Summary of the project

If a person's intentions can be estimated, the ultimate CA (cybernetic avatar)*1, which can be operated as desired, becomes possible. For estimation, not only internal brain activity but also brain surface information and interaction information with others are important clues. By 2050, we will realize the ultimate BMI-CA that can be operated as a person wishes.

Milestone by year 2030

[Realization of a remote mutual assistance society where people can overcome their disabilities and participate in social activities]

People with physical and brain limitations can independently expand their own physical, cognitive, and perceptual abilities using AI-assisted BMI-CA*2 that can decode words and actions in their minds with high accuracy, and can further expand these abilities by collaborating and cooperating with experience-sharing CAs of others with whom they agree. and participate in new cultural, artistic, sporting, and educational activities.

Milestone by year 2025

[Technological Transformation of Communicating Words and Actions Conceived in the Mind to Others]

The expansion of physical, cognitive, and perceptual abilities beyond one person's work capacity and the speed of voice communication can be realized by coordinating and collaborating AI-assisted BMI-CAs that decipher with high precision the words and actions that anyone can conceive in his or her head. In particular, if a person with a disability wishes to have a surgical operation, some functions of the AI-assisted BMI-CA can be extended beyond those of a single person, enabling a new way of life.

*1 Avatars, like teleoperated robots or characters in cyberspace, share sensory perceptions with humans and engage in social activities. In this project, avatars function as semi-autonomous agents, and when combined with AI technology, they are applications that extend human physical, cognitive, and perceptual capabilities.

*2 AI-assisted BMI-CA: Cybernetic Avatar (CA) that can decipher words and actions in the user's head with high precision according to the combination of different types of BMI by machine learning of AI. This project uses non-contact BMI to estimate behavior from environmental sensors, non-invasive BMI to decode words and behavior from EEG on the scalp surface, and invasive BMI to decode from brain information near the dura mater during surgical procedures.

R&D theme progress reports

• 1. Development of IoB Interfaces (IoB: Internet of Brains)
• 2. IoB Middleware
• 3. IoB Core technology
• 4. Study of issues common to the project and R&D for social implementation
• 5. Research and Development of Minimally Invasive BMI

R&D theme structure of the project

Leader's institution

Advanced Telecommunications Research Institute International ATR

R&D institutions

ARAYA Inc., Osaka University, Keio University, AIST, Sony Computer Science Laboratories, Inc., University of Tokyo, Tokyo Institute of Technology, Tokyo Metropolitan Institute of Medical Science, Waseda University, University of Melbourne

PDF Download

• Summary of the project (252KB)
• Progress Report (Download all) (779KB)

News & Events