Temporally extended goal recognition in fully observable non-deterministic domain models
Ramon Fraga Pereira, Francesco Fuggitti, Felipe Rech Meneguzzi, Giuseppe De Giacomo · Applied Intelligence · 2023
Abstract Goal Recognitionis the task of discerning the intended goal that an agent aims to achieve, given a set of goal hypotheses, a domain model, and a sequence of observations (i.e., a sample of the plan executed in the environment). Existing approaches assume that goal hypotheses comprise a single conjunctive formula over a single final state and that the environment dynamics are deterministic, preventing the recognition of temporally extended goals in more complex settings. In this paper, we expand goal recognition totemporally extended goalsinFully Observable Non-Deterministic(fond) planning domain models, focusing on goals on finite traces expressed inLinear Temporal Logic(ltl $$_f$$ f ) andPure-Past Linear Temporal Logic(ppltl). We develop the first approach capable of recognizing goals in such settings and evaluate it using differentltl $$_f$$ f andppltlgoals over sixfondplanning domain models. Empirical results show that our approach is accurate in recognizing temporally extended goals in different recognition settings.