Arterial Path-Level Travel-Time Estimation Using Machine-Learning Techniques
01 May 2017-Journal of Computing in Civil Engineering (American Society of Civil Engineers (ASCE))-Vol. 31, Iss: 3, pp 04016070
TL;DR: A methodology for travel-time prediction on urban arterial networks using data from global positioning system (GPS) probe vehicles, under Indian traffic conditions, is presented.
Abstract: This study presents a methodology for travel-time prediction on urban arterial networks using data from global positioning system (GPS) probe vehicles, under Indian traffic conditions. Give...
Citations
More filters
TL;DR: The new concept of consensual 3D speed maps allows the essence out of large amounts of link speed observations and reveals a global and previously mostly hidden picture of traffic dynamics at the whole city scale, which may be more regular and predictable than expected.
Abstract: In this paper, we investigate the day-to-day regularity of urban congestion patterns. We first partition link speed data every 10 min into 3D clusters that propose a parsimonious sketch of the congestion pulse. We then gather days with similar patterns and use consensus clustering methods to produce a unique global pattern that fits multiple days, uncovering the day-to-day regularity. We show that the network of Amsterdam over 35 days can be synthesized into only 4 consensual 3D speed maps with 9 clusters. This paves the way for a cutting-edge systematic method for travel time predictions in cities. By matching the current observation to historical consensual 3D speed maps, we design an efficient real-time method that successfully predicts 84% trips travel times with an error margin below 25%. The new concept of consensual 3D speed maps allows us to extract the essence out of large amounts of link speed observations and as a result reveals a global and previously mostly hidden picture of traffic dynamics at the whole city scale, which may be more regular and predictable than expected.
221 citations
TL;DR: The Artificial Fish Swarm algorithm is applied to optimize the SVM model parameters, which include the kernel parameter σ, non-sensitive loss function parameter ε, and penalty parameter C, and the results show that the accuracy of the optimized S VM model is 17.27% and 16.44% higher than those of the BP neural network model and the common SVM models, respectively.
Abstract: Freeway travel time is influenced by many factors including traffic volume, adverse weather, accidents, traffic control, and so on. We employ the multiple source data-mining method to analyze freeway travel time. We collected toll data, weather data, traffic accident disposal logs, and other historical data from Freeway G5513 in Hunan Province, China. Using the Support Vector Machine (SVM), we proposed the travel time predicting model founded on these databases. The new SVM model can simulate the nonlinear relationship between travel time and those factors. In order to improve the precision of the SVM model, we applied the Artificial Fish Swarm algorithm to optimize the SVM model parameters, which include the kernel parameter σ, non-sensitive loss function parameter e, and penalty parameter C. We compared the new optimized SVM model with the Back Propagation (BP) neural network and a common SVM model, using the historical data collected from freeway G5513. The results show that the accuracy of the optimized SVM model is 17.27% and 16.44% higher than those of the BP neural network model and the common SVM model, respectively.
9 citations
DOI•
27 Feb 2020
TL;DR: This thesis develops a series of efficient data-driven methods for extracting the mobility patterns of large-scale metropolitan networks and explores some of their applications.
Abstract: Cities are complex, dynamic and ever-evolving. We need to understand how these cities work in order to predict, control or optimize their operations. We have identified some open issues related to network and data complexity that need to be solved to build feasible methods for these purposes. To this end, we first build multiscale graphs automatically to address a problem that is becoming increasingly relevant in the age of big data, where reducing the network complexity could easily determine the viability of the research in real-world applications. Next, we propose different methods from different fields to extract the essence of network dynamics from the vast amount of spatiotemporal traffic data. One such method is a new way of looking at traffic patterns, combining the field of pattern recognition - with a focus on computer vision - with the traffic domain. The inspiration comes from the fact that humans are the most sophisticated pattern recognizer in the world and we use specific visual features to recognize different complex patterns and we explore if these features can also be used to recognize traffic patterns. Finally, we explore different applications of such mobility patterns such as revealing the unknown correlation between supply and demand patterns, evaluating the scalability of the proposed approach by applying the method to the entire Dutch highway network and transferability by building similar network patterns for public transport networks. Thus, this thesis develops a series of efficient data-driven methods for extracting the mobility patterns of large-scale metropolitan networks and explore some of their applications. With the increasing availability of data in the transport domain, the Achilles heel is not data scarcity anymore but rather extracting insights from this massive amount of data. This thesis is a step forward in solving this complex problem by leveraging the increased acceptance of using machine learning as a worthy and effective method for network-wide analysis of traffic patterns.
8 citations
Cites background from "Arterial Path-Level Travel-Time Est..."
...parameterized mathematical models such as (generalized) linear regression[137, 138]; kriging [139]; support vector regression [140]; random forest[141]; Bayesian networks[142]; artificial neural networks, e....
[...]
TL;DR: The root-cause analysis revealed a lack of effective collaboration between the ML and transportation experts, resulting in the most accessible transportation applications being used as a case study to test or enhance a given ML algorithm and not necessarily to enhance a mobility or safety issue.
Abstract: Surface transportation has evolved through technology advancements using parallel knowledge areas such as machine learning (ML). However, the transportation industry has not yet taken full advantage of ML. To evaluate this gap, we utilized a literature review approach to locate, categorize, and synthesize the principal concepts of research papers regarding surface transportation systems using ML algorithms, and we then decomposed them into their fundamental elements. We explored more than 100 articles, literature review papers, and books. The results show that 74% of the papers concentrate on forecasting, while multilayer perceptions, long short-term memory, random forest, supporting vector machine, XGBoost, and deep convolutional neural networks are the most preferred ML algorithms. However, sophisticated ML algorithms have been minimally used. The root-cause analysis revealed a lack of effective collaboration between the ML and transportation experts, resulting in the most accessible transportation applications being used as a case study to test or enhance a given ML algorithm and not necessarily to enhance a mobility or safety issue. Additionally, the transportation community does not define transportation issues clearly and does not provide publicly available transportation datasets. The transportation sector must offer an open-source platform to showcase the sector’s concerns and build spatiotemporal datasets for ML experts to accelerate technology advancements.
4 citations
TL;DR: The performance of the proposed Advanced Time-Space Discterization (AdTSD) method was evaluated with real field data and compared with existing approaches and results show that AdTSD approach was able to perform better than historical average approach with an advantage up to 11% and 5% compared to Base Time Space Discretization (B TSD) approach.
Abstract: Travel time is a variable that varies over both time and space. Hence, an ideal formulation should be able to capture its evolution over time and space. A mathematical representation capturing such variations was formulated from first principles, using the concept of conservation of vehicles. The availability of position and speed data obtained from GPS enabled buses provide motivation to rewrite the conservation equation in terms of speed alone. As the number of vehicles is discrete, the speed-based equation was discretized using Godunov scheme and used in the prediction scheme that was based on the Kalman filter. With a limited fleet size having an average headway of 30 min, availability of travel time data at small interval that satisfy the requirement of stability of numerical solution possess a big challenge. To address this issue, a continuous speed fill matrix spatially and temporally was developed with the help of historic data and used in this study. The performance of the proposed Advanced Time-Space Discterization (AdTSD) method was evaluated with real field data and compared with existing approaches. Results show that AdTSD approach was able to perform better than historical average approach with an advantage up to 11% and 5% compared to Base Time Space Discretization (BTSD) approach. Also, from the results it was observed that the maximum deviation in prediction was in the range of 2–3 min when it is predicted 10 km ahead and the error is close to zero when it is predicted a section ahead i.e. when the bus is close to a bus stop, indicating that the prediction accuracy achieved is suitable for real field implementation.
3 citations
References
More filters
TL;DR: This paper examines how real-time information gathered as part of intelligent transportation systems can be used to predict link travel times for one through five time periods ahead and found that the spectral basis artificial neural network (SNN) gave the best overall results.
Abstract: This paper examines how real-time information gathered as part of intelligent transportation systems can be used to predict link travel times for one through five time periods ahead (of 5-min duration). The study employed a spectral basis artificial neural network (SNN) that utilizes a sinusoidal transformation technique to increase the linear separability of the input features. Link travel times from Houston that had been collected as part of the automatic vehicle identification system of the TranStar system were used as a test bed. It was found that the SNN outperformed a conventional artificial neural network and gave similar results to that of modular neural networks. However, the SNN requires significantly less effort on the part of the modeler than modular neural networks. The results, of the best SNN were compared with conventional link travel time prediction techniques including a Kalman filtering model, exponential smoothing model, historical profile, and real-time profile. It was found that the SNN gave the best overall results.
245 citations
TL;DR: The main object of the following paper is to describe and illustrate some approximation methods that can be used to obtain rough estimates of queue lengths, delays, etc., for various queueing problemis, particularly highway traffic intersection problems, which may be too difficult to solve exactly, or if solved exactly give formulas that are more detailed than one needs for the purpose of making quick estimates.
Abstract: Approximations based upon the representation of the queue as a continuous fluid with either deterministic or stochastic properties are applied to the analysis of models of a fixed-cycle traffic light. These approximations are based upon the use of a law of large numbers or a central limit theorem and are not very sensitive to the detailed stochastic structure of the arrival or departure processes. In the applications considered here these approximations give delays correct to within a few percent. Introduction. The main object of the following paper is to describe and illustrate some approximation methods that can be used to obtain rough estimates of queue lengths, delays, etc., for various queueing problemis, particularly highway traffic intersection problems, which may be too difficult to solve exactly, or if solved exactly give formulas that are more detailed than one needs for the purpose of making quick estimates. The types of approximation we have in mind are those that will apply when the average queue lengths are much larger than 1. They will be asymptotic approximations that strictly speaking are valid only in the limiit of infinite queues, but ones which still give rough estimates for finite but large queues (perhaps of the order of 10). There is an extensive literature on queueing theory including at least a half dozen books and several hundred papers, but the vogue in queueing theory has been to obtain exact solutions of highly idealized models of various processes (exact, however, only in the sense that one usually must invert a few generating functions or Laplace transforms to obtain the quantities one really wants). Consequently the practical value of queueing theory has been severely limited by the lack of approximation methods which one can use to analyse more difficult problems, to estimate errors introduced by the model, or even to compute numbers from very cumbersome exact formulas. A few attempts, however, have been mnade in recent years to break away from the now standard analytic techniques. Particularly noteworthy are the works of Kingman [1], who has shown that properties of nearly saturated queues are rather insensitive to the detailed form of the arrival or service distributions, and of Benes [2], who has shown that many of the results of queueing theory do not depend upon some of the customary statistical independence assumptions. Even here, however, mathematical elegance takes precedence over intuition or applications. In the literature on traffic theory there are already at least 20 references relating to the single problem of delays at a fixed-cycle traffic light, three quarters of which are primarily directed toward a fruitless pursuit of simple exact solutions * Received by the editors October 19, 1964, and in revised form November 19, 1964. t Division of Applied Mathematics, Brown University, Providence, Rhode Island. Part of this work was done while the author was a Fulbright professor at the University of Adelaide, South Australia. The work was also supported in part by the National Science Foundation at Brown University.
243 citations
TL;DR: A three-layer neural network model is proposed to estimate complete link travel time for individual probe vehicle traversing the link and results suggest that the Artificial Neural Network model outperforms the analytical model.
Abstract: In the urban signalized network, travel time estimation is a challenging subject especially because urban travel times are intrinsically uncertain due to the fluctuations in traffic demand and supply, traffic signals, stochastic arrivals at the intersections, etc. In this paper, probe vehicles are used as traffic sensors to collect traffic data (speeds, positions and time stamps) in an urban road network. However, due to the low polling frequencies (e.g. 1 min or 5 min), travel times recorded by probe vehicles provide only partial link or route travel times. This paper focuses on the estimation of complete link travel times. Based on the information collected by probe vehicles, a three-layer neural network model is proposed to estimate complete link travel time for individual probe vehicle traversing the link. This model is discussed and compared with an analytical estimation model which was developed by Hellinga et al. (2008). The performance of these two models are evaluated with data derived from VISSIM simulation model. Results suggest that the Artificial Neural Network model outperforms the analytical model.
218 citations
TL;DR: An analytical formulation based on conditional probability distributions is developed for estimating the expected queue length and its variance and it is found that, for the given settings, only the location information of the last probe vehicle in the queue is sufficient for the estimation.
203 citations
TL;DR: This research provides new possibilities for fully utilizing the partial information obtained from urban taxicab data for estimating network condition, which is not only very useful but also is inexpensive and has much better coverage than traditional sensor data.
Abstract: Taxicabs equipped with Global Positioning System (GPS) devices can serve as useful probes for monitoring the traffic state in an urban area. This paper presents a new descriptive model for estimating hourly average of urban link travel times using taxicab origin–destination (OD) trip data. The focus of this study is to develop a methodology to estimate link travel times from OD trip data and demonstrate the feasibility of estimating network condition using large-scale geo-location data with partial information. The data, collected from the taxicabs in New York City, provides the locations of origins and destinations, travel times, fares and other information of taxi trips. The new model infers the possible paths for each trip and then estimates the link travel times by minimizing the error between the expected path travel times and the observed path travel times. The model is evaluated using a test network from Midtown Manhattan. Results indicate that the proposed method can efficiently estimate hourly average link travel times. This research provides new possibilities for fully utilizing the partial information obtained from urban taxicab data for estimating network condition, which is not only very useful but also is inexpensive and has much better coverage than traditional sensor data.
195 citations