Signal Automatic Modulation Classification and Recognition in View of Deep Learning
With the advancement of 5G technology, wireless communication resources such as channels and spectrum become scarce. This necessitates ensuring the efficiency and security of signal modulation and demodulation, which imposes higher requirements for wireless communication systems. However, signal mod...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10286033/ |
| _version_ | 1797654053811912704 |
|---|---|
| author | Tianpei Xu Ying Ma |
| author_facet | Tianpei Xu Ying Ma |
| author_sort | Tianpei Xu |
| collection | DOAJ |
| description | With the advancement of 5G technology, wireless communication resources such as channels and spectrum become scarce. This necessitates ensuring the efficiency and security of signal modulation and demodulation, which imposes higher requirements for wireless communication systems. However, signal modulation has the problems of large amount of data, low recognition accuracy and various types. In this study, a classification network of automatic modulation classification recognition algorithm for signal-to-noise ratio is proposed to solve the problem that traditional noise reduction algorithms will damage signals with high signal-to-noise ratio, consequently reducing the accuracy of signal recognition. In order to solve the problem of high complexity of network model algorithm, in particular, a signal automatic modulation classification and recognition algorithm based on neural network autoencoder is proposed. Experimental results show that the accuracy of signal automatic modulation classification recognition in the algorithm increases as the increase of modulation signals and tends to be stable. When the modulation signal is 0dB, the recognition accuracy gradually converges to the highest, and reaches 81.6% when the modulation signal is 18 dB. In contrast, the DenseNet algorithm has the lowest recognition accuracy, with only 77.5% recognition accuracy when the signal modulation classification is 18dB, a difference of 4.1%. This indicates that the algorithm performs exceptionally well in automatic signal modulation classification, and its complexity is lower than other comparative network models, providing certain advantages. |
| first_indexed | 2024-03-11T16:53:47Z |
| format | Article |
| id | doaj.art-dae183726bb84d40ae914cf4dda7629a |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-11T16:53:47Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-dae183726bb84d40ae914cf4dda7629a2023-10-20T23:00:31ZengIEEEIEEE Access2169-35362023-01-011111462311463710.1109/ACCESS.2023.332467310286033Signal Automatic Modulation Classification and Recognition in View of Deep LearningTianpei Xu0Ying Ma1https://orcid.org/0009-0005-9920-3784College of Education, Hulunbuir University, Hulunbuir, ChinaCollege of Software, Nanchang Hangkong University, Nanchang, ChinaWith the advancement of 5G technology, wireless communication resources such as channels and spectrum become scarce. This necessitates ensuring the efficiency and security of signal modulation and demodulation, which imposes higher requirements for wireless communication systems. However, signal modulation has the problems of large amount of data, low recognition accuracy and various types. In this study, a classification network of automatic modulation classification recognition algorithm for signal-to-noise ratio is proposed to solve the problem that traditional noise reduction algorithms will damage signals with high signal-to-noise ratio, consequently reducing the accuracy of signal recognition. In order to solve the problem of high complexity of network model algorithm, in particular, a signal automatic modulation classification and recognition algorithm based on neural network autoencoder is proposed. Experimental results show that the accuracy of signal automatic modulation classification recognition in the algorithm increases as the increase of modulation signals and tends to be stable. When the modulation signal is 0dB, the recognition accuracy gradually converges to the highest, and reaches 81.6% when the modulation signal is 18 dB. In contrast, the DenseNet algorithm has the lowest recognition accuracy, with only 77.5% recognition accuracy when the signal modulation classification is 18dB, a difference of 4.1%. This indicates that the algorithm performs exceptionally well in automatic signal modulation classification, and its complexity is lower than other comparative network models, providing certain advantages.https://ieeexplore.ieee.org/document/10286033/Neural network self codingconvolutional long short term memory networkautomatic signal modulationsignal-to-noise ratio classful network |
| spellingShingle | Tianpei Xu Ying Ma Signal Automatic Modulation Classification and Recognition in View of Deep Learning IEEE Access Neural network self coding convolutional long short term memory network automatic signal modulation signal-to-noise ratio classful network |
| title | Signal Automatic Modulation Classification and Recognition in View of Deep Learning |
| title_full | Signal Automatic Modulation Classification and Recognition in View of Deep Learning |
| title_fullStr | Signal Automatic Modulation Classification and Recognition in View of Deep Learning |
| title_full_unstemmed | Signal Automatic Modulation Classification and Recognition in View of Deep Learning |
| title_short | Signal Automatic Modulation Classification and Recognition in View of Deep Learning |
| title_sort | signal automatic modulation classification and recognition in view of deep learning |
| topic | Neural network self coding convolutional long short term memory network automatic signal modulation signal-to-noise ratio classful network |
| url | https://ieeexplore.ieee.org/document/10286033/ |
| work_keys_str_mv | AT tianpeixu signalautomaticmodulationclassificationandrecognitioninviewofdeeplearning AT yingma signalautomaticmodulationclassificationandrecognitioninviewofdeeplearning |