.png){kind=icon}
.png){kind=icon}
Return
HMLP-CNN-based anti-turbulence algorithm for underwater pulse position modulation signal detection
.png){kind=icon}
PRE
AI
W
Z
L
T
Z
DOI：10.1117/1.OE.65.4.048101.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We propose a hybrid architecture, termed hybrid multilayer perceptron-convolutional neural network (HMLP-CNN), specifically designed for an anti-turbulence signal detection algorithm in pulse position modulation (PPM) optical communication systems operating under challenging underwater turbulent conditions. This architecture effectively combines the advantages of multilayer perceptron (MLP) and convolutional neural network (CNN) through a linked structure, ensuring precise and accurate signal detection. Experimental results demonstrate that the HMLP-CNN significantly outperforms least mean square (LMS) equalizer threshold detection, standalone MLP, CNN, recurrent neural network (RNN), gated recurrent unit (GRU), long short-term memory (LSTM) in terms of bit error rate (BER) performance. Under six-channel turbulence conditions, HMLP-CNN reduces the BER by one to two orders of magnitude compared with LMS equalizer threshold detection, even achieving zero BER under certain conditions. Under the channel condition of strongest turbulence, average BER performance of the LMS equalizer threshold detection algorithm, MLP, CNN, GRU, LSTM, RNN, and HMLP-CNN is 1.63 & times; 10 (- 3) , 6.02 & times; 10 (- 4) , 3.25 & times; 10 (- 4) , 8.41 & times; 10 (- 4) , 1.06 & times; 10 (- 3) , 1.33 & times; 10 (- 3 )and 2.31 & times; 10 (- 4) , respectively, at a PPM-8 pulse transmission rate of 200 kb/s. The HMLP-CNN algorithm demonstrates significant BER improvements, achieving 85.84%, 61.63%, 28.92%, 72.53%, 78.30%, and 82.70% enhancement over the other six algorithms. To the best of our knowledge, we pioneer the application of neural network algorithms for PPM modulation signal detection in turbulent underwater channels.
Keywords:
optical communication
HMLP-CNN
signal detection algorithm
anti-turbulence
PPM
Journal
O
IF:
1.2
Papers:
178
Citations:
1.1W
Organization
Cited Papers