高速公路短期旅行时间预测：数据驱动方法的综述

需积分: 10 74 浏览量更新于2024-07-16 收藏 280KB PDF 举报

"短途高速公路旅行时间预测：数据驱动方法的综述" 这篇论文"Short-term travel-time prediction on highway: a review of the data-driven approach"主要探讨了如何利用数据驱动的方法来预测高速公路的短期旅行时间。在当今信息化社会，交通管理系统需要实时、准确地预测旅行时间以优化交通流量、提升道路效率并提供给驾驶者有效的行程规划信息。数据驱动的方法是近年来在交通工程领域发展起来的一种重要技术。它依赖于收集到的各种交通数据，如车辆速度、交通流量、路面状况等，通过高级的数据分析和机器学习算法，构建预测模型。这些模型能够处理大量的历史数据，找出隐藏的模式和趋势，并以此预测未来的旅行时间。论文作者包括Simon Oha、Young-Ji Byon、Kitae Jang和Hwasoo Yeo，他们分别来自韩国科学技术院（KAIST）、哈利法科技大学（KUSTAR）以及Cho Chun Shik绿色交通研究生院。这些研究机构在交通技术和数据分析方面具有显著的专业背景。文章可能涵盖了以下关键知识点： 1. **数据源**：讨论了不同类型的数据源，如交通监控摄像头、车载传感器、GPS设备、浮动车数据等，这些数据源如何用于收集实时交通信息。 2. **数据预处理**：在进行预测模型构建前，通常需要对原始数据进行清洗、整合和标准化，以消除异常值、缺失值并确保数据质量。 3. **预测模型**：介绍了多种数据驱动的预测模型，如时间序列分析、支持向量机（SVM）、随机森林、神经网络等，以及它们在旅行时间预测中的应用和优势。 4. **性能评估**：讨论了评价预测模型准确性的指标，如均方误差（MSE）、平均绝对误差（MAE）和决定系数（R²），以及如何通过交叉验证来评估模型的稳定性和泛化能力。 5. **挑战与未来方向**：论文可能还涉及数据驱动预测面临的挑战，如大数据处理的复杂性、实时性需求、模型的可解释性以及如何应对交通状况的不确定性等问题，并提出未来的研究方向。这篇综述对于理解数据驱动的旅行时间预测方法在交通工程中的应用具有重要意义，不仅为交通管理部门提供了理论参考，也为交通领域的研究者提供了最新的研究进展和潜在的研究问题。

(3) Parametric regression (ARIMA, Kalman ﬁlter), Non-parametric regression

(Nearest neighbourhood), and NNs (Chrobok, 2005)

(4) Naı

ve (Instantaneous, Historical averages, and Cluster analysis), Parametric

(Trafﬁc ﬂow models (Model based), Linear regression, ARIMA, Kalman ﬁlter-

ing), and Non-parametric models (NNs, k-NN, etc.) (van Hinsbergen, van

Lint, & Sanders, 2007)

(5) Regression (Linear regression), Time series (ARIMA, Kalman ﬁlter), and NNs

(Shen, 2008)

(6) Parametric (Regression models, Time series (ARIMA, Kalman Filter)) and

Non-parametric (Artiﬁcial intelligence (ANN), Pattern search (k-NN))

approaches (Yu et al., 2008)

(7) Parametric (Linear regression, Time series, Kalman ﬁlter) and Non-parametric

(NNs, Bayesian models, pattern recognition (k-NN)) methods (Fei et al., 2011)

In compliance with the previous researches’ taxonomy on the data-driven

approach, we propose a set of criteria for classifying and evaluating the data-

driven approaches considering underlying mechanisms and theoretical prin-

ciples. Figure 1 shows the taxonomy of data-driven approaches.

In the parametric approach the functional relationship between the explanatory

and response variables is known, and some unknown parameters may be esti-

mated from the training set. Selecting input variables and estimating coefﬁcients

minimizing errors are the key issues for this approach. The parametric statistical

approaches are known to perform quite accurately despite their simple formu-

lations provided that they have well-established theoretical and mathematical

backgrounds and are validated by transportation engineers. The main drawback

of this approach is that coefﬁcients are site-speciﬁc and it is difﬁcult to implement

in large-scale networks.

NNs are non-parametric models that predict travel-times by training them-

selves with historical data which mimic the mechanisms of a human brain. The

parameters (e.g. weights) have no physical meaning in regard to the problems

to which they are applied. The main advantage of using NNs in transportation

applications is that they can handle complex and non-linear properties that are

inherently embedded in the nature of many transportation engineering problems.

The method has been validated by many researchers with acceptable accuracy.

Complex training with site-speciﬁc limitations and black-box procedures

involved are the main demerits of artiﬁcial NNs (ANNs). Poor logical descriptions

with regards to trafﬁc mechanisms may be questioned by various audiences who

Figure 1. Taxonomy of a data-driven approach to travel-time prediction.

Short-term Travel-time Prediction on Highway 7

Downloaded by [University of Nebraska, Lincoln] at 07:32 08 April 2015

may demand intuitive rationales for their ﬁeld applications regardless of their pre-

diction performances.

Another non-parametric method that is widely used in literatures is the nearest

neighbourhood method. This method matches similar historical patterns with the

current one by searching archived databases. This non-parametric approach is

expected to perform well with large data sets. However, this dependency inevita-

bly requires the integrity of data and sufﬁcient sizes of databases. The following

sections present and describe various data-driven approaches with respect to

their strengths, weaknesses, and performances.

2. Review of Data-driven Approaches

2.1. The Linear Regression and Time Series Modelling Approach

The parametric approach treats travel-time prediction problems with a pre-struc-

tured model by ﬁtting the parameters using data, and there are both merits and

demerits as discussed in previous sections. According to forecasting mechanisms

and underlying rationales, the parametric models can be classiﬁed as linear

regression, ARIMA, of which ARIMA is considered as a time series-based

approach, and Kalman ﬁlter. The descriptions and performances of these

approaches are listed in Tables 1 and 2, respectively.

2.1.1. Linear Regression. Prediction functions in linear regression basically

assume a linear combination of covariates. Several researchers have conducted

regression analyses for deriving future travel times from relevant variables. Due

to their relatively simple structures, the researchers consistently conﬁrm the

high efﬁciency of the method in terms of computations.

Kwon, Coifman, and Bickel (2000) predict travel times using the linear

regression with a stepwise method for covariates using a heterogeneous data

set. The current trafﬁc state is found to be the most inﬂuencing factor for short-

term predictions, while the historical data are more useful in predictions for

longer prognosis horizons. The regression model is fed with observed travel-

times from probe-vehicles as a response variable and others are treated (VDS

data, departure time, and day of week) as covariates. Then the model has been

tested on I-880N&S. The observations show large variations in the metrics in

the day-to-day scenario and their strong correlations with travel times, indicating

the signiﬁcant inﬂuence of the metrics on travel times. It is noted that the input

explanatory variables have been ﬁltered through the stepwise method which is

unique in their works.

For different recurring and non-recurring congestion scenarios, the authors

improve the explanatory power through the use of abnormality measures detect-

ing outlying days from the normal days. The paper ﬁnds that the 20-min predic-

tion time frame is beneﬁcial, producing similar prediction errors for all four

scenarios. On average, the resulting errors are found as 116.75 of root MSPE

(95149 s) and 14.1% of MAPPE (1116.6%). Authors state that relatively short

prediction horizons and small spatial ranges are the major limitations of the pro-

posed model.

TVC, ATHENA, and Bayesian prediction. Zhang and Rice (2003), and Rice and

van Zwet (2004) predict travel times using the method of simple linear regression

8 S. Oh et al.

Downloaded by [University of Nebraska, Lincoln] at 07:32 08 April 2015

剩余30页未读，继续阅读

qq_24890901

粉丝: 19
资源: 4

高速公路短期旅行时间预测：数据驱动方法的综述

[machine_learning_mastery系列]long-short-term-memory-networks-with-python.pdf

long－short－term－memory－networks－with python

Long-term Planning by Short-term Prediction

please write a paragraph that explain the application of transformer in time-series prediction

lstm环境污染的参考文献

de-normalization lstm-neural-network-for-time-series-prediction-master

python time series predict model multi

Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting

复现lstm-neural-network-for-time-series-prediction

最新资源