본문 바로가기
  • Home

Development of a Multi-Agent Simulation Model for Generating Future Pedestrian Flows Based on Real-World CCTV Trajectory Data

  • Industry Promotion Research
  • Abbr : IPR
  • 2026, 11(4), pp.173~185
  • DOI : 10.21186/IPR.2026.11.4.173
  • Publisher : Industrial Promotion Institute
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Received : October 8, 2026
  • Accepted : October 29, 2016
  • Published : October 31, 2026

Jeong-Won Chu 1,  Ju-Han Park 1,  Young Wook Seo 2

1유클리드소프트
2대전대학교

Accredited

ABSTRACT

This study proposes a data-driven multi-agent simulation model for generating future pedestrian flows from real-world multi-camera CCTV trajectory data. Pedestrian trajectories and inter-camera transition patterns were learned from CCTV spatial information, camera topology, and tracking data, and future trajectories were generated by introducing stochastic variations while preserving the spatiotemporal characteristics of observed movements. The generated pedestrians were implemented in a NetLogo 7.0.4-based multi-agent system integrating microscopic movements within CCTV areas and macroscopic transitions across cameras. The proposed approach directly incorporates movement characteristics learned from real-world CCTV observations and provides a data-driven framework for urban pedestrian congestion analysis and pedestrian-safety decision support.

Citation status

* References for papers published after 2025 are currently being built.