Dynamic Intersection Signal Control Optimization (DISCO)
Transforming urban mobility and public safety through revolutionary adaptive traffic control

Dynamic Intersection Signal Control Optimisation (DISCO)

Transforming urban mobility and public safety through revolutionary adaptive traffic control

Despite Hong Kong’s 7.5 million residents and soaring skyscrapers, severe urban traffic congestion remains a primary challenge, diminishing economic productivity and air quality. Traditional "static" preset traffic signal schedules fail to respond to real-time gridlocks.

Developed over a 30-year research journey at The Hong Kong University of Science and Technology (HKUST), the Dynamic Intersection Signal Control Optimization (DISCO) system treats urban traffic as a dynamic, interconnected ecosystem. Replacing rigid schedules with real-time, networked adaptive modeling, DISCO resolves bottleneck queues, reduces wait times, lowers accident rates, and optimizes public safety and transit infrastructure.

Team Members

  • Prof. Hong K. LO — Project Leader
  • Dr. Ka Fai NG — Principal Investigator
  • Dr. Lubing LI — Research Scientist
  • Mr. Chak Yan LO — Assistant Research Engineer
  • Mr. Ho Kit WUN — Assistant Research Engineer
  • Mr. Ho Man SO — Research Assistant
  • Mr. Ka Wa LEUNG — Research Assistant
  • Mr. Hao FANG — Research Assistant
  • Mr. Chengshen GE — Research Assistant
  • Mr. Lok Hei SHAM — Research Assistant
  • Mr. Ching Tat YIM — Research Assistant
  • Mr. Yifan ZHAO — PhD student
  • Mr. Zhiyuan ZHOU   — PhD student

Core Innovation 

DISCO is more than just code; it utilizes a dynamic, cell-based, two-stage stochastic programming approach to model the unpredictable traffic flows. This is done by extending the model to include stochastic traffic features and improved computational efficiency, making it ready for real-world implementation through their DISCO traffic control system. This is enabled by real-time simulations over 1,300 times faster than previous methods. This means ensuring shorter waiting, fewer stops, smoother flows, and less frustration for everyday commuters.

  • Dynamic Network Modeling: Unlike legacy static preset schedules, DISCO accurately models physical queuing, traffic shockwaves, and complex gridlock scenarios where vehicles become trapped within intersections. 
  • Implementation & Application: Deployed area-wide adaptive traffic control across 8 signalized junctions, 2 roundabouts, and more.
  • Stochastic Traffic Programming: Integrates stochastic features to account for the inherent randomness, uncertainty, and expected or unforeseen human behaviors in urban driving patterns. 
  • Real-Time Simulation & Calibration: Employs advanced computational techniques that deliver simulation and calibration speeds up 40 to 1,300 times faster than conventional traffic modeling methods. 
  • Large-Scale Network Equilibrium: Built to efficiently optimize massive network structures comprising over 160,000 variables simultaneously across large urban regions. 
  • Adaptive Signal Control Architecture: Processes live traffic state data to provide real-time, area-wide adaptive signal control across multi-junction networks and roundabouts. 
  • Multimodal Scalability: Extends seamlessly from roadway traffic control to heavy rail transit platforms, supporting predictive passenger heatmaps and crowd movement simulations. 
    Overview of the DISCO Smart Mobility System Deployment and Capabilities. Illustrating the transition from static control to real-time adaptive traffic flow across Hong Kong districts, alongside multimodal railway passenger simulation and police manpower optimization in Kwun Tong.
    Overview of the DISCO Smart Mobility System Deployment and Capabilities. Illustrating the transition from static control to real-time adaptive traffic flow across Hong Kong districts, alongside multimodal railway passenger simulation and police manpower optimization in Kwun Tong.

Impact & Metrics

The transformative efficacy of the DISCO system is demonstrated through measurable real-world outcomes across Hong Kong's urban infrastructure. By replacing static presets with dynamic adaptive control and advanced predictive modeling, DISCO has delivered unprecedented improvements in traffic flow efficiency, road safety, enforcement manpower reduction, and mass transit crowd management. The key operational metrics and achievements across DISCO's regional deployments include:

DISCO Regional Impact Map. Highlighting key performance metrics and deployment locations across Hong Kong, including Pok Oi Roundabout, Kwun Tong, Tung Chung, and the MTR rail network.
DISCO Regional Impact Map. Highlighting key performance metrics and deployment locations across Hong Kong, including Pok Oi Roundabout, Kwun Tong, Tung Chung, and the MTR rail network.

    Case 1 – Pok Oi Success

    Background & Challenge: Pok Oi Roundabout in Yuen Long was long plagued by severe congestion during evening peak hours, causing long vehicle queues and driver delays.
    How DISCO Was Applied: Implemented DISCO's dynamic, real-time adaptive traffic control system to continuously adjust signal timing based on live traffic queue dynamics.
    Results & Impact:

    • 77% Reduction in Waiting Time: Reduced evening peak travel times from a 20-minute crawl down to 5 minutes.   
    • 38% Decrease in Accidents: Significantly enhanced road safety by smoothing traffic flow through the roundabout.   
    ​ ​Effect of DISCO at the Pok Oi Roundabout. Notably, a 77% reduction in the waiting time and 38% decrease in accidents were achieved ​ ​
    ​ Effect of DISCO at the Pok Oi Roundabout. Notably, a 77% reduction in the waiting time and 38% decrease in accidents were achieved. ​ ​

    Case 2 – Kwun Tong Smart Enforcement & Manpower Optimization

    Background & Challenge: Over 50 police officers and traffic wardens were required during evening peak hours to manually manage congestion and kerbside parking in Kwun Tong's industrial district.
    DISCO Solution: Collaborated with the Hong Kong Police Force to conduct traffic simulations, evaluate over 30 patrol routes, and recommend optimized patrol paths.
    Results & Impact: 

    • 88% Manpower Reduction: Replaced a 50-officer manual operation with only 6 officers/silhouettes.  
    • Improved Traffic Flow: Solved a decades-old congestion issue in Kwun Tong while freeing up police resources.
    • San Po Kong Rollout (2025): Successfully introduced to the Wong Tai Sin neighborhood to alleviate traffic congestion in the San Po Kong area.   
    • International Expansion: Partnering with the Bangkok Metropolitan Administration to adapt and introduce DISCO to the crowded urban center of Thailand.   
    Following HKUST’s collaboration with the Hong Kong Police Force on the Smart Traffic Management System, DISCO reduced police manpower required for traffic management in Kwun Tong by 88%
    Following HKUST’s collaboration with the Hong Kong Police Force on the Smart Traffic Management System, DISCO reduced police manpower required for traffic management in Kwun Tong by 88%.

    Case 3 – MTR Heavy Rail Network & Planned Disruption Control

    Background & Challenge: Managing daily flows for millions of transit riders and mitigating severe crowding during unexpected or planned rail service disruptions.
    DISCO Solution: Utilized DISCO's predictive simulation models through the HKUST-MTR Joint Research Lab to simulate platform heatmaps and crowd movement.
    Impact & Metrics

    • 5 Million Daily MTR Passengers: Benefited from predictive rail transit simulation and crowd management.
    Professor Lo’s dynamic simulation model supports MTR operations for 5 million daily passengers, demonstrating multimodal transport applicability.
    Professor Lo’s dynamic simulation model supports MTR operations for 5 million daily passengers, demonstrating multimodal transport applicability.

    Milestones & Status

    Horizontal timeline infographic showing DISCO's evolution from Prof. Hong K. Lo’s 2007 prototype to Hong Kong deployments and Bangkok expansion in 2025.
    Development Milestones of DISCO (2007–2025): Tracing the evolution of DISCO from foundational traffic simulation research at HKUST to award-winning, city-scale deployments in Hong Kong and international adaptation in Bangkok.

      Recognition & Awards

        

      Media Coverage
       

      Image Slideshow
      觀塘商貿區第三季起試用鏡頭及AI分析路況以調整交通燈號及調配人手執法
      https://www.881903.com/news/local/2569465/%E8%AD%A6%E6%96%B9%E6%93%AC%E7%AC%AC%E4%B8%89%E5%AD%A3%E5%BC%95%E5%85%A5AI%E4%BA%A4%E9%80%9A%E7%AE%A1%E7%90%86%E7%B3%BB%E7%B5%B1-%E8%A7%80%E5%A1%98%E7%8E%87%E5%85%88%E6%87%89%E7%94%A8
      Image Slideshow
      https://www.hk01.com/%E7%AA%81%E7%99%BC/60268198/%E9%BB%83%E5%A4%A7%E4%BB%99%E8%AD%A6%E5%8D%80%E6%99%BA%E6%85%A7%E4%BA%A4%E9%80%9A%E7%AE%A1%E7%90%86%E8%A8%88%E5%8A%83%E5%95%9F%E5%8B%95-%E6%96%B0%E8%92%B2%E5%B4%97%E5%95%86%E8%B2%BF%E5%8D%80%E5%AF%A6%E6%B8%AC2%E6%9C%88%E6%94%B9%E5%96%84%E5%A1%9E%E8%BB%8A
      警方智慧交通管理計劃啟動 改善新蒲崗商貿區塞車問題
      黃大仙警區智慧交通管理計劃啟動 助改善新蒲崗商貿區塞車問題
      警新蒲崗推智慧交通管理 車流燈過燈減逾三成延誤時間
      https://hk.news.yahoo.com/%E6%96%B0%E8%92%B2%E5%B4%97%E5%A1%9E%E8%BB%8A%E5%95%8F%E9%A1%8C%E6%8C%81%E7%BA%8C-%E7%A7%91%E5%A4%A7-ai-%E5%88%86%E6%9E%90%E5%8A%A9%E8%AD%A6%E6%B8%9B%E5%9F%B7%E6%B3%95%E4%BA%BA%E6%89%8B-%E8%A1%8C%E8%BB%8A%E5%BB%B6%E8%AA%A4%E6%99%82%E9%96%93%E6%B8%9B%E4%B8%89%E6%88%90%EF%BD%9Cyahoo-040418695.html
      https://www.thestandard.com.hk/news/article/310154/Police-launch-AI-powered-traffic-system-to-tackle-congestion-in-San-Po-Kong
      港鐵科大組模型 預測大型活動車流量
      https://www.youtube.com/watch?v=3sT-xTJ1f_c
      Prev
      Next