A Vehicle Speed Prediction Method Integrating Multi-Source Traffic Information Based on Informer

Vehicle speed prediction is of great significance for intelligent transportation and eco-driving. Currently, mainstream methods for speed prediction rely more on the vehicle's own historical data, ignoring the influence of the surrounding traffic environment. This paper proposes a vehicle speed prediction method based on Informer, which integrates real-time multi-source traffic information to improve prediction accuracy. K-means clustering is used to cluster the following mode and traffic flow mode. During prediction, a back propagation neural network is employed for recognition, and the recognition results are used as inputs to the prediction model, achieving the extraction and integration of traffic information. Experimental results demonstrate that the Informer-based vehicle speed prediction method outperforms current mainstream deep learning methods in prediction accuracy, and the integration of multi-source traffic information in speed prediction surpasses methods that do not integrate traffic information.