Representing and Learning Human Intent and Affordances for Human-Robot Collaboration
Abstract
The objective of this research project was to develop an architecture for reliable early inference of human intent and accordances. This research is motivated by fixC;ndings in experimental psychology, which indicate that exploratory movements and generic actions can be used to infer intent and accordance, without actually observing humans perform the tasks of interest. We proposed to include three systems in our architecture:1. Human Subjects Studies (HSS): to collect motion capture data of human subjects' exploratory movements when they intend to perform specificc tasks, and generic movements not directly related to the tasks they intend to perform. This data was to be analyzed to extract concise and unique mappings from these movements to specixC;fic intent and accordances.2. Incremental and Interactive Learning (IIL): algorithms that use the learned mappings from exploratory movements and generic actions to intent and accordances, and enable robots to incrementally learn probabilistic models of domain objects and human behavior from sensor inputs. These probabilistic models were to be used to revise domain knowledge.3. Knowledge Representation and Reasoning (KRR): algorithms that couple declarative logic programming and probabilistic graphical models to represent and reason with different descriptions of knowledge and uncertainty. These algorithms were to reason with incomplete, generic domain knowledge to guide the analysis of motion capture data and sensor inputs. Robots equipped with such an architecture could oxB;er preemptive assistance to humans whose capabilities do not extend to the tasks they intend to perform. The project was thus to build capabilities for human-robot collaboration in critical domains such as disaster rescue, surveillance, and assistive care.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 12, 2023
- Accession Number
- AD1224960
Entities
People
- Mohan Sridharan
Organizations
- University of Auckland