4.7 Article

Smart Detection Using the Cascaded Artificial Neural Network for OFDM With Subcarrier Number Modulation

Journal

IEEE WIRELESS COMMUNICATIONS LETTERS
Volume 10, Issue 6, Pages 1227-1231

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/LWC.2021.3062686

Keywords

OFDM; Detectors; Computer architecture; Neural networks; Receivers; Equalizers; Signal detection; Artificial neural network (ANN); orthogonal frequency-division multiplexing (OFDM); subcarrier number modulation; low-complexity detection; cascaded architecture

Funding

  1. National Natural Science Foundation of China [61872102, 61661004]
  2. International Collaborative Research Program of Guangdong Science and Technology Department [2020A0505100061]
  3. Pearl River Nova Program of Guangzhou [201806010171]
  4. Fundamental Research Funds for the Central Universities [2019SJ02]
  5. Guangxi Science Key Research and Development Project [AB1850043]
  6. Natural Science Foundation of Guangdong Province [2020A1515110470]

Ask authors/readers for more resources

To address the high detection complexity of OFDM-SNM signals, a cascaded artificial neural network (CANN) is designed to provide a simplified solution, with numerical simulations confirming its effectiveness and efficiency.
Since the proposal of orthogonal frequency-division multiplexing with subcarrier number modulation (OFDM-SNM), the high detection complexity of this novel modulation scheme becomes a major issue degrading its applicability in practice. To provide an easy-to-implement and low-complexity solution to the detection of the OFDM-SNM signal, we design a cascaded artificial neural network (CANN) at the OFDM-SNM receiver to decouple the detection of subcarrier number pattern (SNP) and constellation symbols on the active subcarriers. This is the first time that the CANN design is introduced to assist in detecting parallel signals. Numerical simulations verify the effectiveness and efficiency of the proposed smart detection scheme and show that the CANN based smart detection with adequate training can yield comparable error performance to the optimal maximum-likelihood detection at much lower complexity.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available