A complex exponential waveform embedded in white noise can be optimally detected by matched filtering which is equivalent to Discrete Fourier Transform (DFT). However, if the input includes multiple complex exponentials, the DFT processing is not optimal. The frequency spectrum of the complex exponential signal with finite observation interval is not impulse. The spectrum includes side-lobes called spectral leakage.Because of the strong side-lobes, weak components can be masked, or side-lobes can be interpreted as independent complex exponentials causing false alarms. The conventional approach is to suppress side-lobes by using an amplitude weighting function, called a window function. However, windows reduce the signal to noise ratio at the output and also widens the main lobe in the spectrum resulting in a loss of frequency resolution. In this thesis, the problem of detection of a complex exponential signal under a secondary complex exponential interference is investigated. Alternative windowing techniques such as spatially variant apodization (SVA), dual apodization (DA) are studied in the context of detection of weak sinusoids under strong sinusoidal interference. The detection performance of SVA and DA is compared with the conventional methods and the optimal detector.
| Identifer | oai:union.ndltd.org:METU/oai:etd.lib.metu.edu.tr:http://etd.lib.metu.edu.tr/upload/12612785/index.pdf |
| Date | 01 December 2010 |
| Creators | Balci, Burak |
| Contributors | Candan, Cagatay |
| Publisher | METU |
| Source Sets | Middle East Technical Univ. |
| Language | English |
| Detected Language | English |
| Type | M.S. Thesis |
| Format | text/pdf |
| Rights | To liberate the content for public access |
Page generated in 0.0018 seconds