Optimal Digital Detection of Acoustic Signals in Colored Noise.
Abstract
This thesis addresses optimal methods for the detection of acoustic signals corrupted by colored noise. In achieving this we provide a study of the characteristics of ambient noise in the ocean and the digital techniques which can be used in the process of detecting known acoustic signals which are corrupted by that noise. Various techniques are studied, in particular the use of matrix decomposition techniques applied to the correlation matrix or to a data matrix, and the matched filter for colored noise. Other methods such as the inverse filter, the differential operator, and the adaptive prediction-error filter will also be looked at for their whitening properties. The theoretical foundations of those techniques are presented as well as the application of each method to the problem. Simulations are conducted for each technique in order to provide quantified performance measurements supporting the use of each method.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1995
- Accession Number
- ADA305824
Entities
People
- Martin A. Cloutier
Organizations
- Naval Postgraduate School