Study on spoken dialogue system with safety consideration based on automatic estimation of driving situation

• Project/Area Number: 26730092
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Perceptual information processing
• Research Institution: Okayama University
• Principal Investigator: Hara Sunao 岡山大学, 自然科学研究科, 助教 (50402467)
• Research Collaborator: ABE Masanobu ; YAMAOKA Masaki ; TOBA Shunji
• Project Period (FY): 2014-04-01 – 2017-03-31
• Project Status: Completed (Fiscal Year 2016)
• Budget Amount: ¥3,640,000 (Direct Cost: ¥2,800,000、Indirect Cost: ¥840,000); Fiscal Year 2016: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2015: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
• Keywords: 音声対話 / 情報検索 / 運転行動 / 生体信号 / スマートフォン / 高度道路交通システム（ITS）

Outline of Final Research Achievements

We conducted below:
(1) We used bio signals and driving information signals that are measured by the driver’s body and the vehicle’s body for estimating the driving road from the signals. Furthermore, we estimated driving load assuming use of sensors obtained by smartphones. (2) We evaluated the spoken dialog strategy from the viewpoint of user’s driving load. An objective evaluation by computer simulation was conducted by considering both dialog initiatives and the exitance of confirmation utterances. (3) We conducted a subjective evaluation to evaluate the performance of the proposed dialog strategy as a spoken dialog system. The subjects were asked to drive a simulator during talking with a spoken dialog system. (4) A dialog strategy based on graph search was introduced to realize a dialog strategy which considering the estimation result of the user's mental load. We evaluated the proposed system by objective evaluation and subjective evaluation.

Research Products

• [Presentation] スマートフォンで収録した環境音データベースを用いたCNNによる環境音分類 (2017)
  • Author(s): 鳥羽隼司，原直，阿部匡伸
  • Organizer: 日本音響学会2017年春季研究発表会
  • Place of Presentation: 明治大学生田キャンパス
  • Year and Date: 2017-03-15
• [Presentation] 重複音を含む環境音データベースにおける環境音検出のための特徴量の基本検討 (2016)
  • Author(s): 原直，田中智康，阿部匡伸
  • Organizer: 日本音響学会2016年秋季研究発表会
  • Place of Presentation: 富山大学
  • Year and Date: 2016-09-14
• [Presentation] RNNによる実環境音データからのマルチ音響イベント検出 (2016)
  • Author(s): 鳥羽隼司，原直，阿部匡伸
  • Organizer: 日本音響学会2016年秋季研究発表会
  • Place of Presentation: 富山大学
  • Year and Date: 2016-09-14
• [Presentation] A Spoken Dialog System with Redundant Response to Prevent User Misunderstanding (2015)
  • Author(s): Masaki Yamaoka, Sunao Hara, Masanobu Abe
  • Organizer: APSIPA Annual Summit and Conference 2015
  • Place of Presentation: Hong kong
  • Year and Date: 2015-12-16
  • Int'l Joint Research
• [Presentation] 冗長なシステム応答を用いたユーザの誤認識に頑健な音声対話システムに関する検討 (2015)
  • Author(s): 山岡将綺，原直，阿部匡伸
  • Organizer: 2015年日本音響学会秋季研究発表会
  • Place of Presentation: 会津大学
  • Year and Date: 2015-09-16
• [Presentation] 車載用音声対話システムにおけるユーザ負荷を考慮した対話戦略の検討 (2014)
  • Author(s): 山岡将綺，原直，阿部匡伸
  • Organizer: 情報処理学会音声言語処理研究会
  • Place of Presentation: 東京工業大学（東京都・目黒区）
  • Year and Date: 2014-05-22 – 2014-05-23
• [Remarks] ホームページ等
  • URL: http://www.a.cs.okayama-u.ac.jp/~hara/projects/kaken2014.html
• [Remarks] 自動車運転状況の自動推定に基づく安全性に配慮した音声対話システムの研究 - 岡山大学・原
  • URL: http://www.a.cs.okayama-u.ac.jp/~hara/projects/kaken2014.html