Symbolic Manipulation Planning with Discovered Object and Relational Predicates

• DOI: 10.48550/arxiv.2401.01123
• Publication Date: 2024-01-01
• Authors: Alper Ahmetoğlu; Erhan Öztop; Emre Uğur
• Keywords: Affordance; ENCODE

Discovering the symbols and rules that can be used in long-horizon planning from a robot's unsupervised exploration of its environment continuous sensorimotor experience is challenging task. The previous studies proposed learning single or paired object interactions with these symbols. In this work, we propose system learns discovered relational encode an arbitrary number objects relations between them, converts those to Planning Domain Description Language (PDDL), generates plans involve affordances achieve tasks. We validated our box-shaped different sizes showed develop symbolic knowledge pick-up, carry, place operations, taking into account compounds configurations, such as boxes would carried together larger box they are placed on. also compared method state-of-the-art methods operators defined over gives better performance baselines.