A Novel Approach of Optimal Signal Streaming Analysis Implicated Supervised Feedforward Neural Networks
The analysis and interpretation of enormous amounts of data generated by 5G networks present several challenges related to noise, precision, and feasibility validation. Therefore, this study aims to evaluate the effectiveness of channel equalisation in the network and enhance it by distributing sign...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-01-01
|
| Series: | IET Signal Processing |
| Online Access: | http://dx.doi.org/10.1049/2024/2819057 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The analysis and interpretation of enormous amounts of data generated by 5G networks present several challenges related to noise, precision, and feasibility validation. Therefore, this study aims to evaluate the effectiveness of channel equalisation in the network and enhance it by distributing signals over all subcarriers and symbols. The error-free signal received ensures the reliable transmission of signals in the network connection. These simulations were undertaken to fulfil the needs of and adapt the transmission properties according to the specific conditions of the channel. The dataset consists of artificially generated radio waves to train signals through neural networks (NNs) and machine learning algorithms to detect errors properly. The primary objective is to achieve optimal signal performance. In this regard, an artificial neural network (ANN) was initially employed, explicitly utilising the back-propagation technique and a feedforward multilayer perceptron (MLP). In addition, the signals were subjected to train using a real-time simulator, employing feedforward neural network and support vector machine (SVM) to validate the proposed methodology. Feedforward MLP achieved the highest performance in simulations compared to SVM. The scheme is promising to achieve optimal signal performance in real-time. |
|---|---|
| ISSN: | 1751-9683 |