Modified state activation functions of deep learning-based SC-FDMA channel equalization system
Abstract The most important function of the deep learning (DL) channel equalization and symbol detection systems is the ability to predict the user’s original transmitted data. Generally, the behavior and performance of the deep artificial neural networks (DANNs) rely on three main aspects: the netw...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-11-01
|
| Series: | EURASIP Journal on Wireless Communications and Networking |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13638-023-02326-4 |
| _version_ | 1827604497346592768 |
|---|---|
| author | Mohamed A. Mohamed Hassan A. Hassan Mohamed H. Essai Hamada Esmaiel Ahmed S. Mubarak Osama A. Omer |
| author_facet | Mohamed A. Mohamed Hassan A. Hassan Mohamed H. Essai Hamada Esmaiel Ahmed S. Mubarak Osama A. Omer |
| author_sort | Mohamed A. Mohamed |
| collection | DOAJ |
| description | Abstract The most important function of the deep learning (DL) channel equalization and symbol detection systems is the ability to predict the user’s original transmitted data. Generally, the behavior and performance of the deep artificial neural networks (DANNs) rely on three main aspects: the network structure, the learning algorithms, and the activation functions (AFs) used in each node in the network. Long short-term memory (LSTM) recurrent neural networks have shown some success in channel equalization and symbol detection. The AFs used in the DANN play a significant role in how the learning algorithms converge. Our article shows how modifying the AFs used in the tanh units (block input and output) of the LSTM units can significantly boost the DL equalizer's performance. Additionally, the learning process of the DL model was optimized with the help of two distinct error-measuring functions: default (cross-entropy) and sum of squared error (SSE). The DL model's performance with different AFs is compared. This comparison is conducted using three distinct learning algorithms: Adam, RMSProp, and SGdm. The findings clearly demonstrate that the most frequently used AFs (sigmoid and hyperbolic tangent functions) do not really make a significant contribution to perfect network behaviors in channel equalization. On the other hand, there are a lot of non-common AFs that can outperform the frequently employed ones. Furthermore, the outcomes demonstrate that the recommended loss functions (SSE) exhibit superior performance in addressing the channel equalization challenge compared to the default loss functions (cross-entropy). |
| first_indexed | 2024-03-09T06:01:56Z |
| format | Article |
| id | doaj.art-0df9b6200be04736b6248b7077029c39 |
| institution | Directory Open Access Journal |
| issn | 1687-1499 |
| language | English |
| last_indexed | 2024-03-09T06:01:56Z |
| publishDate | 2023-11-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | EURASIP Journal on Wireless Communications and Networking |
| spelling | doaj.art-0df9b6200be04736b6248b7077029c392023-12-03T12:07:30ZengSpringerOpenEURASIP Journal on Wireless Communications and Networking1687-14992023-11-012023112610.1186/s13638-023-02326-4Modified state activation functions of deep learning-based SC-FDMA channel equalization systemMohamed A. Mohamed0Hassan A. Hassan1Mohamed H. Essai2Hamada Esmaiel3Ahmed S. Mubarak4Osama A. Omer5Department of Electrical Engineering, Faculty of Engineering, Al-Azhar UniversityDepartment of Electrical Engineering, Faculty of Engineering, Al-Azhar UniversityDepartment of Electrical Engineering, Faculty of Engineering, Al-Azhar UniversityDepartment of Electrical Engineering, Faculty of Engineering, Aswan UniversityDepartment of Electrical Engineering, Faculty of Engineering, Aswan UniversityDepartment of Electrical Engineering, Faculty of Engineering, Aswan UniversityAbstract The most important function of the deep learning (DL) channel equalization and symbol detection systems is the ability to predict the user’s original transmitted data. Generally, the behavior and performance of the deep artificial neural networks (DANNs) rely on three main aspects: the network structure, the learning algorithms, and the activation functions (AFs) used in each node in the network. Long short-term memory (LSTM) recurrent neural networks have shown some success in channel equalization and symbol detection. The AFs used in the DANN play a significant role in how the learning algorithms converge. Our article shows how modifying the AFs used in the tanh units (block input and output) of the LSTM units can significantly boost the DL equalizer's performance. Additionally, the learning process of the DL model was optimized with the help of two distinct error-measuring functions: default (cross-entropy) and sum of squared error (SSE). The DL model's performance with different AFs is compared. This comparison is conducted using three distinct learning algorithms: Adam, RMSProp, and SGdm. The findings clearly demonstrate that the most frequently used AFs (sigmoid and hyperbolic tangent functions) do not really make a significant contribution to perfect network behaviors in channel equalization. On the other hand, there are a lot of non-common AFs that can outperform the frequently employed ones. Furthermore, the outcomes demonstrate that the recommended loss functions (SSE) exhibit superior performance in addressing the channel equalization challenge compared to the default loss functions (cross-entropy).https://doi.org/10.1186/s13638-023-02326-4Activation functionsDeep artificial neural networksDeep learningChannel equalizationSymbol detectionLong-short-term memory |
| spellingShingle | Mohamed A. Mohamed Hassan A. Hassan Mohamed H. Essai Hamada Esmaiel Ahmed S. Mubarak Osama A. Omer Modified state activation functions of deep learning-based SC-FDMA channel equalization system EURASIP Journal on Wireless Communications and Networking Activation functions Deep artificial neural networks Deep learning Channel equalization Symbol detection Long-short-term memory |
| title | Modified state activation functions of deep learning-based SC-FDMA channel equalization system |
| title_full | Modified state activation functions of deep learning-based SC-FDMA channel equalization system |
| title_fullStr | Modified state activation functions of deep learning-based SC-FDMA channel equalization system |
| title_full_unstemmed | Modified state activation functions of deep learning-based SC-FDMA channel equalization system |
| title_short | Modified state activation functions of deep learning-based SC-FDMA channel equalization system |
| title_sort | modified state activation functions of deep learning based sc fdma channel equalization system |
| topic | Activation functions Deep artificial neural networks Deep learning Channel equalization Symbol detection Long-short-term memory |
| url | https://doi.org/10.1186/s13638-023-02326-4 |
| work_keys_str_mv | AT mohamedamohamed modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem AT hassanahassan modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem AT mohamedhessai modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem AT hamadaesmaiel modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem AT ahmedsmubarak modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem AT osamaaomer modifiedstateactivationfunctionsofdeeplearningbasedscfdmachannelequalizationsystem |