Complex Controller Synthesis Framework for Autonomous Robots in Unknown Environments

In recent years, advances in control theory and artificial intelligence (AI) have driven significant progress in autonomous mobile robots. However, challenges remain in navi-gating in unknown environments, handling unreliable task plans, and implementing complex actions, which affect system reliability and flexibility. This paper addresses these challenges by proposing an automated and unified control program synthesis framework that combines formal synthesis with machine learning based techniques. Using GR(1) reactive synthesis, which generates high-level task planning with guaranteed safety, and learning based techniques, which enable adaptive low-level control, this unified framework generates reliable and adaptive control programs for autonomous robots operating in unknown environments with complex actions.