Mohammdreza Khajeh Hosseini, an engineer and venture investor, has a Ph.D. in Transportation Engineering from the University of Illinois and a background in business management. With a focus on Mobility, Robotics, and Automation, he has a unique perspective on the intersection of technology and business. Reza has pursued multidisciplinary research projects sponsored by prestigious institutes such as the NSF, DOE, and SAE and has presented and published multiple papers in the field of connected and automated vehicles and human-vehicle interactions, as well as the development of autonomous vehicles. At Plug and Play tech center, Reza leverages his expertise to identify promising startups and investment opportunities in the mobility industry. His passion for technology and business drives him to stay on the cutting edge of industry trends and innovative solutions.
Accurate traffic state prediction is crucial for effective traffic management, but conventional data sources like loop detectors and stationary sensors have historically made it difficult. The dynamic nature of traffic flow and the interactions between vehicles and their environment have contributed to the difficulty. However, connected and automated vehicles have the potential to address these limitations by providing accurate vehicle trajectory data that can be transformed into a time-space diagram to capture vehicle interactions. The study proposes two traffic state prediction methodologies based on convolutional neural networks (CNN) that can utilize the time-space diagram. The first methodology directly uses the time-space diagram in the prediction process, while the second combines microscopic and macroscopic predictions to directly use the interactions among vehicles. The study adopts a probabilistic approach to predict the location of individual vehicles based on different maneuvers and convert these predictions into aggregated traffic state predictions (i.e. flow, space-mean speed, and density). The methodologies are data-driven and require accurate and comprehensive training datasets, which the study acquires through simulation-based and real-world vehicle trajectory datasets
شما می توانید ویدئوی جلسه برگزار شده را مشاهده نموده و مستندات مربوط به آن را از طریق لینک زیر دانلود نمایید.
