题目：GE-GAN: a deep learning framework for virtual road traffic state detector design
报告人：魏臣臣
摘要：Traffic state estimation is a crucial elemental function in Intelligent Transportation Systems (ITS) and needs a large number of road traffic state detectors. However, there are redundancy characteristics of road traffic state data, and it is costly to install new road traffic state detectors. In this paper, a virtual road traffic state detector is designed to use information from adjacent links to generate traffic state information. First, the representation of the road network is realized based on graph embedding (GE). Second, with this representation information, the generative adversarial network (GAN) is applied to generate the road traffic state information of the virtual detector. Finally, one typical road network in Caltrans District 7 is adopted as a case study. Experimental results indicate that the generated road traffic state data have higher accuracy than the data generated by other models.