Robust Normalization Algorithms for Low-Frequency Active Sonar Signals
Abstract
Normalization is the process of forming a signal-to-noise ratio (SNR) from the raw signal power received by a sonar system. It is considered to be an essential part of sonar systems, especially under non-alerted conditions, and it may contribute significantly to the overall effectiveness of a surveillance system. However, the effectiveness of the normalization process is reliant upon an accurate estimate of the noise only reference; under real circumstances, the received data in general contains both noise and signal elements that can be difficult to separate. This report considers two algorithms for providing an adaptive estimate of the noise only reference for the normalization processors of active sonar signals. A 'soft' background noise estimation approach rather than a selective rejection of false echoes is chosen. Owing to their specific design, these algorithms a particularly good candidates to normalize the data stream before applying processing methods based on ping history. Their performances and their robustness against signal components in the ref- cells are examined. Recommendations about the implementation of the algorithms are derived. The results obtained from applying the algorithms to a real data set show that they perform well.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1995
- Accession Number
- AD1120698
Entities
People
- X. M. De Gastines
Organizations
- Centre for Maritime Research and Experimentation