學術活動 NAVIGATION
Real-Time Signal Control for Large Heterogeneous Traffic Networks
時間：2024-12-23
來源：薄翠梅
作者：
攝影： 編輯：馮羽璐
上傳：
報告題目：Real-Time Signal Control for Large Heterogeneous Traffic Networks
報告人：Prof. Rong Su
報告人單位：新加坡南洋理工大學
報告時間：2024年12月24日（周二）14:30-16:30
會議地點：電氣工程與控制科學學院416報告廳（崇德樓D座416）
舉辦單位：電氣工程與控制科學學院
報告人簡介：RongSu，教授，博導，新加坡南洋理工大學電氣與電子工程學院計算機控制與自動化碩士項目負責人、KTH-NTU 聯合博士項目管理委員會 NTU負責人。研究方向包括多智能體系統、離散事件系統理論、基于模型的故障診斷、網絡安全分析和綜合、復雜網絡的控制和優化，以及在柔性制造、智能交通、人機界面、電源管理和綠色建筑中的應用，擁有330多篇期刊和會議出版物、2部專著、18項已授予/申請的專利，曾獲得多項最佳論文獎，包括IEEE/CAA Journal of Automatica Sinica 2021年Hsue-shen Tsien論文獎等。 目前，他擔任IEEE Transactions on Cybernetics、Automatica (IFAC)、Journal of Discrete Event Dynamic Systems: Theory and Applications和Journal of Control and Decision的副主編。
報告摘要：Traffic congestion in urban areas significantly increases the commuting time for passengers and introduces unnecessary fuel burns and carbon emissions to the fragile urban ecosystem. Traffic lights, which is introduced to improve the order in traffic systems, may harm the traveling efficiency if the green times are not properly assigned for each approach. Sensors and controllers are implemented in modern intelligent transportation systems to generate traffic-responsive signal plans, which highly depends on the topological parameter estimation and traffic model based optimization. In this paper, to fulfill the requirements of constructing a V2X-enabled traffic light control scheme, a closed-loop traffic light scheduling strategy is proposed. A macroscopic model is introduced to depict the traffic movements in the network, which involves the traffic flow dynamics and the prediction of speed variations. A mixed integer linear model is elaborated to generate optimal traffic light plans. Topological parameters, such as turning ratios from each approach, are required to precisely depict the traffic movements. A learning-based parameter estimator is designed to on-line predict the turning ratios based on historical traffic data and traffic light assignments. Simulations show that the convergence is achieved under constant cyclic flow profiles, and our proposed closed-loop traffic light scheduling strategy could achieve significant reduction on key performance indices.
審核人：薄翠梅
