Accurate and timely traffic flow information is important for the successful deployment of intelligent transportation systems. Over the last few years, traffic data have been exploding, and we have truly entered the era of big data for transportation. Existing traffic flow prediction methods mainly use shallow traffic prediction models and are still unsatisfying for many real-world applications. This situation inspires us to rethink the traffic flow prediction problem based on deep architecture models with big traffic data. In this paper, a novel deep-learning-based traffic flow prediction method is proposed, which considers the spatial and temporal correlations inherently. A stacked autoencoder model is used to learn generic traffic flow features, and it is trained in a greedy layerwise fashion. To the best of our knowledge, this is the first time that a deep architecture model is applied using autoencoders as building blocks to represent traffic flow features for prediction. Moreover, experiments demonstrate that the proposed method for traffic flow prediction has superior performance.

Traffic Flow Prediction With Big Data: A Deep Learning Approach

The development of the Internet in recent years has made it possible and useful to access many different information systems anywhere in the world to obtain information. While there is much research on the integration of heterogeneous information systems, most commercial systems stop short of the actual integration of available data. Data fusion is the process of fusing multiple records representing the same real-world object into a single, consistent, and clean representation.This article places data fusion into the greater context of data integration, precisely defines the goals of data fusion, namely, complete, concise, and consistent data, and highlights the challenges of data fusion, namely, uncertain and conflicting data values. We give an overview and classification of different ways of fusing data and present several techniques based on standard and advanced operators of the relational algebra and SQL. Finally, the article features a comprehensive survey of data integration systems from academia and industry, showing if and how data fusion is performed in each.

Data fusion

Highway Capacity Manual改訂の動向--テイラ-教授の講演より

Short-term traffic forecast is one of the essential issues in intelligent transportation system. Accurate forecast result enables commuters make appropriate travel modes, travel routes, and departure time, which is meaningful in traffic management. To promote the forecast accuracy, a feasible way is to develop a more effective approach for traffic data analysis. The availability of abundant traffic data and computation power emerge in recent years, which motivates us to improve the accuracy of short-term traffic forecast via deep learning approaches. A novel traffic forecast model based on long short-term memory (LSTM) network is proposed. Different from conventional forecast models, the proposed LSTM network considers temporal-spatial correlation in traffic system via a two-dimensional network which is composed of many memory units. A comparison with other representative forecast models validates that the proposed LSTM network can achieve a better performance.

https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/iet-its.2016.0208

LSTM network: a deep learning approach for short-term traffic forecast

Traffic flow prediction is a fundamental problem in transportation modeling and management. Many existing approaches fail to provide favorable results due to being: 1) shallow in architecture; 2) hand engineered in features; and 3) separate in learning. In this paper we propose a deep architecture that consists of two parts, i.e., a deep belief network (DBN) at the bottom and a multitask regression layer at the top. A DBN is employed here for unsupervised feature learning. It can learn effective features for traffic flow prediction in an unsupervised fashion, which has been examined and found to be effective for many areas such as image and audio classification. To the best of our knowledge, this is the first paper that applies the deep learning approach to transportation research. To incorporate multitask learning (MTL) in our deep architecture, a multitask regression layer is used above the DBN for supervised prediction. We further investigate homogeneous MTL and heterogeneous MTL for traffic flow prediction. To take full advantage of weight sharing in our deep architecture, we propose a grouping method based on the weights in the top layer to make MTL more effective. Experiments on transportation data sets show good performance of our deep architecture. Abundant experiments show that our approach achieved close to 5% improvements over the state of the art. It is also presented that MTL can improve the generalization performance of shared tasks. These positive results demonstrate that deep learning and MTL are promising in transportation research.

Deep Architecture for Traffic Flow Prediction: Deep Belief Networks With Multitask Learning

The capability to forecast traffic volume in an operational setting has been identified as a critical need for intelligent transportation systems (ITS). In particular, traffic volume forecasts will support proactive, dynamic traffic control. However, previous attempts to develop traffic volume forecasting models have met with limited success. This research effort focused on developing traffic volume forecasting models for two sites on Northern Virginia's Capital Beltway. Four models were developed and tested for the freeway traffic flow forecasting problem, which is defined as estimating traffic flow 15 minutes into the future. They were the historical average, time-series, neural network, and nonparametric regression models. The nonparametric regression model significantly outperformed the other models. A Wilcoxon signed-rank test revealed that the nonparametric regression model was easy to implement, and proved to be portable, performing well at two distinct sties. Based on its success, research is ongoing to refine the nonparametric regression model and to extend it to produce multiple interval forecasts.

Traffic flow forecasting: comparison of modeling approaches

Much of the current activity in the area of intelligent vehicle-highway systems (IVHSs) focuses on one simple objective: to collect more data. Clearly, improvements in sensor technology and communication systems will allow transportation agencies to more closely monitor the condition of the surface transportation system. However, monitoring alone cannot improve the safety or efficiency of the system. It is imperative that surveillance data be used to manage the system in a proactive rather than a reactive manner. Proactive traffic management will require the ability to predict traffic conditions. Previous predictive modeling approaches can be grouped into three categories: (a) historical, data-based algorithms; (b) time-series models; and (c) simulations. A relatively new mathematical model, the neural network, offers an attractive alternative because neural networks can model undefined, complex nonlinear surfaces. In a comparison of a backpropagation neural network model with the more traditional approaches of an historical, data-based algorithm and a time-series model, the backpropagation model was clearly superior, although all three models did an adequate job of predicting future traffic volumes. The backpropagation model was more responsive to dynamic conditions than the historical, data-based algorithm, and it did not experience the lag and overprediction characteristics of the time-series model. Given these advantages and the backpropagation model's ability to run in a parallel computing environment, it appears that such neural network prediction models hold considerable potential for use in real-time IVHS applications.

Short-term traffic flow prediction: neural network approach

Traffic flow prediction models are expected to play an important role in intelligent vehicle highway systems. This paper demonstrate that the nearest neighbour models have the potential to serve as accurate and portable traffic flow prediction models. Furthermore the models have the advantages of being easily understood by field personnel. >

Short-term traffic flow prediction models-a comparison of neural network and nonparametric regression approaches

Freeway traffic flow forecasting will play an important role in intelligent transportation systems. The TRB Committee on Freeway Operations has included freeway flow forecasting in its 1995 research program. Much of the past research in traffic flow forecasting has addressed short-term, single-interval predictions. Such limited forecasting models will not support the development of the longer-term operational strategies needed for such events as hazardous material incidents. A multipleinterval freeway traffic flow forecasting model has been developed that predicts traffic volumes in 15-min intervals for several hours into the future. The nonparametric regression modeling technique was chosen for the multiple-interval freeway traffic flow forecasting problem. The technique possesses a number of attractive qualities for traffic forecasting. It is intuitive and uses a data base of past conditions to generate forecasts. It can also be implemented as a generic algorithm and is easily calibrated at field locati...

Multiple-Interval Freeway Traffic Flow Forecasting:

With the recent advances in communications and information technology, real-time traffic routing has emerged as a promising approach to alleviating congestion. Existing approaches to developing real-time routing strategies, however, have limitations. This study examines the potential for using case-based reasoning (CBR), an emerging artificial intelligence paradigm, to overcome such limitations. CBR solves new problems by reusing solutions of similar past problems. To illustrate the feasibility of the approach, the study develops and evaluates a prototype CBR routing system for the interstate network in Hampton Roads, Virginia. Cases for building the system’s case-base are generated using a heuristic dynamic traffic assignment (DTA) model designed for the region. Using a second set of cases, the study evaluates the performance of the prototype system by comparing its solutions to those of the DTA model. The evaluation results demonstrate that the prototype system is capable of running in real-time , and of producing high quality solutions using case-bases of reasonable size.

Michael J Demetsky

Papers

Traffic flow forecasting: comparison of modeling approaches

Short-term traffic flow prediction: neural network approach

Short-term traffic flow prediction models-a comparison of neural network and nonparametric regression approaches

Multiple-Interval Freeway Traffic Flow Forecasting:

A prototype case-based reasoning system for real-time freeway traffic routing