First Translation from LTL to LTLf+ for Infinite Trace Objectives

Researchers have introduced the first translation from Linear Temporal Logic (LTL) to LTLf+, a logic that lifts finite-trace LTL to infinite traces while preserving LTL's expressive power. LTL is widely used in AI for specifying temporal objectives in reactive synthesis, stochastic planning, and reinforcement learning. Traditionally, solving these problems requires translating LTL to nondeterministic automata on infinite words and then determinizing them, a step that is notoriously difficult. LTLf+ retains the advantages of LTLf, including canonical minimal representation and efficient determinization for finite automata on finite words. The translation normalizes the LTL specification and converts it into LTLf+, enabling more efficient reasoning. The work is available on arXiv under the identifier 2608.02454.