Energy-based Regularization for Learning Residual Dynamics in Neural MPC for Omnidirectional Aerial Robots

2026-04-16

AI Summary

This work addresses the lack of physical consistency—such as inertia and energy conservation—in learned residual dynamics models commonly employed in neural model predictive control (MPC), which can compromise control stability. To mitigate this issue, the authors propose an energy-conserving regularization method that incorporates system-energy-based constraints during the training of deep neural networks, thereby enhancing the physical plausibility of the residual model. The regularized model is integrated into an MPC framework for trajectory tracking of an omnidirectional aerial robot. Experimental results demonstrate that the proposed approach reduces the mean absolute position error (MAE) by 23% on average across three real-world flight tasks compared to analytical MPC, and further decreases MAE by up to 15% relative to unregularized neural MPC, while significantly improving flight stability.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Natural Language Processing: Safety and Robustness
• Machine Learning: Adversarial Learning & Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Data-driven Model Predictive Control (MPC) has lately been the core research subject in the field of control theory. The combination of an optimal control framework with deep learning paradigms opens up the possibility to accurately track control tasks without the need for complex analytical models. However, the system dynamics are often nuanced and the neural model lacks the potential to understand physical properties such as inertia and conservation of energy. In this work, we propose a novel energy-based regularization loss function which is applied to the training of a neural model that learns the residual dynamics of an omnidirectional aerial robot. Our energy-based regularization encourages the neural network to cause control corrections that stabilize the energy of the system. The residual dynamics are integrated into the MPC framework and improve the positional mean absolute error (MAE) over three real-world experiments by 23% compared to an analytical MPC. We also compare our method to a standard neural MPC implementation without regularization and primarily achieve a significantly increased flight stability implicitly due to the energy regularization and up to 15% lower MAE. Our code is available under: https://github.com/johanneskbl/jsk_aerial_robot/tree/develop/neural_MPC.

Problem

• Neural MPC
• residual dynamics
• energy conservation
• omnidirectional aerial robots
• flight stability

Innovation

• energy-based regularization
• neural MPC
• residual dynamics
• omnidirectional aerial robots
• model predictive control