Deep Learning Short-Term Traffic Flow Prediction Based on Lane Changing Behavior Recognition

摘要

Short term traffic flow prediction is of great significance for reasonable traffic control and easing traffic congestion. Most of the existing methods are based on the traditional time-space parameters of traffic flow or feature extraction through deep neural network to predict short-term traffic flow. With the increase of road traffic volume, the influence of lane changing behavior on short-term traffic flow is greater. Combined with deep learning and image processing technology, a deep learning short-term traffic flow prediction method based on vehicle lane changing behavior recognition is proposed. The prediction results on real data sets show that the model has high prediction accuracy.