Improving object reference recognition through human-robot alignment

2016 Fiscal Year Final Research Report

• Project/Area Number: 15K16075
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Intelligent robotics
• Research Institution: Advanced Telecommunications Research Institute International
• Principal Investigator: Iio Takamasa 株式会社国際電気通信基礎技術研究所, 知能ロボティクス研究所, 連携研究員 (70642958)
• Project Period (FY): 2015-04-01 – 2017-03-31
• Keywords: 知能ロボティクス / ヒューマンロボットインタラクション

Outline of Final Research Achievements

When referring to an object, a person sometimes use a reference with insufficient information for specifying the object. Such ambiguous references decrease a performance of object reference recognition.
In this research, I developed an algorithm of robot's confirmation behaviors to entrain humans to use enough information in their references based on human-robot alignment, which is a phenomenon that human behaviors are elicited to robot behavior, and implemented the algorithm into the object reference recognition system. Through an experiment, the proposed algorithm made the performance of object reference recognition better than basic one.

Free Research Field

知能ロボティクス