MotiveMob: Motivation as Semantic Action for Closed-Loop Human Mobility Generation
MotiveMob generates human mobility trajectories using a 'why–where–when' decision process driven by motivation. It uses two components, a Motivation Predictor (PM) and a State Predictor (PS), to first predict the next move's intent and then ground that intent in a feasible location and time. This framework outperforms existing methods by learning from both observed movements and generated alternative transitions.