Interactive Robot
Research Team

Research Summary

This team aims at developing a social robot that can naturally interact with humans and provide casual and modest support for human daily activities. We explore a mechanism underlies natural interaction and robot behaviors that subconsciously change human behaviors.

Main Research Fields
Human-Robot Interaction
Keywords
  • Humanoid
  • Android robot
  • Humanlike expression
  • Natural interaction
  • Non-verbal interaction
  • Multimodal recognition and expression
Research theme
  • Modeling human behaviors in interaction and generating natural robot behavior
  • Recognizing a state of human activity based on multimodal information
  • Design of robot behaviors to lead humans’ subconscious behavioral change

Takashi Minato

Takashi Minato

History

2001
Japan Science and Technology Agency (JST)
2002
Osaka University
2006
Japan Science and Technology Agency (JST)
2011
Advanced Telecommunications Research Institute International (ATR)
2020
RIKEN

Award

2016
ATR Prize for excellent study
2020
ATR Prize for excellent study

Members

Carlos Toshinori Ishi
Senior Scientist
Chaoran Liu
Research Scientist
Kurima Sakai
Research Scientist
Tomo Funayama
Technical Staff I
Yuka Nakayama
Technical Staff I
Takamasa Iio
Visiting Scientist
Bowen Wu
Research Part-time Worker I and Student Trainee
Jiaqi Shi
Research Part-time Worker I and Student Trainee
Taiken Shintani
Research Part-time Worker I and Student Trainee
Houjian Guo
Research Part-time Worker II and Student Trainee
Shudai Deguchi
Research Part-time Worker II and Student Trainee
Niina Okuda
Research Part-time Worker II and Student Trainee
Namika Taya
Administrative Part-time Worker II and Student Trainee
Mayuka Yamasaki
Administrative Part-time Worker II and Student Trainee
Taichi Matsuoka
Administrative Part-time Worker II and Student Trainee
Minjeong Bae
Administrative Part-time Worker II and Student Trainee
Rin Takahira
Student Trainee
Simon Andreas Piorecki
Student Trainee
Nabeela Khanum Khan
Student Trainee

Former member

Ryusuke Mikata
Technical Staff I(2022/08-2023/07)
Changzeng Fu
Visiting Scientist(2022/05-2023/03)
Alexander Diel
Visiting Researcher(2022/07-2023/05)
Yuki Hori
Student Trainee (2020/08-2022/03)
Kakeru Sakuma
Research Intern(2023/09)

Research results

Upper body gestures

For the generation of natural robot movements in interaction, we are studying the generation of upper body gestures synchronized with the robot's voice. The Conditional GAN (Generative Adversarial Networks) model is trained based on the data of natural human gestures while speaking. This model can generate the body/hand gestures from the input speech signals. So far, we have implemented the model in a CG agent, and have shown that human-like gestures are generated through subject experiments.

Upper body gestures

Gaze control

We are studying gaze control in face-to-face interactions between robots and multiple people to generate natural robot motions. Based on the data from multi-person dialogues, we modeled occurrence probabilities of eye contact, looking away, and the direction of looking away depending on the role of the dialogue (speaker, main listener, sub-listener, etc.), and developed a gaze behavior model. We implemented this model to control the gaze of a small robot, and through subjective experiments, we have shown that the model generates more human-like behavior than conventional methods.

Gaze control

Selected Publications

  1. Alexander Diel, Wataru Sato, Chun-Ting Hsu, and Takashi Minato, Differences in configural processing for human versus android dynamic facial expressions, Scientific Reports, Vol.13, No. 16952, 2023.
  2. Carlos T. Ishi, Chaoran Liu, and Takashi Minato, An attention-based sound selective hearing support system: evaluation by subjects with age-related hearing loss, Proc. of IEEE/SICE International Symposium on System Integrations, 2023.
  3. Xinyue Li, Carlos T. Ishi, and Ryoko Hayashi, Prosodic and voice quality analyses of filled pauses in Japanese spontaneous conversation by Chinese learners and Japanese native speakers, Proc. of International Conference on Speech Prosody, 2022.
  4. Dongsheng Yang, Qianying Liu, Takashi Minato, Shushi Namba, and Shin'ya Nishida, Optimizing facial expressions of an android robot effectively: a Bayesian optimization approach, Proc. of IEEE-RAS International Conference on Humanoid Robots, 2022.
  5. Takashi Minato, Kurima Sakai, Takahisa Uchida, and Hiroshi Ishiguro, A study of interactive robot architecture through the practical implementation of conversational android, Frontiers in Robotics and AI, Vol.9, No.905030, 2022.
  6. Changzeng Fu, Chaoran Liu, Carlos T. Ishi, and Hiroshi Ishiguro, C-CycleTransGAN: A non-parallel controllable cross-gender voice conversion model with CycleGAN and Transformer, Proc. of Asia Pacific Signal and Information Processing Association Annual Summit and Conference, 2022.
  7. Taiken Shintani, Carlos T. Ishi, and Hiroshi Ishiguro, Expression of personality by gaze movements of an android robot in multi-party dialogues, Proc. of IEEE International Conference on Robot and Human Interactive Communication, 2022.
  8. Changzeng Fu, Chaoran Liu, Carlos T. Ishi, and Hiroshi Ishiguro, An adversarial training based speech emotion classifier with isolated Gaussian regularization, IEEE Transactions on Affective Computing, Vol.14, No.8, 2022.
  9. Bowen Wu, Jiaqi Shi, Chaoran Liu, Carlos T. Ishi, and Hiroshi Ishiguro, Controlling the impression of robots via GAN-based gesture generation, Proc. of IEEE/RSJ International Conference on Intelligent Robots and Systems, 2022.
  10. Wataru Sato, Shushi Namba, Dongsheng Yang, Shin'ya Nishida, Carlos T. Ishi, and Takashi Minato, An android for emotional interaction: Spatiotemporal validation of its facial expressions, Frontiers in Psychology, Vol.12, 2022.
  11. Jiaqi Shi, Chaoran Liu, Carlos T. Ishi, and Hiroshi Ishiguro, 3D skeletal movement-enhanced emotion recognition networks, APSIPA Transactions on Signal and Information Processing, Vol.10, No.E12, 2021.

Links

Contact Information

takashi.minato [at] riken.jp

Top