No more Itaewon crowd crush disaster: Safety in dense crowds transformed by ‘KOSMOS’ service with AI reinforcement learning

No more Itaewon crowd crush disaster: Safety in dense crowds transformed by ‘KOSMOS’ service with AI reinforcement learning

“The 2022 Itaewon crushing incident raised public safety concerns. In response, we developed KOSMOS, an AI service addressing “”crowd turbulence””. It employs a reinforcement learning-based AI, designed to suggest routes for crowd dispersion and collision minimization. Our model’s effectiveness in suggesting safe, efficient routes was verified through a simulation that recreated the Itaewon incident using Seoul’s road data. The KOSMOS application assesses real-time crowd density and alerts users to potential accident sites, offering safe travel routes. This paper aims to prevent crushing incidents and protect lives via this AI model. Our goal is to apply this technology in various real-life environments, particularly in crowded places.”

  • Connector.

    Participant(s)

    Taegeonwoo Yeo
    Dohui Ryu

  • Connector.

    Country

    South Korea

  • Connector.

    Category

    Social sciences

Share this Project