The Study of Traffic Flow Information Completion Based on GAN Algorithm
Min Li, Wang Li, Jiaqing Yan, Haibo Zhang, Lili Zhang, Lingyu Zhang · 2018 IEEE 7th Data Driven Control and Learning Systems Conference (DDCLS) · 2018
In urban road traffic, detectors often cause incomplete data and missing data as a result of inadequate coverage or equipment damage and other reasons. Therefore, the data needs to be repaired to ensure data support for the traffic management service. This paper regards traffic flow data from section geomagnetic detectors as the object, processing graphically section flow information. And the missing data of network is predicted and complemented by the idea of generating network analysis images. This paper analyzes the influence of missing area size and loss at random of data on the accuracy of complete information. The results prove the feasibility and applicability of this method.