Waseda University Researchers Develop AI for Enhanced 5G Multicast Reliability

Researchers at Waseda University (https://www.waseda.jp/top/en/news/89346) have introduced a lightweight artificial intelligence (AI) methodology designed to improve the reliability of 5G multicast broadcasting. The approach aims to mitigate temporary video freezing, which often results from data loss during wireless transmission.

The study, spearheaded by Professor Jiro Katto and PhD candidate Kasidis Arunruangsirilert, focuses on using AI to predict wireless network conditions proactively. This predictive capability allows for real-time adjustments to transmission settings, potentially leading to more consistent video delivery.

Waseda University: Study Details and Performance Metrics

The AI model underwent training using an extensive dataset comprising approximately 26 million measurements. These data points were collected from a commercial 5G network at 0.5-millisecond intervals.

Key findings from the research indicate a significant improvement in transmission reliability:
* The AI-driven approach accurately selected error-free transmission settings for approximately 87% of video segments. This contrasts with a 32% success rate observed when using a conventional speed-oriented transmission method.
* The model demonstrated efficient processing, operating in under 0.07 milliseconds on smartphone chipsets released from 2020 onward. This performance metric suggests its feasibility for real-time implementation on consumer devices.
* Researchers anticipate that this methodology could facilitate more dependable high-definition television delivery over local 5G networks, particularly in environments where the installation of fiber-optic cables is logistically challenging or cost-prohibitive.

Broader Implications for AI-Native Networks

Professor Jiro Katto characterized the research as a practical demonstration of integrating AI into future wireless networks. “Our study stands as a practical example of AI-native wireless communication, in which AI takes on decision-making responsibilities in next-generation networks, and represents a meaningful step toward the long-sought convergence of broadcasting and broadband on a single, spectrally efficient wireless platform,” Katto stated.

Beyond conventional television broadcasting, the lightweight AI approach holds potential applications across various sectors, including satellite communications, autonomous vehicles, industrial control systems, and scientific exploration. The integration of AI in such critical infrastructure highlights evolving [industry advancements in wireless technology (https://technosports.co.in/).

The research findings were formally presented at the 2026 IEEE 104th Vehicular Technology Conference (VTC2026-Fall), held in Boston.