Towards Safe Learning-Based Non-Linear Model Predictive Control through Recurrent Neural Network Modeling

2026-03-25

AI Summary

This work addresses the challenge of deploying nonlinear model predictive control (NMPC) on embedded platforms, where its online computational complexity is prohibitive, and existing learning-based approximation methods suffer from high training costs and heavy reliance on expert demonstration data. To overcome these limitations, the authors propose Sequential-AMPC, a novel approach employing a recurrent neural policy with shared parameters to generate candidate control sequences, thereby substantially reducing dependence on expert trajectories. The method further incorporates a safety-enhanced online evaluation and fallback mechanism to ensure closed-loop safety. Compared to feedforward baselines, Sequential-AMPC achieves faster training in high-dimensional systems, requires significantly less expert data, yields more feasible candidate sequences, and demonstrates consistently improved and stable validation performance.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Multiagent Systems: Multiagent Learning

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The practical deployment of nonlinear model predictive control (NMPC) is often limited by online computation: solving a nonlinear program at high control rates can be expensive on embedded hardware, especially when models are complex or horizons are long. Learning-based NMPC approximations shift this computation offline but typically demand large expert datasets and costly training. We propose Sequential-AMPC, a sequential neural policy that generates MPC candidate control sequences by sharing parameters across the prediction horizon. For deployment, we wrap the policy in a safety-augmented online evaluation and fallback mechanism, yielding Safe Sequential-AMPC. Compared to a naive feedforward policy baseline across several benchmarks, Sequential-AMPC requires substantially fewer expert MPC rollouts and yields candidate sequences with higher feasibility rates and improved closed-loop safety. On high-dimensional systems, it also exhibits better learning dynamics and performance in fewer epochs while maintaining stable validation improvement where the feedforward baseline can stagnate.

Problem

• nonlinear model predictive control
• online computation
• learning-based control
• expert dataset
• embedded hardware

Innovation

• Sequential-AMPC
• safe learning-based control
• recurrent neural network modeling
• nonlinear model predictive control
• parameter sharing across horizon