A novel distance estimation algorithm for periodic surface vibrations based on frequency band energy percentage feature. (December 2018)

• Record Type: Journal Article
• Authors: Cao, Jiuwen; Wang, Tianlei; Shang, Luming; Wang, Jianzhong; Vong, Chi-Man; Yin, Chun; Huang, Xuegang
• Is Part Of: Mechanical systems and signal processing. Volume 113(2018)
• Page Start: 222
• Page End: 236
• Publication Date: 2018-12
• Subjects: Periodic vibration signal processing -- Distance prediction system -- Fundamental frequency estimation -- Frequency band energy percentage -- Cepstrum; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2017.10.016
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A novel propagation distance prediction method for surface periodic vibration is developed. A new fundamental frequency estimation algorithm based on cepstrum is proposed. The fundamental frequency energy percentage feature (FBEP) is developed for vibration signal representation. Analyses on the FBEP feature distribution and its relationship to propagation distance are presented. Abstract: Earth surface vibration signal processing becomes attractive recently due to its significance in source detection and localization, which can be adopted in a number of real world applications, such as footstep detection, underground pipeline network surveillance, etc. In this paper, we investigate the distance estimation problem for earth surface periodic vibration signal localization. The signal attenuation principle between the propagation distance and the signal frequency is exploited and a novel frequency band energy percentage (FBEP) feature is developed to characterize the energy distribution property within different frequency bands of different propagation distances. To obtain the fundamental frequency of periodic vibrations, the cepstrum approach is employed. An enhanced computationally efficient k nearest neighborhood (EH-kNN) algorithm is developed to perform the distance estimation. Experiments on real periodic vibration signals generated by an electric hammer under different collecting distances and transmission medias are conducted to show the superiority of the proposed distance estimation method in this paper.