Ground vehicle traffic jams are a serious issue in today’s society. Traffic jams cause significant economic and environmental problems ranging from reduction in the efficiency of corporations to the increase of both fuel consumption and air pollution related with vehicle emissions. Hence, numerous efforts have been devoted to model the complex dynamics of traffic and to predict traffic jams. Traffic jams are fundamentally affected by a variety of factors, from driver behavior to weather, which are difficult to model. As a result, accurate models are very difficult to obtain and the results of predictions and real-world traffic phenomena might not be consistent. In recent years, the growing availability of traffic data and advances in computational techniques have motivated researchers to study and forecast traffic jams using data assisted and data-driven techniques. These methods, however, usually demand extensive data for model building and prediction. Hence, such methods perform well on specific locations where data is available, but the applicability of the models remains a question in places where data availability is an issue. As a result, despite advances in traffic flow predictions in recent years, predicting traffic jams before they occur is still a challenge. Recently, novel techniques have been developed in complex systems theory to enable forecasting emergent behaviors in dynamical systems. Forecasting methods have been developed based on exploiting the phenomenon of critical slowing down which occurs in dynamical systems near certain types of bifurcations and phase transitions. Previous studies have shown that traffic jams are the result of a bifurcation where the traffic dynamics loses stability once passing a tipping point. Hence, one can take advantage of these developments to forecast upcoming traffic instabilities on roads or use new features in current data-assisted algorithms to improve their accuracy and applicability. Herein, we introduce tools for tipping point forecasting in complex systems, namely early warning indicators and bifurcation forecasting methods, and investigate their application to predict traffic jams on roads. The proposed methods are model-free representing substantial advantages for applications such as traffic flow systems where an accurate mathematical model is not easily available. Measurements required for forecasting are recorded dynamical features of the system such as headways between cars in traffic or density of cars on road. Forecasting approaches are applied to simulated and experimental traffic flow conditions. Results show that one can successfully predict the proximity of the traffic to a congestion and the traffic dynamics after this critical point using the proposed approaches. The methodologies presented can be used as a basis to analyze the stability of complex traffic models and to address challenges related to the complexity of traffic dynamics.


Data-Driven Methods for Detecting Traffic Jams in Vehicular Traffic Systems

Division / Sponsors: 
Aerospace, Bioengineering, Applied Mechanics, Design Engineering Division, Advanced Energy Systems, Manufacturing Engineering Division, Noise Control and Acoustics, Nondestructive Evaluation, Diagnosis, and Prognosis (NDPD), Nuclear Engineering Division, Process Industries Division, Fluids Engineering, Heat Transfer, Materials, Microelectromechanical Systems, Safety Engineering and Risk Analysis, Computers & Information in Engineering


Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2020/volume/84478

IMECE2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.