The invention discloses a deep learning traffic state estimation method and system based on Internet of Vehicles sparse information, and provides a low-cost solution for road network traffic state estimation. After a sparse data set of vehicle driving information is collected, vehicles are matched to road sections according to geometric constraints, and the average speed of each road section is obtained to serve as a traffic state estimation value; and the real-time traffic state of the traffic road network is recovered through a TGASA model so as to correct the inaccuracy and instability of the traffic state estimation value. Compared with a method for detecting the traffic state based on a roadside electronic eye, the method can achieve the real-time monitoring of the traffic state of the road section level of all road networks through the sparse mobile sensing data under the condition that the vehicle data is limited, and has the advantages of being lower in cost and wider in coverage. The TGASA model can capture the collaborative correlation of traffic data in time and space to improve the robustness of the TGASA model, is suitable for a dynamically changing graph structure, and has generalization learning ability.
本发明公开了一种基于车联网稀疏信息的深度学习交通状态估计方法及系统,为路网交通状态估计提供了一个低成本的解决方案。本发明在采集车辆行驶信息的稀疏数据集之后,根据几何约束将车辆匹配到路段,得出各路段的平均速度作为其交通状态估计值;再通过TGASA模型恢复出交通路网的实时交通状态,以修正交通状态估计值的不准确性和不稳定性。相比于基于路侧电子眼检测交通状态的方法,本发明能够在车辆数据有限的情况下,使用稀疏的移动感知数据对全部路网实现路段级的交通状态实时监测,具有更低成本,更广覆盖的优点。所提TGASA模型能捕捉交通数据在时空上的协同相关性以提高自身的鲁棒性,适用于动态变化的图结构,具有泛化学习能力。
Deep learning traffic state estimation method and system based on car networking sparse information
基于车联网稀疏信息的深度学习交通状态估计方法及系统
2023-01-31
Patent
Electronic Resource
Chinese
DEEP LEARNING BASED TRAFFIC STATE ESTIMATION SYSTEM AND METHOD THEREOF
European Patent Office | 2025
|A Deep Learning Based Traffic State Estimation Method for Mixed Traffic Flow Environment
DOAJ | 2022
|Traffic state estimation method based on clustering and deep sequence learning
European Patent Office | 2020
|