Advancing Transportation Safety and Sustainability Using Big Data and Deep Learning

点击翻译按钮获取中文摘要

Transportation systems around the world continue to face significant challenges such as persistent safety problems, rapidly rising traffic congestion, and deteriorating air quality. These challenges could be potentially addressed with the latest developments in information and sensor technologies, which are making available massive volumes of data from dedicated sensors embedded in the road network, connected vehicles with vehicle-to-vehicle and vehicle-to-infrastructure communication, and various elements of the transportation systems, as well as the traveling public via their personal communication and wearable devices. Availability of these data along with significant advances in artificial intelligence (AI) affords transportation authorities new opportunities to estimate and predict travel demand and traffic conditions more accurately, speed up transportation services, optimize network supply and improve the overall experience and health of the travelling public. However, before these potentials can be realized fully, many technical challenges must be addressed. Specific challenges for transportation applications include incompleteness and heterogeneity of available data, responsiveness and timeliness of decision-making and information provision, and needs to consideration traveler behavior and responses. The proposed research program focuses on development and advancement of big data analytics and machine learning techniques for monitoring, managing and controlling the road transportation system for maximum safety, efficiency, and sustainability. Specifically, it includes three research themes as follows: (i) Development and Evaluation of Innovative Sensor Technologies: design novel dedicated as well as crowdsourcing techniques for collecting real-time road and traffic condition data, develop new integrated driver, road condition, and vehicle monitoring system, (ii) Estimation and Prediction with Big Data: develop models and algorithms for estimating and predicting travel related decisions and road, traffic, air and driver conditions using various big data sources such as video traffic surveillance systems, loop detectors, Bluetooth/Wifi detectors, mobile phones, and social medias, (iii) Enabling Intelligent Decision Support: develop systematic optimization models and algorithms for intelligently managing travel demand, traffic, and travel experience, such as, adaptive traffic signal control, advanced travel information provision, autonomous driver, and ramp-metering and changeable message signs. The research will provide a foundation for developing a wide array of innovative, application-driven technologies for ubiquitous road weather and traffic condition monitoring, integrated driver information and traffic network management, and advanced traffic control schemes.