kth.sePublications
Change search
Link to record
Permanent link

Direct link
Koutsopoulos, Haris N.ORCID iD iconorcid.org/0000-0003-3830-9794
Publications (10 of 88) Show all publications
Zhang, P., Koutsopoulos, H. N. & Ma, Z. (2023). DeepTrip: A Deep Learning Model for the Individual Next Trip Prediction With Arbitrary Prediction Times. IEEE transactions on intelligent transportation systems (Print), 24(6), 5842-5855
Open this publication in new window or tab >>DeepTrip: A Deep Learning Model for the Individual Next Trip Prediction With Arbitrary Prediction Times
2023 (English)In: IEEE transactions on intelligent transportation systems (Print), ISSN 1524-9050, E-ISSN 1558-0016, Vol. 24, no 6, p. 5842-5855Article in journal (Refereed) Published
Abstract [en]

The increasing availability of travel trajectory data allows for a better understanding of travel behavior. In the individual mobility analysis, the problem of next trip prediction assumes a central role and is beneficial for applications such as personalized services and mobility management. This paper addresses the next trip prediction problem with arbitrary prediction times (the time when the prediction is made). This problem has not been studied adequately in the literature and it is important for applications driven by system events, such as proactive travel recommendations under disruptions or crowding in transport systems. It predicts an individual's next trips given their historical trip sequences and the prediction time. We formulate the next trip prediction problem as on-board and off-board predictions depending on an individual's travel status (i.e. on-board/off-board). Using historical/real-time travel trajectories, a DeepTrip model is proposed based on a trip sequence-to-sequence deep learning structure coupled with an attention mechanism. A novel overlapped embedding method is proposed to represent continuous travel attributes capturing simultaneously the categorical and numerical feature information. We also develop a random-sampling training algorithm to learn the impact of the prediction time. The model is validated using trip data in urban rails. The results show that DeepTrip outperforms statistical-based models by more than 10% in terms of accuracy and other deep learning models by 2%-3%. The impact analysis shows that different representations are appropriate for the two prediction cases (on-board/off-board), and the prediction performance does not monotonically improve as the prediction time approaches the next trip.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2023
Keywords
deep learning, individual mobility, metro systems, Next trip prediction
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-330939 (URN)10.1109/TITS.2023.3252043 (DOI)000995467900001 ()2-s2.0-85151512857 (Scopus ID)
Note

QC 20230704

Available from: 2023-07-04 Created: 2023-07-04 Last updated: 2024-01-10Bibliographically approved
Tympakianaki, A., Koutsopoulos, H. N. & Jenelius, E. (2019). Anatomy of tunnel congestion: Causes and implications for tunnel traffic management. Tunnelling and Underground Space Technology, 83, 498-508
Open this publication in new window or tab >>Anatomy of tunnel congestion: Causes and implications for tunnel traffic management
2019 (English)In: Tunnelling and Underground Space Technology, ISSN 0886-7798, E-ISSN 1878-4364, Vol. 83, p. 498-508Article in journal (Refereed) Published
Abstract [en]

Tunnel congestion is an important safety problem and is often dealt with using disruptive traffic management strategies, such as closures. The paper proposes an approach to identify the underlying causes of recurrent congestion in tunnels and tests the hypothesis that the cause may vary from day to day. It also suggests that the appropriate tunnel management strategy to deploy depends on the cause. Utilizing traffic sensor data the approach consists of: (i) cluster analysis of historical traffic data to identify distinct congestion patterns; (ii) in-depth analysis of the underlying demand patterns and associated bottlenecks; (iii) simulation to evaluate alternative strategies for each demand pattern; (iv) on-line classification analysis which is able to identify, in real time, the emerging congestion pattern, and inform the type of mitigation strategy to be implemented. The methodology is demonstrated for a congested tunnel in Stockholm, Sweden revealing two different spatio-temporal congestion patterns. The results show that, if the current strategy of closures is to be used, the timing should depend on the congestion pattern. However, metering is the most promising strategy. The on-line classification of the emerging congestion pattern is effective and can inform appropriate strategy proactively. The analysis emphasizes that the effectiveness of tunnel traffic management can be increased by identifying the causes of congestion on a given day.

Place, publisher, year, edition, pages
PERGAMON-ELSEVIER SCIENCE LTD, 2019
Keywords
Tunnel traffic management, Data-driven analysis, Clustering, Simulation
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-242252 (URN)10.1016/j.tust.2018.10.015 (DOI)000454963800043 ()2-s2.0-85056468137 (Scopus ID)
Note

QC 20190130

Available from: 2019-01-30 Created: 2019-01-30 Last updated: 2024-01-04Bibliographically approved
Burghout, W. & Koutsopoulos, H. (2019). Hybrid Traffic Simulation Models: Vehicle loading at meso-micro boundaries. In: Transport Simulation: (pp. 27-42). CRC Press
Open this publication in new window or tab >>Hybrid Traffic Simulation Models: Vehicle loading at meso-micro boundaries
2019 (English)In: Transport Simulation, CRC Press , 2019, p. 27-42Chapter in book (Other academic)
Abstract [en]

Traffic simulation models, especially microscopic ones, are becoming increasingly popular and are being used to address a wide range of problems, from planning to operations. However, for applications with large-scale networks, microscopic models are impractical because of input data and calibration requirements. Hybrid models that combine simulation models at different levels of detail have the potential to address these practical issues. This chapter presents a framework for implementing meso-micro hybrid models which facilitates a consistent representation of traffic dynamics. Furthermore, the chapter carries out a detailed examination of an important element impacting the consistent representation of traffic dynamics, i.e., the loading of vehicles from the meso- to the micro-model. A new loading method is presented demonstrating a superior performance as compared to existing approaches. The method is useful not only in the context of hybrid models, but also for microscopic models on their own. A case study illustrates the importance of the method in improving the fidelity of both hybrid and pure microscopic models.

Place, publisher, year, edition, pages
CRC Press, 2019
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-10751 (URN)10.1201/9780429093258-2 (DOI)2-s2.0-85075690794 (Scopus ID)9781420095098 (ISBN)9781439808016 (ISBN)
Note

QC 20111109

Available from: 2009-07-16 Created: 2009-07-16 Last updated: 2024-03-18Bibliographically approved
Koutsopoulos, H. N. & Wang, Z. (2019). Simulation of urban rail operations: Models and calibration methodology. In: Transport Simulation: Beyond Traditional Approaches: (pp. 153-170). Taylor & Francis Group
Open this publication in new window or tab >>Simulation of urban rail operations: Models and calibration methodology
2019 (English)In: Transport Simulation: Beyond Traditional Approaches, Taylor & Francis Group, 2019, p. 153-170Chapter in book (Other academic)
Abstract [en]

SimMETRO is a microscopic, dynamic, stochastic simulator of urban rail operations (METRO), specifically designed for service performance analysis, and the evaluation of operations and strategies for a real-time control of subway systems. SimMETRO employs a detailed representation of the network, rolling stock, signal control, demand, schedule and dispatching. In particular, the various sources of stochasticity in operations are explicitly captured. This chapter presents approaches for the calibration of model parameters and input (such as dynamic arrival and alighting rates), and a case study illustrates the applicability of the model and the proposed calibration methodology. After calibration, the RMSE of block run times (relative to actual times as reported by the train detection system) was reduced by 60% as compared to the default values. The calibrated demand was in good agreement with recent counts of arrival rates at the stations.

Place, publisher, year, edition, pages
Taylor & Francis Group, 2019
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-268291 (URN)10.1201/9780429093258-10 (DOI)2-s2.0-85075660920 (Scopus ID)
Note

QC 20200318

Available from: 2020-03-18 Created: 2020-03-18 Last updated: 2024-01-04Bibliographically approved
Balakrishna, R., Ben-Akiva, M. & Koutsopoulos, H. N. (2019). Time-dependent origin-destination estimation without assignment matrices. In: Transport Simulation: Beyond Traditional Approaches: (pp. 201-213). Taylor & Francis Group
Open this publication in new window or tab >>Time-dependent origin-destination estimation without assignment matrices
2019 (English)In: Transport Simulation: Beyond Traditional Approaches, Taylor & Francis Group, 2019, p. 201-213Chapter in book (Other academic)
Abstract [en]

Time-dependent origin-destination (OD) flows are crucial inputs to dynamic traffic assignment (DTA) models. However, they are often unobserved, and must be estimated from indirect traffic measurements collected from the study network. Approaches to estimate OD flows from link counts traditionally rely on assignment matrices that map the OD flow variables onto the counts. However, this method (a) approximates the complex relationship between OD flows and counts with a linear function, (b) is restricted to the use of only counts, and cannot exploit richer data such as speeds, densities or travel times, and (c) cannot estimate route choice and supply parameters that critically impact the OD estimates. This chapter presents a dynamic OD estimation method that is accurate and flexible in the use of general traffic data. Moreover, it simultaneously estimates all parameters with an impact on OD estimation, and can be applied to any traffic assignment model.

Place, publisher, year, edition, pages
Taylor & Francis Group, 2019
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-268290 (URN)10.1201/9780429093258-12 (DOI)2-s2.0-85075676214 (Scopus ID)
Note

QC 20200318

Available from: 2020-03-18 Created: 2020-03-18 Last updated: 2024-01-04Bibliographically approved
Tympakianaki, A., Koutsopoulos, H. N., Jenelius, E. & Cebecauer, M. (2018). Impact analysis of transport network disruptions using multimodal data: A case study for tunnel closures in Stockholm. Case Studies on Transport Policy, 6(2), 179-189
Open this publication in new window or tab >>Impact analysis of transport network disruptions using multimodal data: A case study for tunnel closures in Stockholm
2018 (English)In: Case Studies on Transport Policy, ISSN 2213-624X, E-ISSN 2213-6258, Vol. 6, no 2, p. 179-189Article in journal (Refereed) Published
Abstract [en]

The paper explores the utilization of heterogeneous data sources to analyze the multimodal impacts of transport network disruptions. A systematic data-driven approach is proposed for the analysis of impacts with respect to two aspects: (a) spatiotemporal network changes, and (b) multimodal effects. The feasibility and benefits of combining various data sources are demonstrated through a case study for a tunnel in Stockholm, Sweden which is often prone to closures. Several questions are addressed including the identification of impacted areas, and the evaluation of impacts on network performance, demand patterns and performance of the public transport system. The results indicate significant impact of tunnel closures on the network traffic conditions due to the redistribution of vehicles on alternative paths. Effects are also found on the performance of public transport. Analysis of the demand reveals redistribution of traffic during the tunnel closures, consistent with the observed impacts on network performance. Evidence for redistribution of travelers to public transport is observed as a potential effect of the closures. Better understanding of multimodal impacts of a disruption can assist authorities in their decision-making process to apply adequate traffic management policies.

Place, publisher, year, edition, pages
ELSEVIER SCIENCE BV, 2018
Keywords
Transport system disruptions, Data-driven analysis
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-231201 (URN)10.1016/j.cstp.2018.05.003 (DOI)000434260300001 ()2-s2.0-85047071116 (Scopus ID)
Note

QC 20180629

Available from: 2018-06-29 Created: 2018-06-29 Last updated: 2024-01-04Bibliographically approved
Tympakianaki, A., Koutsopoulos, H. N. & Jenelius, E. (2018). Robust SPSA algorithms for dynamic OD matrix estimation. In: The 9th International Conference on Ambient Systems, Networks and Technologies (ANT 2018) / The 8th International Conference on Sustainable Energy Information Technology (SEIT-2018) / Affiliated WorkshopsThe 9th International Conference on Ambient Systems, Networks and Technologies (ANT 2018) / The 8th International Conference on Sustainable Energy Information Technology (SEIT-2018) / Affiliated Workshops: . Paper presented at 9th International Conference on Ambient Systems, Networks and Technologies May 8-11, 2018, Porto, Portugal (pp. 57-64). Elsevier, 130
Open this publication in new window or tab >>Robust SPSA algorithms for dynamic OD matrix estimation
2018 (English)In: The 9th International Conference on Ambient Systems, Networks and Technologies (ANT 2018) / The 8th International Conference on Sustainable Energy Information Technology (SEIT-2018) / Affiliated WorkshopsThe 9th International Conference on Ambient Systems, Networks and Technologies (ANT 2018) / The 8th International Conference on Sustainable Energy Information Technology (SEIT-2018) / Affiliated Workshops, Elsevier, 2018, Vol. 130, p. 57-64Conference paper, Published paper (Refereed)
Abstract [en]

The Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm has been used for solving the off-line dynamic origin-destination (OD) estimation problem. While the algorithm can be used with very general formulations of the problem, it can also be unstable. The paper proposes methods and evaluates their effectiveness in improving the SPSA performance at two levels: a) scaling the step size and using a hybrid gradient estimation; and b) proposing alternative clustering strategies to be used with the c-SPSA version of the algorithm, where OD flows are estimated in clusters. The proposed enhancements are evaluated through simulation experiments on a real network.

Place, publisher, year, edition, pages
Elsevier, 2018
Series
Procedia Computer Science, ISSN 1877-0509 ; 130130
Keywords
SPSA, c-SPSA, Origin-Destination (OD) matrix estimation, stochastic approximation
National Category
Transport Systems and Logistics
Research subject
Transport Science
Identifiers
urn:nbn:se:kth:diva-221856 (URN)10.1016/j.procs.2018.04.012 (DOI)000684379100007 ()2-s2.0-85051266286 (Scopus ID)
Conference
9th International Conference on Ambient Systems, Networks and Technologies May 8-11, 2018, Porto, Portugal
Note

QC 20180129

Available from: 2018-01-28 Created: 2018-01-28 Last updated: 2024-03-18Bibliographically approved
Jenelius, E. & Koutsopoulos, H. (2018). Urban network travel time prediction based on a probabilistic principal component analysis model of probe data. IEEE transactions on intelligent transportation systems (Print), 19(2), 436-445
Open this publication in new window or tab >>Urban network travel time prediction based on a probabilistic principal component analysis model of probe data
2018 (English)In: IEEE transactions on intelligent transportation systems (Print), ISSN 1524-9050, E-ISSN 1558-0016, Vol. 19, no 2, p. 436-445Article in journal (Refereed) Published
Abstract [en]

This paper proposes a network travel time prediction methodology based on probe data. The model is intended as a tool for traffic management, trip planning, and online vehicle routing, and is designed to be efficient and scalable in calibration and real-time prediction; flexible to changes in network, data, and model extensions; and robust against noisy and missing data. A multivariate probabilistic principal component analysis (PPCA) model is proposed. Spatio-temporal correlations are inferred from historical data based on MLE and an efficient EM algorithm for handling missing data. Prediction is performed in real time by computing the expected distribution of link travel times in future time intervals, conditional on recent current-day observations. A generalization of the methodology partitions the network and applies a distinct PPCA model to each subnetwork. The methodology is applied to the network of downtown Shenzhen, China, using taxi probe data. The model captures variability over months and weekdays as well as other factors. Prediction with PPCA outperforms k-nearest neighbors prediction for horizons from 15 to 45 min, and a hybrid method of PPCA and local smoothing provides the highest accuracy.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2018
Keywords
Travel time prediction, PPCA, probe data
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-219362 (URN)10.1109/TITS.2017.2703652 (DOI)000424060200011 ()2-s2.0-85020405032 (Scopus ID)
Funder
Swedish Transport AdministrationTrenOp, Transport Research Environment with Novel Perspectives
Note

QC 20171212

Available from: 2017-12-04 Created: 2017-12-04 Last updated: 2024-03-18Bibliographically approved
Toledo, T., Jenelius, E., Cats, O., Koutsopoulos, H. N., Manasra, H. & Leffler, D. (2017). ADAPT‐IT: Improved transit service through real‐time operations control and traveller informationFinal report.
Open this publication in new window or tab >>ADAPT‐IT: Improved transit service through real‐time operations control and traveller informationFinal report
Show others...
2017 (English)Report (Other academic)
Publisher
p. 186
National Category
Transport Systems and Logistics
Identifiers
urn:nbn:se:kth:diva-333759 (URN)
Funder
Vinnova, 2014-03874
Note

QC 20230811

Available from: 2023-08-10 Created: 2023-08-10 Last updated: 2024-01-04Bibliographically approved
Tympakianaki, A., Koutsopoulos, H. N. & Jenelius, E. (2017). Anatomy of tunnel congestion: causes and implications for tunnel traffic management.
Open this publication in new window or tab >>Anatomy of tunnel congestion: causes and implications for tunnel traffic management
2017 (English)Manuscript (preprint) (Other academic)
Abstract [en]

Tunnel congestion is an important safety problem and is often dealt with using disruptive traffic management strategies, such as closures. The paper proposes an approach to identify the underlying causes of recurrent congestion in tunnels and tests the hypothesis that the cause may vary from day to day. It also suggests that the appropriate tunnel management strategy to deploy depends on the cause. Utilizing traffic sensor data the approach consists of: (i) cluster analysis of historical traffic data to identify distinct congestion patterns; (ii) in-depth analysis of the underlying demand patterns and associated bottlenecks; (iii) simulation to evaluate alternative strategies for each demand pattern; (iv) on-line classification analysis which is able to identify, in real time, the emerging congestion pattern, and inform the type of mitigation strategy to be implemented. The methodology is demonstrated for a congested tunnel in Stockholm, Sweden revealing two different spatiotemporal congestion patterns. The results show that, if the current strategy of closures is to be used, the timing should depend on the congestion pattern. However, metering is the most promising strategy. The on-line classification of the emerging congestion pattern is effective and can inform appropriate strategy proactively. The analysis emphasizes that the effectiveness of tunnel traffic management can be increased by identifying the causes of congestion on a given day. 

Keywords
Tunnel traffic management; data-driven analysis; clustering; simulation
National Category
Transport Systems and Logistics
Research subject
Transport Science
Identifiers
urn:nbn:se:kth:diva-221855 (URN)
Note

QC 20180129

Available from: 2018-01-28 Created: 2018-01-28 Last updated: 2024-03-18Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0003-3830-9794

Search in DiVA

Show all publications