Vehicle acoustic signal signature recognition based on Karhunen-Loeve transform

Zhenshan Li*, Jianqun Wang, Xuejun Ran, Guozhong Yao

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

A method of vehicle acoustic signal signature recognition based on Karhunen-Loeve transform is proposed. First, the vehicle acoustic signal is preprocessed, and through Karhunen-Loeve transform, the signature of the signal is obtained, and the Eigen-vehicle is built. The sample to be recognized is projected onto the Eigen-vehicle subspace, and the distance between the sample signature and the Eigen-vehicle is calculated. This distance is then used as a basis for signature recognition. The experimental results revealed that the method presented in this paper for vehicle acoustic signal signature recognition is valid, and the recognition accuracy is 92.1 percent on average.

Original languageEnglish
Title of host publicationICCTP 2010
Subtitle of host publicationIntegrated Transportation Systems: Green, Intelligent, Reliable - Proceedings of the 10th International Conference of Chinese Transportation Professionals
Pages2239-2246
Number of pages8
DOIs
Publication statusPublished - 2010
Event10th International Conference of Chinese Transportation Professionals - Integrated Transportation Systems: Green, Intelligent, Reliable, ICCTP 2010 - Beijing, China
Duration: 4 Aug 20108 Aug 2010

Publication series

NameICCTP 2010: Integrated Transportation Systems: Green, Intelligent, Reliable - Proceedings of the 10th International Conference of Chinese Transportation Professionals
Volume382

Conference

Conference10th International Conference of Chinese Transportation Professionals - Integrated Transportation Systems: Green, Intelligent, Reliable, ICCTP 2010
Country/TerritoryChina
CityBeijing
Period4/08/108/08/10

Keywords

  • Acoustic techniques
  • Traffic signals
  • Vehicles

Fingerprint

Dive into the research topics of 'Vehicle acoustic signal signature recognition based on Karhunen-Loeve transform'. Together they form a unique fingerprint.

Cite this