor, we benchmark our hybrid model against a state-of-the-art Gaussian Process (GP) model and show that the hybrid model provides more accurate predictions of the quadrotor dynamics and is able to generalize beyond the training data. To improve closed-loop performance, the hybrid model is integrated into a novel MPC framework, known as KNODE-MPC. Results show that the integrated framework achieves 60.2% improvement in simulations and more than 21% in physical experiments, in terms of trajectory tracking performance.