Short-term Travel Time Prediction Research Articles

Short-term bus travel time prediction for transfer synchronization with intelligent uncertainty handling

This paper presents two novel approaches for uncertainty estimation adapted and extended for the multi-link bus travel time problem. The uncertainty is modeled directly as part of recurrent artificial neural networks, but using two fundamentally different approaches: one based on Deep Quantile Regression and the other on Bayesian neural network. Both approaches use a recurrent neural network to predict multiple time steps into the future, but handle the time-dependent uncertainty estimation differently. We present a novel sampling technique in order to aggregate quantile estimates for link level travel time to yield the multi-link travel time distribution needed for a vehicle to travel from its current position to a specific downstream stop point or transfer site.To motivate the relevance of uncertainty-aware models in the domain, we focus on the connection protection application as a case study: An expert system to determine whether a bus driver should hold and wait for a connecting service, thus ensuring the connection, or break the connection and reduce its own delay. Our results show that the proposed quantile sampling method performs overall best for the 80%, 90% and 95% prediction intervals, both for a 15 min time horizon into the future (t+1), but also for the 30 and 45 min time horizon (t+2 and t+3), with a constant, but very small underestimation of the uncertainty interval (1–4 pp.). However, we also show, that the Bayesian model still can outperform the DQR for specific cases. Lastly, we demonstrate how a simple decision support system can take advantage of our uncertainty-aware travel time models to prioritize the difference in travel time uncertainty for bus holding at strategic points, thus reducing the introduced delay for the connection protection application.

An attention-based recurrent learning model for short-term travel time prediction.

With the advent of Big Data technology and the Internet of Things, Intelligent Transportation Systems (ITS) have become inevitable for future transportation networks. Travel time prediction (TTP) is an essential part of ITS and plays a pivotal role in congestion avoidance and route planning. The novel data sources such as smartphones and in-vehicle navigation applications allow traffic conditions in smart cities to be analyzed and forecast more reliably than ever. Such a massive amount of geospatial data provides a rich source of information for TTP. Gated Recurrent Unit (GRU) has been successfully applied to traffic prediction problems due to its ability to handle long-term traffic sequences. However, the existing GRU does not consider the relationship between various historical travel time positions in the sequences for traffic prediction. We propose an attention-based GRU model for short-term travel time prediction to cope with this problem enabling GRU to learn the relevant context in historical travel time sequences and update the weights of hidden states accordingly. We evaluated the proposed model using FCD data from Beijing. To demonstrate the generalization of our proposed model, we performed a robustness analysis by adding noise obeying Gaussian distribution. The experimental results on test data indicated that our proposed model performed better than the existing deep learning time-series models in terms of Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Coefficient of Determination (R2).

Constructing Cooperative Intelligent Transport Systems for Travel Time Prediction With Deep Learning Approaches

Drivers and traffic system planners require accurate forecasting of future travel times. To alleviate traffic congestion on Taiwan’s Provincial Highway No. 61, this study considers temporal (weekdays, weekends or continuous holidays) and spatial characteristics (road types), and establishes short-term and long-term travel time prediction models. Data pre-processing is accomplished using the Google Maps API and floating car method to verify travel time comparisons, finding that post-processing travel times are reliable, credible and reasonable. This study develops a collaborative intelligent transportation system (CITS) based on 9 different algorithms for the prediction of current and future travel-time. The results show that <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> NN-R provides the most accurate short-term predictions, with average prediction error within 20 seconds per kilometer. For long-term forecasting, SARIMAX and fbProphet provide the most accurate results for weekday and continuous holiday modes. Travel time prediction can assist traffic management agencies in the timely implementation of appropriate traffic management measures. CITS also provides real-time traffic query and travel time prediction functions to help drivers avoid traffic congestion.

Short-Term Travel-Time Prediction using Support Vector Machine and Nearest Neighbor Method

This paper presents an investigation into the performance of support vector machine (SVM) in short-term travel-time prediction in comparison with baseline methods, including the historical mean, current time based, and time varying coefficient predictors. To demonstrate the SVM performance, 1-month time-series speed data on a section of Pan-Island Expressway in Singapore were used to estimate the travel time for training and testing the SVM model. The results show that the SVM method significantly outperforms the baseline methods in both normal and recurring congestion over a wide range of prediction intervals. In studying SVM prediction behavior under incident situations, the results show that all the predictors are not responsive enough using 15-minute aggregated field data, but the SVM predicted outcome follows the test data profile closely for 2-minute aggregated simulated data. Finally, to improve the prediction performance, an empirical k-nearest neighbor method is introduced to retrieve patterns closest to the test vector for SVM training. The results show that k-Nearest Neighbor is an attractive tool for SVM travel-time prediction. In retrieving the most similar patterns for SVM training, k-nearest neighbor allows dramatic reduction of training size to accelerate the training task while maintaining prediction accuracy.

Spatial-temporal attention-based seq2seq framework for short-term travel time prediction

Short-term travel time prediction is an important research direction of the intelligent transportation system. However, due to the complex temporal and spatial correlation of the traffic data, the more accurate and efficient the prediction of the short-term travel time, the harder it is to achieve. In order to conquer the problems and accelerate the construction of smart cities, this study first proposes the seq2seq model based on LSTM network with strong time series processing capability and can avoid manual feature extraction. Then, a seq2seq model based on multi-level attention mechanism is proposed. Finally, the method proposed is evaluated and compared with the traditional prediction methods through experimental simulations based on the real traffic dataset. The results verify the accuracy and effectiveness of the proposed models, demonstrate that the utilisation of multi-level attention mechanism can effectively integrate temporal and spatial features of the dataset, and improve the efficiency of the original seq2seq model.

Travel time forecasting on a freeway corridor: a dynamic information fusion model based on the random forests approach

Purpose Metropolitan areas suffer from frequent road traffic congestion not only during peak hours but also during off-peak periods. Different machine learning methods have been used in travel time prediction, however, such machine learning methods practically face the problem of overfitting. Tree-based ensembles have been applied in various prediction fields, and such approaches usually produce high prediction accuracy by aggregating and averaging individual decision trees. The inherent advantages of these approaches not only get better prediction results but also have a good bias-variance trade-off which can help to avoid overfitting. However, the reality is that the application of tree-based integration algorithms in traffic prediction is still limited. This study aims to improve the accuracy and interpretability of the models by using random forest (RF) to analyze and model the travel time on freeways. Design/methodology/approach As the traffic conditions often greatly change, the prediction results are often unsatisfactory. To improve the accuracy of short-term travel time prediction in the freeway network, a practically feasible and computationally efficient RF prediction method for real-world freeways by using probe traffic data was generated. In addition, the variables’ relative importance was ranked, which provides an investigation platform to gain a better understanding of how different contributing factors might affect travel time on freeways. Findings The parameters of the RF model were estimated by using the training sample set. After the parameter tuning process was completed, the proposed RF model was developed. The features’ relative importance showed that the variables (travel time 15 min before) and time of day (TOD) contribute the most to the predicted travel time result. The model performance was also evaluated and compared against the extreme gradient boosting method and the results indicated that the RF always produces more accurate travel time predictions. Originality/value This research developed an RF method to predict the freeway travel time by using the probe vehicle-based traffic data and weather data. Detailed information about the input variables and data pre-processing were presented. To measure the effectiveness of proposed travel time prediction algorithms, the mean absolute percentage errors were computed for different observation segments combined with different prediction horizons ranging from 15 to 60 min.

Bus arrival time prediction and reliability analysis: An experimental comparison of functional data analysis and Bayesian support vector regression

To maintain the stability and punctuality of bus systems, an accurate forecast of arrival time is essential to devise control strategies to prevent bus bunching especially under congested traffic conditions. Transit agencies provide travelers with accurate and reliable bus arrival times to downstream stations to improve transit service quality so as to attract more transit riders. Varieties of approaches have been dedicated to providing high prediction accuracy while the measure of the associated uncertainty is ignored. Noting that the quantification of uncertainty is vital for robust performance, this paper proposes data-driven approaches based on the Functional Data Analysis (FDA) and Bayesian Support Vector Regression (BSVR) for short-term bus travel time prediction while anticipating various uncertainties. To capture spatial–temporal dynamic traffic conditions along the route so as to increase the accuracy of the journey time prediction and to capture the skewness in journey time distribution, a probabilistic nested delay operator is adopted. Journey time reliability analysis is then conducted using the skewness of dynamic journey time distribution. An empirical study is carried out by fusing the bus transit date of No. 261 bus route and Floating Car Data (FCD) in Guangzhou. The proposed FDA and BSVR methods applied in conjunction with the probabilistic nested delay operator turn out to be highly competitive when performing forecasts under various traffic conditions. Comparative studies indicate that FDA provides more accurate prediction results and tends to anticipate uncertainties in journey time distribution more effectively.

Machine Learning Based Short-Term Travel Time Prediction: Numerical Results and Comparative Analyses

Due to the increasing traffic volume in metropolitan areas, short-term travel time prediction (TTP) can be an important and useful tool for both travelers and traffic management. Accurate and reliable short-term travel time prediction can greatly help vehicle routing and congestion mitigation. One of the most challenging tasks in TTP is developing and selecting the most appropriate prediction algorithm using the available data. In this study, the travel time data was provided and collected from the Regional Integrated Transportation Information System (RITIS). Then, the travel times were predicted for short horizons (ranging from 15 to 60 min) on the selected freeway corridors by applying four different machine learning algorithms, which are Decision Trees (DT), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Long Short-Term Memory neural network (LSTM). Many spatial and temporal characteristics that may affect travel time were used when developing the models. The performance of prediction accuracy and reliability are compared. Numerical results suggest that RF can achieve a better prediction performance result than any of the other methods not only in accuracy but also with stability.

Freeway Short-Term Travel Time Prediction Based on LightGBM Algorithm

In order to realize the prediction of freeway travel time, a short-term travel time prediction model based on LightGBM (Light Gradient Boosting Machine) is proposed under the influence of weather factors, time period factors, and traffic factors. These factors are called as the features used for increase prediction accuracy. The travel time of a single vehicle is determined by license plate recognition data of two adjacent video monitors in Shaoxing section of Shanghai-Hangzhou-Ningbo Freeway, and a better travel time data set is constructed by data preprocessing. The feature data are determined by Pearson correlation. Based on the analysis of main optimization parameters in LightGBM, the short-term average travel time is predicted, the MAE (Mean Absolute Error) and MAPE (Mean Absolute Percentage Error) obtained by experiments are satisfying, indicating that LightGBM model has high accuracy and good fit. Finally, through comparison with KNN (K-Nearest Neighbor) model and GBDT (Gradient Boosting Decision Tree) model, the prediction accuracy and training speed both show that LightGBM has good advantages in predicting short-term freeway travel time.

Short-term prediction of freeway travel times by fusing input-output vehicle counts and GPS tracking data

ABSTRACT Short-term travel time prediction on freeways is the most valuable information for drivers when selecting their routes and departure times. Furthermore, this information is also essential at traffic management centers in order to monitor the network performance and anticipate the activation of traffic management strategies. The importance of reliable short-term travel time predictions will even increase with the advent of autonomous vehicles, when vehicle routing will strongly rely on this information. In this context, it is important to develop a real-time method to accurately predict travel times. The present paper uses vehicle accumulation, obtained from input-output diagrams constructed from loop detector data, to predict travel times on freeway sections. Loop detector count drift, which typically invalidates vehicle accumulation measurements, is corrected by means of a data fusion algorithm using GPS measurements. The goodness of the methodology has been proven under different boundary conditions using simulated data.

A dynamic approach to predict travel time in real time using data driven techniques and comprehensive data sources

Having access to accurate travel time is of great importance for both highway network users and traffic engineers. The travel time which is currently reported on highways is usually estimated by employing naïve methods and using limited sources of data. This could result in unreliable and inaccurate travel time prediction which causes inconvenience for travelers. Despite abundant effort on predicting travel time, the dynamic nature of travel time was neglected in many studies. Therefore, the main objective of this study is to predict the short-term travel time of highways dynamically, using multiple data sources including loop detectors, probe vehicles, weather condition, geometry, accidents, road works, special events, and sun glare. To this end, three well known machine learning methods, Artificial Neural Network, K-Nearest Neighbors and Random Forest, are employed to predict and compare short-term travel time for prediction horizons of 5 min in one hour ahead. The results confirm satisfying performance of models, especially Random Forest, in short-term travel time prediction. Our analyses also show that traffic variables, especially occupancy, are the most effective variables in predicting the travel time. Finally, we proposed an approach for a dynamic prediction of travel time for a corridor. Interestingly, the results of our dynamic approach improved the accuracy of travel time prediction over the snapshot travel time prediction.

Urban Traffic Travel Time Short-Term Prediction Model Based on Spatio-Temporal Feature Extraction

In order to improve the accuracy of short-term travel time prediction in an urban road network, a hybrid model for spatio-temporal feature extraction and prediction of urban road network travel time is proposed in this research, which combines empirical dynamic modeling (EDM) and complex networks (CN) with an XGBoost prediction model. Due to the highly nonlinear and dynamic nature of travel time series, it is necessary to consider time dependence and the spatial reliance of travel time series for predicting the travel time of road networks. The dynamic feature of the travel time series can be revealed by the EDM method, a nonlinear approach based on Chaos theory. Further, the spatial characteristic of urban traffic topology can be reflected from the perspective of complex networks. To fully guarantee the reasonability and validity of spatio-temporal features, which are dug by empirical dynamic modeling and complex networks (EDMCN), for urban traffic travel time prediction, an XGBoost prediction model is established for those characteristics. Through the in-depth exploration of the travel time and topology of a particular road network in Guiyang, the EDMCN-XGBoost prediction model’s performance is verified. The results show that, compared with the single XGBoost, autoregressive moving average, artificial neural network, support vector machine, and other models, the proposed EDMCN-XGBoost prediction model presents a better performance in forecasting.

Improving Coverage Rate for Urban Link Travel Time Prediction Using Probe Data in the Low Penetration Rate Environment.

Short-term travel time prediction is an important consideration in modern traffic control and management systems. As probe data technology has developed, research interest has moved from highways to urban roads. Most research has only focused on improving the prediction accuracy on urban roads because it is the key index of evaluating a model. However, the low penetration rate of probe vehicles at urban networks may result in the low coverage rate which restricts prediction models from practical applications. This research proposed a non-parametric model based on Bayes’ theorem and a resampling process to predict short-term urban link travel time, which can enhance the coverage rate while maintaining the prediction accuracy. The proposed model used data from vehicles in both the target link and its crossing direction, so its coverage rate can be expanded, especially when the data penetration rate is low. In addition, the utilization of relationships between vehicles in both directions can reflect the influence of signal timing. The proposed model was evaluated in a computer simulation to test its robustness and reliability under different data penetration rates. The results implied that the proposed model has a high coverage rate, demonstrating stable and acceptable performance at different penetration rates.

Short-term prediction of speed contour map and travel time by matching similar spatiotemporal traffic patterns

Short-term prediction of speed contour map and travel time by matching similar spatiotemporal traffic patterns

Long-term travel time prediction using gradient boosting

Reliable long-term travel time prediction would be effective support to traffic management, for example, traffic flow control or the pricing of tolls. Gradient boosting (GB) has been suggested as an excellent tool for short-term travel time prediction problems. This paper shows that GB with modifications can also work for long-term prediction. We introduce key variables, regarded as multiple factors, from various sources into a GB model and apply the Fourier filtering process to reduce noise. Those key variables include time of the day, day of the week, holidays, big events or activities, promotion of tolls, and narrowing of roadways. This paper takes the first step in applying key variables (from various sources) to predict long-term travel time. The electronic toll collection (ETC) data of Taiwan Freeway No.1 are used to train and to test in our process. Results demonstrate that the prediction ability of the GB model with Fourier filtering is the best. This paper shows a research direction of long-term travel time prediction, and the relatively important key variables.

Short-Term Urban Link Travel Time Prediction Using Dynamic Time Warping With Disaggregate Probe Data

There is increasing demand for short-term urban link travel time prediction to build an advanced intelligent transportation system (ITS). With the development of data collection technology, probe data are receiving more attention but the penetration rate of probe vehicles capable of sending probe data is still limited. Most research pertaining to short-term travel time prediction tends to aggregate probe data to obtain useful samples when the penetration rate is low. However, as a result, the prediction can only provide a general description of the travel time and changes in travel time during a short time interval are neglected. To overcome this limitation, a non-parametric model using disaggregate probe data based on dynamic time warping (DTW) was developed in this study. Data from the crossing direction are introduced to separate the data into different signal phases instead of identifying the exact signal pattern. A classical k-nearest neighbor (KNN) model and a naive model were compared with the proposed model. The models were tested in three scenarios: a computer simulation and two real cases from Nagoya, Japan. The results showed that the proposed model outperforms the other two models under different data penetration rates because it can reflect changes in travel time during a traffic signal cycle. Moreover, the proposed model has wider applicability than the KNN model because it is free from the equal time interval constraint.

A hybrid method for short-term freeway travel time prediction based on wavelet neural network and Markov chain

Short-term travel time prediction is an essential input to intelligent transportation systems. Timely and accurate traffic forecasting is necessary for advanced traffic management systems and advanced traveler information systems. Despite several short-term travel time prediction approaches have been proposed in the past decade, especially for hybrid models that consist of machine learning models and statistical models, few studies focus on the over-fitting problem brought by hybrid models. The over-fitting problem deteriorates the prediction accuracy especially during peak hours. This paper proposes a hybrid model that embraces wavelet neural network (WNN), Markov chain (MAR), and the volatility (VOA) model for short-term travel time prediction in a freeway system. The purpose of this paper is to provide deeper insights into underlining dynamic traffic patterns and to improve the prediction accuracy and robustness. The method takes periodical analysis, error correction, and noise extraction into consideration and improve the forecasting performance in peak hours. The proposed methodology predicts travel time by decomposing travel time data into three components: a periodic trend presented by a modified WNN, a residual part modeled by Markov chain, and the volatility part estimated by the modified generalized autoregressive conditional heteroscedasticity model. Forecasting performance is investigated with freeway travel time data from Houston, Texas and examined by three measures: mean absolute error, mean absolute percentage error, and root mean square error. The results show that the travel times predicted by the WNN-MAR-VOA method are robust and accurate. Meanwhile, the proposed method is able to capture the underlying periodic characteristics and volatility nature of travel time data.

Short-term travel-time prediction on highway: A review on model-based approach

Emerging technologies provide a venue on which on-line traffic controls and management systems can be implemented. For such applications, having access to accurate predictions on travel-times are mandatory for their successful operations. Transportation engineers have developed numerous approaches including model-based approaches. The model-based approaches consider underlying traffic mechanisms and behaviors in developing the prediction procedures and they are logically intuitive unlike datadriven approaches. Because of this explanation power, the model-based approaches have been developed for the on-line control purposes. For departments of transportation (DOTs), it is still a challenge to choose a specific approach that meets their requirements. In efforts to develop a unique guideline for transportation engineers and decision makers when considering for implementing modelbased approaches for highways, this paper reviews model-based travel-time prediction approaches by classifying them into four categories according to the level of details involved in the model: Macroscopic, Mesoscopic, CA-based, and Microscopic. Then each method is evaluated from five main perspectives: Prediction range, Accuracy, Efficiency, Applicability, and Robustness. Finally, this paper concludes with evaluations of model-based approaches in general and discusses them in relation to data-driven approaches along with future research directions.

The influence of alternative data smoothing prediction techniques on the performance of a two-stage short-term urban travel time prediction framework

ABSTRACTThis article investigates the impact of alternative data smoothing and traffic prediction methods on the accuracy of the performance of a two-stage short-term urban travel time prediction framework. Using this framework, we test the influence of the combination of two different data smoothing and four different prediction methods using travel time data from two substantially different urban traffic environments and under both normal and abnormal conditions. This constitutes the most comprehensive empirical evaluation of the joint influence of smoothing and predictor choice to date. The results indicate that the use of data smoothing improves prediction accuracy regardless of the prediction method used and that this is true in different traffic environments and during both normal and abnormal (incident) conditions. Moreover, the use of data smoothing in general has a much greater influence on prediction performance than the choice of specific prediction method, and this is independent of the specific smoothing method used. In normal traffic conditions, the different prediction methods produce broadly similar results but under abnormal conditions, lazy learning methods emerge as superior.

Highway travel time estimation using multiple data sources

Travel time is considered the most useful travel related information as it is the best indicator of the level of service on the road stretch and is completely understandable to all users. Various technologies for measuring traffic flow parameters provide the optimal background for the implementation of data fusion schemes to gain the maximum accuracy from the combination of the available data. The objective of the data fusion is to gain knowledge of predicted departure based travel time from the two outdated accurate measurements. In this paper a new and simple algorithm is proposed for short-term highway travel time prediction by fusing direct travel time measurements estimated by vehicle reidentification, indirect travel time estimated by the extrapolation of spot speed measurements and additional qualitative data in terms of the level of service. The proposed algorithm has been in operation on the A1 highway in Slovenia for more than two years and has shown robust behaviour in the real world environment. The algorithm is capable of providing short-term travel time prediction in real time with a 9 % better accuracy than the presently used travel time prediction algorithms.