Deep representation of raw traffic data: An embed-and-aggregate framework for high-level traffic analysis

Woosung Choi, Jonghyeon Min, Taemin Lee, Kyeongseok Hyun, Taehyung Lim, Soon Young Jung

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In Intelligent Transportation Systems (ITS), it is widely used to extract a fixed-size feature vector from raw traffic data for high-level traffic analysis. In several existing works, the statistical approach has been used for extracting feature vectors, which directly extracts features by averaging speed or travel time of each vehicle. However, we can achieve a better representation by taking advantage of state-of-the-art machine learning algorithms instead of the statistical approach. In this paper, we propose a two-phase framework named embed-and-aggregate framework for extracting features from raw traffic data, and a feature extraction algorithm (Traffic2Vec) based on our framework exploiting state-of-the-art machine learning algorithms such as deep learning. We also implement a traffic flow prediction system based on Traffic2Vec as a proof-of-concept. We conducted experiments to evaluate the applicability of the proposed algorithm, and show its superior performance in comparison with the prediction system based on the statistical feature extraction method.

Original languageEnglish
Title of host publicationAdvances in Computer Science and Ubiquitous Computing - CSA-CUTE 17
PublisherSpringer Verlag
Pages1383-1390
Number of pages8
Volume474
ISBN (Print)9789811076046
DOIs
Publication statusPublished - 2018 Jan 1
EventInternational Conference on Computer Science and its Applications, CSA 2017 - Taichung, Taiwan, Province of China
Duration: 2017 Dec 182017 Dec 20

Publication series

NameLecture Notes in Electrical Engineering
Volume474
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Other

OtherInternational Conference on Computer Science and its Applications, CSA 2017
CountryTaiwan, Province of China
CityTaichung
Period17/12/1817/12/20

Fingerprint

Learning algorithms
Learning systems
Feature extraction
Travel time
Experiments
Deep learning

Keywords

  • Embedding
  • Feature extraction
  • High-level traffic analysis
  • Traffic data
  • Trajectory

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering

Cite this

Choi, W., Min, J., Lee, T., Hyun, K., Lim, T., & Jung, S. Y. (2018). Deep representation of raw traffic data: An embed-and-aggregate framework for high-level traffic analysis. In Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17 (Vol. 474, pp. 1383-1390). (Lecture Notes in Electrical Engineering; Vol. 474). Springer Verlag. https://doi.org/10.1007/978-981-10-7605-3_220

Deep representation of raw traffic data : An embed-and-aggregate framework for high-level traffic analysis. / Choi, Woosung; Min, Jonghyeon; Lee, Taemin; Hyun, Kyeongseok; Lim, Taehyung; Jung, Soon Young.

Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17. Vol. 474 Springer Verlag, 2018. p. 1383-1390 (Lecture Notes in Electrical Engineering; Vol. 474).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Choi, W, Min, J, Lee, T, Hyun, K, Lim, T & Jung, SY 2018, Deep representation of raw traffic data: An embed-and-aggregate framework for high-level traffic analysis. in Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17. vol. 474, Lecture Notes in Electrical Engineering, vol. 474, Springer Verlag, pp. 1383-1390, International Conference on Computer Science and its Applications, CSA 2017, Taichung, Taiwan, Province of China, 17/12/18. https://doi.org/10.1007/978-981-10-7605-3_220
Choi W, Min J, Lee T, Hyun K, Lim T, Jung SY. Deep representation of raw traffic data: An embed-and-aggregate framework for high-level traffic analysis. In Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17. Vol. 474. Springer Verlag. 2018. p. 1383-1390. (Lecture Notes in Electrical Engineering). https://doi.org/10.1007/978-981-10-7605-3_220
Choi, Woosung ; Min, Jonghyeon ; Lee, Taemin ; Hyun, Kyeongseok ; Lim, Taehyung ; Jung, Soon Young. / Deep representation of raw traffic data : An embed-and-aggregate framework for high-level traffic analysis. Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17. Vol. 474 Springer Verlag, 2018. pp. 1383-1390 (Lecture Notes in Electrical Engineering).
@inproceedings{472890f28ef248e3b0188cbade79c1f6,
title = "Deep representation of raw traffic data: An embed-and-aggregate framework for high-level traffic analysis",
abstract = "In Intelligent Transportation Systems (ITS), it is widely used to extract a fixed-size feature vector from raw traffic data for high-level traffic analysis. In several existing works, the statistical approach has been used for extracting feature vectors, which directly extracts features by averaging speed or travel time of each vehicle. However, we can achieve a better representation by taking advantage of state-of-the-art machine learning algorithms instead of the statistical approach. In this paper, we propose a two-phase framework named embed-and-aggregate framework for extracting features from raw traffic data, and a feature extraction algorithm (Traffic2Vec) based on our framework exploiting state-of-the-art machine learning algorithms such as deep learning. We also implement a traffic flow prediction system based on Traffic2Vec as a proof-of-concept. We conducted experiments to evaluate the applicability of the proposed algorithm, and show its superior performance in comparison with the prediction system based on the statistical feature extraction method.",
keywords = "Embedding, Feature extraction, High-level traffic analysis, Traffic data, Trajectory",
author = "Woosung Choi and Jonghyeon Min and Taemin Lee and Kyeongseok Hyun and Taehyung Lim and Jung, {Soon Young}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-981-10-7605-3_220",
language = "English",
isbn = "9789811076046",
volume = "474",
series = "Lecture Notes in Electrical Engineering",
publisher = "Springer Verlag",
pages = "1383--1390",
booktitle = "Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17",

}

TY - GEN

T1 - Deep representation of raw traffic data

T2 - An embed-and-aggregate framework for high-level traffic analysis

AU - Choi, Woosung

AU - Min, Jonghyeon

AU - Lee, Taemin

AU - Hyun, Kyeongseok

AU - Lim, Taehyung

AU - Jung, Soon Young

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In Intelligent Transportation Systems (ITS), it is widely used to extract a fixed-size feature vector from raw traffic data for high-level traffic analysis. In several existing works, the statistical approach has been used for extracting feature vectors, which directly extracts features by averaging speed or travel time of each vehicle. However, we can achieve a better representation by taking advantage of state-of-the-art machine learning algorithms instead of the statistical approach. In this paper, we propose a two-phase framework named embed-and-aggregate framework for extracting features from raw traffic data, and a feature extraction algorithm (Traffic2Vec) based on our framework exploiting state-of-the-art machine learning algorithms such as deep learning. We also implement a traffic flow prediction system based on Traffic2Vec as a proof-of-concept. We conducted experiments to evaluate the applicability of the proposed algorithm, and show its superior performance in comparison with the prediction system based on the statistical feature extraction method.

AB - In Intelligent Transportation Systems (ITS), it is widely used to extract a fixed-size feature vector from raw traffic data for high-level traffic analysis. In several existing works, the statistical approach has been used for extracting feature vectors, which directly extracts features by averaging speed or travel time of each vehicle. However, we can achieve a better representation by taking advantage of state-of-the-art machine learning algorithms instead of the statistical approach. In this paper, we propose a two-phase framework named embed-and-aggregate framework for extracting features from raw traffic data, and a feature extraction algorithm (Traffic2Vec) based on our framework exploiting state-of-the-art machine learning algorithms such as deep learning. We also implement a traffic flow prediction system based on Traffic2Vec as a proof-of-concept. We conducted experiments to evaluate the applicability of the proposed algorithm, and show its superior performance in comparison with the prediction system based on the statistical feature extraction method.

KW - Embedding

KW - Feature extraction

KW - High-level traffic analysis

KW - Traffic data

KW - Trajectory

UR - http://www.scopus.com/inward/record.url?scp=85039444045&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85039444045&partnerID=8YFLogxK

U2 - 10.1007/978-981-10-7605-3_220

DO - 10.1007/978-981-10-7605-3_220

M3 - Conference contribution

AN - SCOPUS:85039444045

SN - 9789811076046

VL - 474

T3 - Lecture Notes in Electrical Engineering

SP - 1383

EP - 1390

BT - Advances in Computer Science and Ubiquitous Computing - CSA-CUTE 17

PB - Springer Verlag

ER -