Abstract

Distant speech recognition technology for a robot was implemented to communicate with humans at a natural distance. There were two issues to address. First, surrounding reverberation and noise act to reduce recognition performance. In an attempt to resolve this, dereverberation and sound source identification technology were introduced in addition to the existing developed technology. Second, speech frequency response characteristics and wording differ from person to person. To enhance recognition performance, a deep neural network was used in the acoustic model, in addition to class-based N-grams accounting for the properties of speech recognition tasks in the language model. These technologies were then integrated into a robot system to perform a themed experimental demonstration taking reception calls.The demonstration achieved a reception success rate of 98.3% and, through analysis of the causes of erroneous speech recognition, clarified further issues to address for improving performance.

References

1. (1) Ministry of Economy: New Robot Strategy – Vision, Strategy, Action Plan -, (2015) (in Japanese)
2. (2) Okuno, H. G.: Robot Audition: Present and Future, Journal of the Robotics Society of Japan, Vol. 28, No. 1, p. 2-5, (2010) (in Japanese)
3. (3) Kawahara, T.: Recent Progress of Speech Recognition Technology, The journal of the Institute of Electrical Engineers of Japan, Vol. 133, No. 6, p. 364-367, (2013) (in Japanese)
4. (4) Delcroix, M., Yoshioka, T., Ogawa, A., Kubo, Y., Fujimoto, M., Ito, N., Kinoshita, K., Espi, M., Hori, T., Nakatani, T., Nakamura, A.: Linear Prediction-based Dereverberation with Advanced Speech Enhancement and Recognition Technologies for the Reverb Challenge, REVERB Workshop, (2014)
5. (5) Lee, A., Nakamura, K., Nishimura, R., Saruwatari, H., Shikano, K.: Noise Robust Real World Spoken Dialogue System Using GMM Based Rejection of Unintended Inputs, Proceedings of International Conference on Spoken Language Processing, Vol. 1, p. 173-176, (2004)
6. (6) Yoshioka, T., Ito, N., Delcroix, M., Ogawa, A., Kinoshita, K., Yu, M. F. C., Fabian, W. J., Espi, M., Higuchi, T., Araki, S., Nakatani, T.: The NTT CHiME-3 system: Advances in speech enhancement and recognition for mobile multi-microphone devices, IEEE ASRU, (2015)
7. (7) Shigemi, S., Kawabe, K., Nakamura, T.: Development of New ASIMO – Realization of Autonomous Machine -, Honda R&D Technical Review, Vol. 24, No. 1, p. 34-41
8. (8) Gomez, R., Kawahara, T.: Robust Speech Recognition Based on Dereverberation Parameter Optimization Using Acoustic Model Likelihood, IEEE Transactions on Audio, Speech, and Language Processing, Vol. 18, No. 7, p. 1708-1716, (2010)
9. (9) Shikano, K., Ito, K., Kawahara, T., Takeda, K., Yamamoto, M.: Onseininshiki Sisutemu, Ohmsha, 198p., (2001) (in Japanese)
10. (10) Hinton, G., Deng, L., Yu, D., Dahl, G., Mohamed, A., Jaitly, N., Senior, A., Vanhoucke, V., Nguyen, P., Sainath, T., Kingsbury, B.: Deep neural networks for acoustic modeling in speech recognition, IEEE Signal Processing Magazine, Vol. 29, No. 6, p. 82-97, (2012)
11. (11) Brown, P. F., Petra, V. J. D., de Souza, P. V., Lai, J. C., Mercer, R. L.: Class-Based n-gram Models of Natural Language, Computational Linguistics, Vol. 18, No. 4, p. 467-479, (1992)
12. (12) Hyumanoid Robotto EMIEW3 wo Katsuyoshita Jissho Jikken, Hitachi Hyoron, Vol. 98, No. 12, p. 742-743, (2016) (in Japanese)
13. (13) https://www.tokyo-airport-bldg.co.jp/hanedaroboticslab/
14. (14) Nakagawa, S., Takagi, H.: Statistical methods for comparing pattern recognition algorithms and comments on evaluating speech recognition performance, The Journal of the Acoustical Society of Japan, Vol. 50, No. 10, p. 849-854, (1994) (in Japanese)
15. (15) Nishimura, Y., Shinozaki, T., Iwano, K., Furui, S.: Shuhasutaiikigoto no Omomitsukiyudo wo Mochiita Onseininshiki no Kento, Acoustical Society of Japan 2004 Shunki Kenkyuhappyokai Kouenronbunshu, p. 117-118, (2004) (in Japanese)
16. (16) Povey, D., Ghoshal, A., Boulianne, G., Burget, L., Glembek, O., Goel, N., Hannemann, M., Motlicek, P., Qian, Y., Schwarz, P., Silovsky, J.: The Kaldi Speech Recognition Toolkit, IEEE 2011 Workshop on Automatic Speech Recognition and Understanding, p. 1-4, (2011)
17. (17) Okuno, H. G., Nakadai, K.: Robot Audition Open-Sourced Software HARK, Journal of the Robotics Society of Japan, Vol. 28, No. 1, p. 6-9, (2010) (in Japanese)

Author (organization or company)

• Shunichi YAMAMOTO（R&D Center X）
• Naoaki SUMIDA（Honda Research Institute Japan Co., Ltd.）
• Kazuhiro NAKADAI（Honda Research Institute Japan Co., Ltd.）

Publication Information

Article of Honda R&D Technical Review Vol.29 No.2 (116 - 123)

PDF

Citation Information