Research Projects

India and the world are striving for sustainable, safe, and smart cities (SDG 11), with urban transportation focusing on reducing emissions, minimizing crashes, and implementing intelligent systems to improve traffic flow. However, attaining these goals are challenging in complex, human-influenced traffic systems, particularly under heterogeneous conditions. Advancements in vehicle technologies, driver behavior data collection, and digital twin technology offer new opportunities to address these challenges. Given these advancements, reimagining traditional microscopic traffic flow models is crucial, as current approaches fail to capture real-world driving behavior comprehensively, overlook crash risks and emissions, and lack insights into driver responses to information assistance. To overcome these limitations, this study employs a transformative approach—one that integrates physical laws, behavioral theories, and machine learning methodologies to simultaneously estimate acceleration, crash risk, and emissions. The primary objectives of this project are: • To develop a Physical-Behavioral-Machine Learning (PBM) modeling framework to accurately replicate microscopic driver behavior under heterogeneous traffic conditions • To advance the PBM model to enable simultaneous estimation of acceleration, safety metrics, and emissions at each time instant, enhancing its predictive and analytical capabilities • To build an open-source SUMO-based simulation platform to emulate traffic dynamics at a minor network level (a small-scale digital twin for real-world applications) • To design a driving style-based alert algorithm using insights from the PBM model, and implement it as a proof-of-concept hardware device (TRL 3). Method: A test-track experiment involving 100 drivers will replicate heterogeneous traffic scenarios to capture diverse driving behaviors and the effects of information assistance. Data collected on vehicle trajectories, driver actions, demographics, emissions, and psychological factors will undergo rigorous analysis to refine the model. The project includes creating an open-source SUMO-based simulation platform to evaluate traffic flow, safety, and emissions at a minor network level, and testing ITS-based technologies. Additionally, a driver style-based alert system algorithm will be developed and validated on hardware as a proof of concept through field trials. Contributions: Theoretically, it provides a comprehensive understanding of real-world driving behavior, particularly under heterogeneous traffic conditions with information assistance, and introduces an innovative PBM model. Practically, it delivers an open-source simulation platform, alongside a driving style-based alert system demonstrated through a proof-of-concept hardware device (TRL 3). These contributions will set a strong foundation for future innovations in sustainable and smart transportation systems.