The project’s goal is to develop a hierarchical and distributed hybrid twin to support urban traffic management systems while leveraging Artificial Intelligence (AI), edge cloud computing, and next generation communication networks. A hybrid twin consists of a virtual (i.e., existing traffic simulation) and a digital twin, which integrate physics-based models and assimilate data acquired from infrastructure and in-vehicle sensors for traffic modeling, prediction, and management. The foundational research contributions are data analytics and machine learning including real-time learning for control. The traffic management system will be validated and evaluated via computer simulation and experimentation in the NSF PAWR COSMOS city-scale wireless testbed that is being deployed in West Harlem next to the Columbia campus. This unique urban testbed will provide a realistic environment for the system design and evaluation process, and will also serve as a platform for local community outreach.
Madison Ihrig <email@example.com>, PhD student with Prof. Qiang Du
Igor Kadota <firstname.lastname@example.org>, Postdoc with Prof. Gil Zussman
Kuang Huang <email@example.com>, Postdoc with Prof. Qiang Du
Mahshid Ghasemi Dehkordi <firstname.lastname@example.org>, PhD student with Prof. Gil Zussman
Yongjie Fu <email@example.com>, PhD student with Prof. Xuan Sharon Di
Abhishek Adhikari email, PhD student with Prof. Gil Zussman
Mehmet Kerem Turkcan, firstname.lastname@example.org, postdoc with Prof. Zoran Kostić
Click to see our interactive interface for NYC Traffic Data
(It may take some time to load data)
Video.1 Digital Twin Simulation Applications
In Video.1, we utilize various simulator environments, such as SUMO, CARLA, and NS3, to develop a range of applications including real-world data simulation, V2V and V2X communication, vehicular driving assistant apps, and virtual driving. These simulators are integrated for co-simulation.
Video.2 Connected vehicle and pedestrian simulation using SUMO and NS3
Video.2 shows the implementation of the MMWAVE communication between a traffic infrastructure and vehicles and pedestrians. The traffic infrastructure is implicitly placed in the center of the intersection. The connected vehicles and persons receive UDP packets from the traffic infrastructure at a constant rate. Accidentally, the vehicles and pedestrians receive warning messages. Different colors indicate the states of the vehicles and pedestrians.
Video.3 The "radar screen" system
Video.4 Detection and tracking at COSMOS pilot site
As shown in Video.3, we devised the “radar screen” system and experiment. The system is envisioned to provide real-time evolving snapshots of velocity vectors of all objects in the intersection. The radar screen is composed by AI learning algorithms dynamically distributed across edge and cloud computing resources based on latency requirements and available communications bandwidth. The radar screen will be be wirelessly broadcast to participants in the intersection within a time that can support safety critical applications.Video.4 shows the detection and tracking of objects at COSMOS pilot site (Amsterdam Av. and 120th St., New York City) based on the 12th floor camera feed (birds-eye view). We can keep tracking each vehicle and pedestrian inside the intersection and use computer vision method to give the prediction results.
Video.5 Platform developed with kentyou
1. M. Ghasemi, S. Kleisarchaki, T. Calmant, J. Lu, S. Ojha, Z. Kostic, L. Gürgen, G. Zussman, and J. Ghaderi, “Demo: Real-time Multi-Camera Analytics for Traffic Information Extraction and Visualization,” in Proc. IEEE PerCom’23, 2023.
2. Mo, Zhaobin, and Xuan Di. "Uncertainty quantification of car-following behaviors: Physics-informed generative adversarial networks." in the 28th ACM SIGKDD in conjunction with the 11th International Workshop on Urban Computing (UrbComp2022), 2022.
3. Mo, Zhaobin, Wangzhi Li, Yongjie Fu, Kangrui Ruan, and Xuan Di. "CVLight: Decentralized learning for adaptive traffic signal control with connected vehicles." in Transportation Research Part C: Emerging Technologies, 2022.
4. Bautista-Montesano, Rolando, Renato Galluzzi, Kangrui Ruan, Yongjie Fu, and Xuan Di. "Autonomous navigation at unsignalized intersections: A coupled reinforcement learning and model predictive control approach." in Transportation research part C: emerging technologies 139 (2022): 103662.
5. Mo, Zhaobin, Yongjie Fu, Daran Xu, and Xuan Di. "TrafficFlowGAN: Physics-informed Flow based Generative Adversarial Network for Uncertainty Quantification" in European Conference on Machine Learning and Data Mining (ECML PKDD) , 2022.
6. Ghasemi, Mahshid and Kleisarchaki, Sofia and Calmant, Thomas and Gürgen, Levent and Ghaderi, Javad and Kostic, Zoran and Zussman, Gil "Real-time camera analytics for enhancing traffic intersection safety" in Proc. ACM MobiSys’22, 2022.
7. Yang, Z., Sun, M., Ye, H., Xiong, Z., Zussman, G., & Kostic, Z. (2022, May). "Bird's-eye view social distancing analysis system". In 2022 IEEE International Conference on Communications Workshops (ICC Workshops), 2022.
8. Z. Kostic, A. Angus, Z. Yang, Z. Duan, I. Seskar, G. Zussman, and D. Raychaudhuri, “Smart city intersections: Intelligence nodes for future metropolises,” in IEEE Computer, Special Issue on Smart and Circular Cities , 2022.
9. Duan, Zhuoxu and Yang, Zhengye and Samoilenko, Richard and Oza, Dwiref Snehal and Jagadeesan, Ashvin and Sun, Mingfei and Ye, Hongzhe and Xiong, Zihao and Zussman, Gil and Kostic, Zoran "Smart City Traffic Intersection: Impact of Video Quality and Scene Complexity on Precision and Inference" in Proc. 19th IEEE Int. Conf. on Smart City, 2021.