Detecting taxi speeding from sparse and low-sampled trajectory data

Xibo ZHOU*, Qiong LUO, Dian ZHANG, Lionel M. NI

*Corresponding author for this work

Research output: Book Chapters | Papers in Conference ProceedingsConference paper (refereed)peer-review

2 Citations (Scopus)


Taxis are a major means of public transportation in large cities, and speeding is a common problem among motor vehicles, including taxis. Unless caught by sensors or patrol officers, many speeding incidents go unnoticed, which pose potential threat to road safety. In this paper, we propose to detect speeding behaviors of individual taxis from taxi trajectory data. Such detection results are useful for driver risk analysis and road safety management. However, the taxi trajectory data are geographically sparse and the sample rate is low. Furthermore, existing methods mainly deal with the estimation of collective road speeds whereas we focus on the speeds of individual vehicles. As such, we propose to use a two-fold collective matrix factorization (CMF)-based model to estimate the individual vehicle speed. We have evaluated our method on real-world datasets, and the results show the effectiveness of our method in detecting taxi speeding behaviors.

Original languageEnglish
Title of host publicationWeb and Big Data : Second International Joint Conference, APWeb-WAIM 2018 : proceedings
EditorsYi CAI, Yoshiharu ISHIKAWA, Jianliang XU
Place of PublicationCham
PublisherSpringer-Verlag GmbH and Co. KG
Number of pages9
ISBN (Electronic)9783319968933
ISBN (Print)9783319968926
Publication statusPublished - 2018
Externally publishedYes
Event2nd Asia Pacific Web and Web-Age Information Management Joint Conference on Web and Big Data, APWeb-WAIM 2018 - Macau, China
Duration: 23 Jul 201825 Jul 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10988 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference2nd Asia Pacific Web and Web-Age Information Management Joint Conference on Web and Big Data, APWeb-WAIM 2018

Bibliographical note

This work is supported in part by the Guangdong Pre-national project 2014GKXM054 and the Guangdong Province Key Laboratory of Popular High Performance Computers 2017B030314073.


  • Collective matrix factorization
  • Speeding
  • Trajectory


Dive into the research topics of 'Detecting taxi speeding from sparse and low-sampled trajectory data'. Together they form a unique fingerprint.

Cite this