European Investor
26 Feb 2026, 13:37
The University of Tokyo, NTT and NEC have jointly demonstrated real-time augmented reality (AR) assistance by integrating three newly proposed technologies on a 6G/IOWN platform, aimed at enabling the widespread use of AI agents for safety and security applications.
The collaboration combines streaming semantic communication technology (University of Tokyo), AI-oriented media control technology (NEC) and In-Network Computing (INC) architecture technology (NTT). Together, these technologies optimize high-volume data transmission and computational processing while reducing end-to-end latency for continuously operating AI agents.
In trials using AR glasses and video data simulating critical situations, the integrated system significantly reduced communication traffic and computational load, maintaining stable end-to-end latency without degrading AI inference accuracy. The results address key infrastructure challenges such as limited wireless bandwidth, rising computational demand and increased power consumption associated with large-scale AI models.
The initiative has been selected for exhibition at the Japan Pavilion at Mobile World Congress 2026. The partners plan to accelerate research and development toward the social implementation of AI agents and next-generation ICT infrastructure that enhance real-time response capabilities in disaster prevention, accident response and cybersecurity.
The collaboration combines streaming semantic communication technology (University of Tokyo), AI-oriented media control technology (NEC) and In-Network Computing (INC) architecture technology (NTT). Together, these technologies optimize high-volume data transmission and computational processing while reducing end-to-end latency for continuously operating AI agents.
In trials using AR glasses and video data simulating critical situations, the integrated system significantly reduced communication traffic and computational load, maintaining stable end-to-end latency without degrading AI inference accuracy. The results address key infrastructure challenges such as limited wireless bandwidth, rising computational demand and increased power consumption associated with large-scale AI models.
The initiative has been selected for exhibition at the Japan Pavilion at Mobile World Congress 2026. The partners plan to accelerate research and development toward the social implementation of AI agents and next-generation ICT infrastructure that enhance real-time response capabilities in disaster prevention, accident response and cybersecurity.