Towards Safe Learning-Based Non-Linear Model Predictive Control through Recurrent Neural Network Modeling
This work addresses computational limitations in deploying NMPC on embedded hardware, offering a more efficient and safer learning-based approximation for control systems.
The paper tackles the high computational cost of nonlinear model predictive control (NMPC) by proposing Sequential-AMPC, a sequential neural policy that reduces the need for expert data and improves safety, achieving higher feasibility rates and better learning dynamics compared to a baseline.
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.