Good-for-Games (GFG) automata constitute a sound alternative to determinism as a way to model specifications in the Church synthesis problem. Typically, inputs for the synthesis problem are in the form of LTL formulas. However, the only known examples where GFG automata present an exponential gap in succinctness compared to deterministic ones are not LTL-definable. We show that GFG automata still enjoy exponential succinctness for LTL-definable languages. We introduce a class of properties called "eventually safe" together with a specification language E νTL for this class. We finally give an algorithm to produce a Good-for-Games automaton from any E νTL formula, thereby allowing synthesis for eventually safe properties.

The design of functionally correct autonomous systems which operate in an unknown environment and that satisfy reliability, availability, maintainability, and safety (RAMS) requirements is a challenge. In this paper we focus on the detection and reconfiguration features these systems must provide. Indeed, evolving in an unknown environment can invalidate the assumptions made during the design phase. In particular, different hardware components might fail and provide erroneous inputs to the system, which will pass in a degraded mode where the expected RAMS do not hold anymore. Such faults need to be detected as early as possible and reconfiguration strategies must be applied to bring the system back into a nominal mode where the RAMS are satisfied. We propose an automated design process based on formal methods to develop Fault Detection, Isolation and Recovery (FDIR) components targeting partially observable timed systems.We describe how to automatically synthesize runtime monitors, design reconfiguration strategies, and obtain full-fledged FDIR components. We illustrate the approach on a case study inspired from autonomous robotics applications.

No abstract

No abstract