Inference of Human-derived Specifications of Object Placement via Demonstration
This addresses the challenge for robots in tasks like packing and sorting to interpret human spatial rules, representing an incremental improvement in expressiveness over existing methods.
The paper tackles the problem of robots understanding human-acceptable object configurations by introducing PARCC, a formal logic framework for describing spatial relationships, and an inference algorithm to learn these specifications from demonstrations, with results from a human study showing it captures human intent better than human-provided specifications.
As robots' manipulation capabilities improve for pick-and-place tasks (e.g., object packing, sorting, and kitting), methods focused on understanding human-acceptable object configurations remain limited expressively with regard to capturing spatial relationships important to humans. To advance robotic understanding of human rules for object arrangement, we introduce positionally-augmented RCC (PARCC), a formal logic framework based on region connection calculus (RCC) for describing the relative position of objects in space. Additionally, we introduce an inference algorithm for learning PARCC specifications via demonstrations. Finally, we present the results from a human study, which demonstrate our framework's ability to capture a human's intended specification and the benefits of learning from demonstration approaches over human-provided specifications.