A Parametric Model for Near-Optimal Online Synthesis with Robust Reach-Avoid Guarantees

2025-04-01

AI Summary

Existing online controller synthesis methods for high-integrity cyber-physical systems lack formally verifiable robustness guarantees. Method: This paper proposes a pretraining-free, real-time deterministic online synthesis framework. It introduces a novel parametric modeling and synthesis mechanism that integrates discretized game-theoretic reasoning with dynamic programming, incorporating hybrid game automata modeling, range-adaptive discrete dynamic programming, look-ahead shielding, and mode-based synthesis. Contribution/Results: The framework synthesizes controllers that are near-optimal while ensuring formal reach-avoid robustness. Evaluated on autonomous aerial vehicle simulations, the synthesized controllers satisfy safety-critical real-time constraints (millisecond-scale response), admit formal robustness verification, and significantly outperform black-box approaches—such as reinforcement learning—in interpretability, verifiability, and reliability.

Technology Category

Multiagent Systems: Adversarial AgentsPlanning, Routing, and Scheduling: Mixed Discrete/Continuous PlanningHumans and AI: Human-Aware Planning and Behavior Prediction

Application Category

Responsible Web: Machine-in-the-loop, human agency and autonomyEconomics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applicationsSystems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Objective: To obtain explainable guarantees in the online synthesis of optimal controllers for high-integrity cyber-physical systems, we re-investigate the use of exhaustive search as an alternative to reinforcement learning. Approach: We model an application scenario as a hybrid game automaton, enabling the synthesis of robustly correct and near-optimal controllers online without prior training. For modal synthesis, we employ discretised games solved via scope-adaptive and step-pre-shielded discrete dynamic programming. Evaluation: In a simulation-based experiment, we apply our approach to an autonomous aerial vehicle scenario. Contribution: We propose a parametric system model and a parametric online synthesis.

Problem

• Develop near-optimal online controllers for cyber-physical systems
• Ensure robust reach-avoid guarantees without prior training
• Use hybrid game automata and dynamic programming

Innovation

• Hybrid game automaton models application scenarios
• Scope-adaptive dynamic programming ensures robustness
• Parametric online synthesis without prior training