Generalizable and Fast Surrogates: Model Predictive Control of Articulated Soft Robots using Physics-Informed Neural Networks

2025-02-04

AI Summary

Soft robotic systems face a fundamental trade-off between prediction speed and generalizability in real-time control, particularly for pneumatic jointed soft robots where first-principles models are computationally expensive and data-driven surrogates require excessive real-world data. Method: This paper proposes a data-efficient, physics-informed neural network (PINN) surrogate model—first successfully applied to pneumatic jointed soft robots—trained on only two hours of single-domain experimental data. Contribution/Results: The PINN achieves 466× faster dynamics prediction than first-principles models and, when integrated with nonlinear model predictive control (MPC), delivers an average joint tracking error of just 1.3° across nine dynamic tasks, enabling real-time high-precision control. Crucially, the approach overcomes longstanding bottlenecks in applying PINNs to complex soft robotic systems, demonstrating strong cross-domain generalization and robustness to parametric variations—thereby drastically reducing reliance on large-scale experimental datasets.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• 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
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Soft robots can revolutionize several applications with high demands on dexterity and safety. When operating these systems, real-time estimation and control require fast and accurate models. However, prediction with first-principles (FP) models is slow, and learned black-box models have poor generalizability. Physics-informed machine learning offers excellent advantages here, but it is currently limited to simple, often simulated systems without considering changes after training. We propose physics-informed neural networks (PINNs) for articulated soft robots (ASRs) with a focus on data efficiency. The amount of expensive real-world training data is reduced to a minimum - one dataset in one system domain. Two hours of data in different domains are used for a comparison against two gold-standard approaches: In contrast to a recurrent neural network, the PINN provides a high generalizability. The prediction speed of an accurate FP model is improved with the PINN by up to a factor of 466 at slightly reduced accuracy. This enables nonlinear model predictive control (MPC) of the pneumatic ASR. In nine dynamic MPC experiments, an average joint-tracking error of 1.3{deg} is achieved.

Problem

• Real-time control of soft robots
• Data efficiency in machine learning
• Improving prediction speed and accuracy

Innovation

• Physics-informed neural networks (PINNs)
• Data efficiency in training
• Nonlinear model predictive control (MPC)