Effective source number enumeration approach under small snapshot numbers
Direction of Arrival (DOA) estimation of signal sources is one of the research hotspots in the field of array signal processing. However, traditional DOA estimation methods usually require many snapshots, a high Signal-to-Noise Ratio (SNR), and a Gaussian white noise background, which are often d...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/119946/ http://psasir.upm.edu.my/id/eprint/119946/1/119946.pdf |
| Summary: | Direction of Arrival (DOA) estimation of signal sources is one of the research hotspots
in the field of array signal processing. However, traditional DOA estimation methods
usually require many snapshots, a high Signal-to-Noise Ratio (SNR), and a Gaussian
white noise background, which are often difficult to meet in actual environments. To
solve this problem, this study proposes a signal source number estimation method
based on supplementary empirical mode decomposition (SEMD). The method first
uses the SEMD method to decompose the array signal, decomposing the complex
signal into several Intrinsic Mode Functions (IMFs), and then extracts features through
these IMFs to estimate the number of signal sources. To verify the performance of the
proposed SEMD method, this study designs a series of experiments, using theoretical
data and measured data from a radio frequency anechoic chamber laboratory as
research objects. The experimental conditions cover different snapshot numbers,
SNRs, and noise backgrounds, aiming to simulate various complex environments in
actual applications. Experimental results show that the SEMD-based method performs
significantly better than the traditional signal source number estimation algorithm in
these complex environments, especially under a small number of snapshots, the
SEMD method can still maintain a high estimation accuracy. This study also makes a
significant contribution to data science by providing a comprehensive method for
estimating the number of signal sources, which is integrated with a machine learning
model. This method overcomes the limitations of traditional methods in complex
environments by combining signal processing problems with pattern recognition
problems, significantly improves the accuracy of data analysis in complex
environments, and provides an innovative solution for signal processing and pattern
recognition in data science. |
|---|