POPL 2027
Sun 10 - Sat 16 January 2027 Mexico City, Mexico

Temporal reasoning in discrete time commonly decomposes a temporal formula into what must hold now and what must hold next, forming the basis of tableau constructions and automata-theoretic methods for logics such as LTL. However, this decomposition does not directly extend to temporal logics interpreted over continuous-time signals, such as Signal Temporal Logic (STL), which is widely used to specify properties of continuous and hybrid dynamical systems. To address this challenge, we introduce \emph{interval predicates}, which express the truth of STL formulas over time intervals rather than at individual instants. They enable interval-based reasoning: temporal operators unfold across a finite partition of a signal, one interval at a time. We formalize this unfolding as expansion rules, establish their correctness, and construct finite tableaux for STL formulas over finite signals. We translate such a tableau into a hybrid automaton, enabling robust STL model checking of hybrid systems through reachability analysis. The translation is independent of any particular reachability backend; experiments with four off-the-shelf analyzers on benchmarks with linear and nonlinear dynamics show substantial gains over state-of-the-art SMT-based robust STL model checking.