arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.10253cs.LGeess.SP
收藏

使用物理感知神经网络的短期交通预测

Short-term traffic prediction using physics-aware neural networks

Chalmers University of Technology(查尔姆斯理工大学);University of Gothenburg(哥德堡大学)

Mike Pereira, Annika Lang, Balázs Kulcsár

AI总结:

本文提出一种物理感知循环神经网络算法,将宏观交通流模型嵌入网络架构,结合LSTM与简单RNN估计时空交通参数,实现基于环形检测器原始数据的短期车流量预测,并对输入进行物理约束平滑。

摘要:

在这项工作中,我们提出了一种算法,利用过去的车流量测量值对一段道路上的车辆流量进行短期预测。该算法基于一种物理感知的循环神经网络。宏观交通流模型(使用所谓的Traffic Reaction Model)的离散化被嵌入到网络架构中,并基于估计和预测的时空相关交通参数产生流量预测。这些参数本身是通过一系列LSTM和简单循环神经网络获得的。此外,在预测之上,该算法还对其输入进行平滑处理,该平滑处理同样受到宏观交通流模型的物理约束。该算法在从环形检测器获得的原始流量测量数据上进行了测试。

英文摘要:

In this work, we propose an algorithm performing short-term predictions of the flux of vehicles on a stretch of road, using past measurements of the flux. This algorithm is based on a physics-aware recurrent neural network. A discretization of a macroscopic traffic flow model (using the so-called Traffic Reaction Model) is embedded in the architecture of the network and yields flux predictions based on estimated and predicted space-time dependent traffic parameters. These parameters are themselves obtained using a succession of LSTM ans simple recurrent neural networks. Besides, on top of the predictions, the algorithm yields a smoothing of its inputs which is also physically-constrained by the macroscopic traffic flow model. The algorithm is tested on raw flux measurements obtained from loop detectors.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑