MRAC-LLM Toolbox: An interactive model reference adaptive control enhanced with large language models

This paper presents the MRAC-LLM Toolbox, a MATLAB-based software that integrates classical Model Reference Adaptive Control (MRAC) with Large Language Model (LLM) assistance for interactive control system design. The software enables users to specify performance requirements through a graphical user interface, which are translated into reference model configurations with LLM guidance. MRAC-LLM focuses exclusively on classical MRAC and supports simulation and real-time parameter adaptation within a master–slave framework. The LLM operates strictly at an advisory level, assisting with reference model selection and adaptation guidance without modifying the underlying MRAC control laws or their stability guarantees. The software features a modular architecture supporting simulation, visualization, and reporting, and is released as open-source under the MIT License at https://github.com/halilokur91/MRAC_LLM.