Resultado de búsqueda
23 de abr. de 2024 · However, it is not considered in previous works how to model the influence of different traffic observation on the task of traffic flow prediction. To address the above issues, in this paper we propose a deep multi-view channel-wise spatial-temporal network model named MVC-STNet for traffic flow prediction.
26 de abr. de 2024 · Traffic flow prediction. Traffic flow prediction is to use the spatial–temporal data collected by road sensors to make as accurate a prediction as possible for the traffic state in...
14 de abr. de 2024 · The traffic flow theory shows a correlation between flow, speed, and occupancy. Hence, we propose combining the proposed prediction model with TL to obtain knowledge transfer from traffic volume to occupancy/speed prediction tasks, achieving better performance.
16 de abr. de 2024 · interpretability in traffic prediction models remains to be a challenge due to the complexity of traffic data and the inherent opacity of deep learning models. To tackle these challenges, we pro-pose a novel approach, Traffic Flow Prediction LLM (TF-LLM), which leverages large language models (LLMs) to generate interpretable traffic flow ...
26 de abr. de 2024 · Qian et al. designed a novel traffic flow prediction model named MDRGCN, which utilizes two relation matrices to capture various traffic flow characteristics. By combining multi-modal dynamic graph convolutions with gated recurrent units, it achieves the integration of spatiotemporal interactions.
25 de abr. de 2024 · Traffic data was collected by sensors spread throughout the city, which recorded how many vehicles passed every five minutes. It also includes longitude, latitude and timestamp. Zhang and Kabuka create a traffic flow forecast model based on traffic data, collected in real time, from over 39,000 individual detectors in the state of California.
24 de abr. de 2024 · Accurate traffic flow prediction is crucial for optimizing traffic management, enhancing road safety, and reducing environmental impacts. Existing models face challenges with long sequence data, requiring substantial memory and computational resources, and often suffer from slow inference times due to the lack of a unified summary state.
