Multi-Source Data Fusion for Traffic Prediction Based on Fog/Cloud Computing and Deep Learning

Traffic congestion in developing countries is an issue where the aspects of the problem are different from those of developed countries. Such an issue comes from the rapid demographic growth and economic stagnation that is matched with weak state infrastructure or the slow rate of infrastructure development. This results in longer trip times, lower air quality, and increased fuel wastage, which in turn affects the overall quality of life. The main objective of this study is to introduce a robust Internet of Vehicles (IoV) architecture based on fog/cloud to reduce data dissemination problems and save network bandwidth. In the proposed architecture, the Deep Learning (DL) technique is applied locally within the fog layer. This operation speeds up the prediction time and improves performance in vehicular networks. The novelty of this study can be spanning: (i) a semantically intelligent expert system for detecting accidents based on the road, weather, and traffic data, (ii) an optimized feature set and optimized deep learning pipelines based on the current.