.png){kind=icon}
.png){kind=icon}
Return
Efficient Automatic Composite-Modulation Classifier Using Cyclic-Paw-Print Features
.png){kind=icon}
PRE
AI
DOI：10.1109/LCOMM.2024.3350670.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A novel efficient automatic composite-modulation (CM) classification approach based on the cyclic-paw-print (CPP) features is introduced for future cognitive space communications. In the proposed new scheme, the new image representation of CM signals, a.k.a. CPP named after its biomimetic shape, is constructed from the normalized second-order cyclic-spectrum of the received CM signal. Then, the discrete cosine transform (DCT) is further applied to the obtained CPP matrix to extract the essential CM feature vector, which is further digested by use of linear discriminant analysis (LDA). Finally, the random forest (RF) is employed to identify the CM scheme of the received signal by taking the aforementioned CM feature vector as the corresponding attributes. Monte Carlo simulation results demonstrate that the proposed new scheme can greatly outperform the existing CM classifiers while the proposed scheme requires a low computational-complexity.
Keywords:
Automatic composite-modulation classification (ACMC)
cyclic-paw-print (CPP)
discrete cosine transform (DCT)
linear discriminant analysis (LDA)
random forest (RF)
Journal
IF:
4.4
Papers:
1.2W
Citations:
2.2W