Rakuten Mobile, AWL and Vissel Kobe Selected for MIC Project Using Edge AI to Improve Safety in Large-scale Facilities

Tokyo, June 24, 2025 – Rakuten Mobile, AWL, Inc. and Rakuten Vissel Kobe today announced they have been selected for the Ministry of Internal Affairs and Communications’ "Regional Digital Transformation (DX) Promotion Package Project (AI Verification Type)," in an initiative managed by Mitsubishi Research Institute^*1. This month the companies launched a joint demonstration in Kobe City focused on using Edge AI technology to optimize communication loads in large-scale venues^*2.

The initiative integrates AWL's intelligent camera solutions, Rakuten Mobile's advanced network infrastructure and innovative AI algorithms at events held at Noevir Stadium Kobe, home of Vissel Kobe, the Rakuten Group's professional soccer club.

Project overview
The security sector is currently facing significant challenges related to labor shortages and increasing personnel expenses. Digital transformation and AI-driven automation offer promising solutions to these issues. However, the deployment of cloud-based AI for monitoring and analysis in large-scale environments, such as stadiums, can strain existing communication networks due to the high volume of video data transmission, leading to increased operational costs and potential infrastructure limitations.

This collaborative project seeks to mitigate these challenges by validating a solution that utilizes Edge AI to optimize anomaly detection and dynamically manage communication load from stadium surveillance cameras. The project also explores the potential of the RAN Intelligent Controller (RIC) platform – an AI-powered management and control system for the Radio Access Network (RAN) – to further enhance network performance and scalability.

Key project areas:

• Demonstration 1: Communication load reduction
  Far-edge devices^*3 deployed at surveillance camera locations will perform real-time risk assessment and analyze video feeds using AI-powered image recognition. These devices will then dynamically compress video data into three levels (text, image and video clip) based on both the identified risk level and real-time network congestion levels received from edge servers.
  
  Edge servers connecting multiple far-edge devices will aggregate and summarize the surveillance data using a small-scale language model (SLM) to generate actionable reports for security personnel. This approach aims to optimize mobile network performance in bandwidth-constrained environments while streamlining security operations and paving the way for RIC platform integration for enhanced RAN management.
• Demonstration 2: AI versatility and scalability
  The project will evaluate the performance of image recognition AI in identifying events such as falls and aggressive behavior, ensuring its robustness and reliability in various environmental conditions. The goal is to develop a versatile AI solution that can be readily deployed and scaled to meet the evolving security needs of different venues.
• Demonstration 3: Portable and scalable security solutions
  Looking ahead, the project aims to develop highly portable and environment-agnostic security solutions that minimize the impact on existing communication infrastructure. By combining portable camera systems with integrated AI processing and mobile network technologies, the project seeks to create versatile security solutions suitable for a wide range of applications, including large-scale event venues, amusement parks, transportation hubs, retail stores and even sparsely populated areas. This will enable cost-effective and scalable security deployments in diverse environments, leveraging the power of mobile networks and Edge AI to enhance security and operational efficiency.

Project timeline:
June 2025 – January 2026 (planned)

Rakuten Mobile, AWL and Vissel Kobe are committed to contributing to safer and more secure communities throughout Japan through this innovative project.

Notes:
*1 For details on the selection results, please refer to the following page on the Ministry of Internal Affairs and Communications website: "Selection Results for the AI Verification Type of the Digital Transformation (DX) Promotion Package for Regional Societies Project"
https://www.soumu.go.jp/menu_news/s-news/01ryutsu06_02000432.html (*Japanese page)
*2 A technology that utilizes AI in edge computing. By performing AI processing on edge devices such as sensors and cameras, real-time processing and communication volume reduction can be expected.
*3 A type of edge device in edge computing. It refers to a solution that performs computational processing close to the endpoints, such as individual equipment.

*Product names and service names published in this press release are generally trademarks or registered trademarks of their respective companies. Trademarks such as TM and ® may be omitted when trademarks are listed in this press release.