Graph Neural Model Predictive Control for High-Dimensional Systems

2026-02-19

AI Summary

This work addresses the challenge of balancing modeling complexity and computational efficiency in real-time control of high-dimensional systems, such as soft robots. The authors propose a novel framework that integrates a graph neural network (GNN)-based dynamics model with a structure-aware model predictive controller. By leveraging graph representations to preserve the sparsity of local interactions and introducing a state compression algorithm with linear complexity—accelerated via GPU parallelization—the method achieves 100 Hz closed-loop real-time control on systems with up to a thousand nodes for the first time. Experiments demonstrate that the approach enables whole-body obstacle avoidance and high-precision trajectory tracking in both simulation and physical platforms, achieving sub-centimeter accuracy in hardware tests and outperforming baseline methods by 63.6%.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Hardware-aware ML

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The control of high-dimensional systems, such as soft robots, requires models that faithfully capture complex dynamics while remaining computationally tractable. This work presents a framework that integrates Graph Neural Network (GNN)-based dynamics models with structure-exploiting Model Predictive Control to enable real-time control of high-dimensional systems. By representing the system as a graph with localized interactions, the GNN preserves sparsity, while a tailored condensing algorithm eliminates state variables from the control problem, ensuring efficient computation. The complexity of our condensing algorithm scales linearly with the number of system nodes, and leverages Graphics Processing Unit (GPU) parallelization to achieve real-time performance. The proposed approach is validated in simulation and experimentally on a physical soft robotic trunk. Results show that our method scales to systems with up to 1,000 nodes at 100 Hz in closed-loop, and demonstrates real-time reference tracking on hardware with sub-centimeter accuracy, outperforming baselines by 63.6%. Finally, we show the capability of our method to achieve effective full-body obstacle avoidance.

Problem

• high-dimensional systems
• real-time control
• soft robots
• computational tractability
• complex dynamics

Innovation

• Graph Neural Networks
• Model Predictive Control
• High-Dimensional Systems
• Real-Time Control
• Soft Robotics