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

arXiv: https://arxiv.org/abs/2603.24503 · HTML: https://mlsanity.com/paper/2603.24503/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 19 references that are in the map.
- First posted on arXiv 25 Mar 2026.

## Builds on (most cited first, 11 of 11)

- [Efficient representation and approximation of model predictive control laws via deep learning](https://mlsanity.com/paper/1806.10644/) (2018)
- [Learning an Approximate Model Predictive Controller with Guarantees](https://mlsanity.com/paper/1806.04167/) (2018)
- [Large Scale Model Predictive Control with Neural Networks and Primal Active Sets](https://mlsanity.com/paper/1910.10835/) (2019)
- [Learning for MPC with Stability & Safety Guarantees](https://mlsanity.com/paper/2012.07369/) (2020)
- [On Recurrent Neural Networks for learning-based control: recent results and ideas for future developments](https://mlsanity.com/paper/2111.13557/) (2021)
- [Approximate non-linear model predictive control with safety-augmented neural networks](https://mlsanity.com/paper/2304.09575/) (2023)
- [Learning-based MPC from Big Data Using Reinforcement Learning](https://mlsanity.com/paper/2301.01667/) (2023)
- [Approximate Predictive Control Barrier Functions using Neural Networks: A Computationally Cheap and Permissive Safety Filter](https://mlsanity.com/paper/2211.15104/) (2022)
- [Imitation Learning of MPC with Neural Networks: Error Guarantees and Sparsification](https://mlsanity.com/paper/2501.03671/) (2025)
- [Model predictive control design for dynamical systems learned by Long Short-Term Memory Networks](https://mlsanity.com/paper/1910.04024/) (2019)
- [Recurrent Model Predictive Control: Learning an Explicit Recurrent Controller for Nonlinear Systems](https://mlsanity.com/paper/2102.10289/) (2021)

## Abstract (arXiv)

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.