Lexical learning from interaction data for service robots

2015 Fiscal Year Final Research Report

• Project/Area Number: 25730111
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing
• Research Institution: Nagoya Institute of Technology
• Principal Investigator: Taguchi Ryo 名古屋工業大学, 工学(系)研究科(研究院), 助教 (70508415)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 音声情報処理 / 言語獲得 / シンボルグラウンディング / ロボット / 発話理解

Outline of Final Research Achievements

In order to understand speech of a user and act adequately, service robots need to have the pre-designed linguistic knowledge, such as lexicon, grammar and the relationship between the meaning of speech and actions of robots. However, the robot cannot adequately act when a user uttered the term that is not included in the pre-designed knowledge. To solve this problem, in this research, we developed the method to extend the linguistic knowledge by statistically analyzing phoneme sequences of the new wordings. On the other hand, utterances for navigation often includes relative spatial concepts. However, reference objects which are required to understanding relative spatial concepts may not be included in the utterances. In this research, we proposed a method for learning relative spatial concepts while estimating the reference objects. In this method, a robot extracts candidates of reference object from an occupancy grid map.

Free Research Field

情報工学