The Applications of Some Statistical Procedures for Analyzing Spectral Data.
Abstract
The implementation of the fast Fourier transform and the availability of the high-speed digital computer has resulted in the widespread usage of spectral analysis methods in the analysis of random processes. Generally, most spectral analysis methods have involved the uses of power spectrum. This research is concerned with developing some useful statistical methods for analyzing such spectral estimates. Road profile data from transportation or highway engineering is used for exemplifying these methods. Three standard statistical procedures are illustrated, two applying to coherence estimates, and the third to power spectral estimates. A new statistical test based around a normalized coherence difference distribution is also included and enables the use of the standard normal distribution tables for testing for similar coherence samples. (Modified author abstract)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 10, 1972
- Accession Number
- AD0778833
Entities
People
- B. F. Womack
- Roger Sherwood Walker
Organizations
- University of Texas at Austin