Tez özetleri Astronomi ve Uzay Bilimleri Anabilim Dalı 2

Wideband Underwater Acoustic Channel Estimation by Using Compressive sensing
Recently, like other communication systems there are great developmets in underwater communication system. In that thesis searching and estimation of underwater communication channel is studied by using compressed sensing approach.
According to the wireless electromagnetic communication channel damaging effects of underwater channels is much more. In addition, because of the difficulty of using radio signal in underwater channel, acoustic signal is used. Disadvantage of the acoustic signal is low speed of transmission.
Accordingly, delay of environmetal trasmission becomes large and predicting channel in receiver is more difficult. Because of falling of transmission speed and delay of broadcasting, in that study Orthogonal Frequency Division Multiplexing Technic is used. Simulation model of multi-way wireless channel structure has been used successfully in MATLAB environment. MATLAB model of wireless underwater channel hasn’t been developed but close MATLAB model can be created by add in some parameters into present channel parameters. In the study, Rayleigh channel is choosed as a channel model. Distorting affects of underwater channel isn’t less like wireless channel. Underwater noise is not total white Gauss. It is a changeable noise according to frequency.
Underwater signal can be considered sparse signal because of affects of acoustic signal speed and structure of underwater channel.
Compressive detection is an algorithmal whole by based on less number of observation by compressing in sparse signal. The purpose is lashing back sparse signal by using less number of samples than the classical sampling rate.
It uses various algorithms (MP OMP) to get the signal back. While these algorithms are used, distance between vectors is not count by Euclidian distance. In the study Matching Pursuit is used in order to obtain signal reached to the receiver.
As forecasting channel with Compressed Sensing method, error permormance changed so much related to chosing of Espansion Matrix.
Although Compressed Sensing Method does not give perfect result in error performance and channel broadcasting under the Expansion Coefficient 2, it gives perfect result near the theoric border for bigger than coefficient 2.