Use this URL to cite or link to this record in EThOS:
Title: Sonar data characterisation and analysis
Author: Levonen, Mika
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 2005
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This thesis is concerned with sonar signal processing, in particular the statistical characterisation of sonar data. It proposes a number of signal processing methods that are suitable for sonar data. More specificity it deals with the signal processing of time series, from a sonar system, which were collected during a number of experiments in the Baltic Sea. The return signals in sonar can be viewed as a mixture of deterministic and stochastic parameters. Consequently, instead of trying to model the wave propagating environment itself this thesis will illustrate a variety of characteristic properties of the signals and propose suitable methods for solving the problems. In Chapter 3 a statistical characterisation of sonar data is presented, different aspects of statistical properties of sonar data are addressed. The focus is on the level of stationarity of both active and passive sonar. For active sonar the issue of ping to ping stationarity, for both the actual ping and for the reverberation tail, are examined and found to be non-stationary. However, for some cases it is seen that there are consecutive pings that are stationary. This leads to the conclusion that it can be beneficial to use several pings from the same target. However it is necessary to remember that there is a large variation in the number of consecutive stationary pings. For passive sonar both ambient noise and tonals emitted from surface vessels are investigate. It is found that for all cases tested the sonar data is non-stationary. That is speaking of stationarity in the strict sense. There are however parts in the data that exhibit more stationary behaviour then the data in general. The stationarity length of the ambient noise data is also examined, using data from a multisensor trial (almost 800 data files, see Section 2.5.1). A large proportion of the data set had a stationarity time of roughly 0.4 seconds, or slightly longer. This is also seen in data from the fibre glass boat trial (Section 2.5.6). The data seems to be stationary for about 0.4 seconds. Testing the passive data for symmetry and linearity show that the data is mostly linear, and symmetric.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available