Learning-Aided Warmstart of Model Predictive Control in Uncertain Fast-Changing Traffic

TL;DR

This paper introduces a learning-aided warmstart framework for Model Predictive Control in uncertain, fast-changing traffic, using neural networks and sampling to improve trajectory initialization and avoid local minima.

Contribution

It proposes a novel multimodal neural network predictor combined with sampling to enhance warmstarts in MPC under uncertainty and dynamic environments.

Findings

Improved initial trajectory guesses in traffic scenarios.

Enhanced convergence and safety in MPC.

Effective identification of multiple local minima.

Abstract

Model Predictive Control lacks the ability to escape local minima in nonconvex problems. Furthermore, in fast-changing, uncertain environments, the conventional warmstart, using the optimal trajectory from the last timestep, often falls short of providing an adequately close initial guess for the current optimal trajectory. This can potentially result in convergence failures and safety issues. Therefore, this paper proposes a framework for learning-aided warmstarts of Model Predictive Control algorithms. Our method leverages a neural network based multimodal predictor to generate multiple trajectory proposals for the autonomous vehicle, which are further refined by a sampling-based technique. This combined approach enables us to identify multiple distinct local minima and provide an improved initial guess. We validate our approach with Monte Carlo simulations of traffic scenarios.