Deep-learning based self-interference cancellation for full-duplex network
Elimination of self-interference in full duplex systems has always been a significant challenge due to the huge power differences between the interference and desired signals. Many traditional methods have been implemented to tackle this problem, however they struggle to effectively cancel out unwan...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/176187 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Elimination of self-interference in full duplex systems has always been a significant challenge due to the huge power differences between the interference and desired signals. Many traditional methods have been implemented to tackle this problem, however they struggle to effectively cancel out unwanted signals completely. With the advancements in technology, deep learning has become a popular topic of interest and has shown many promising results in today’s communications system. In this report, three different deep learning architectures, namely LSTM, GRU and 1dCNN are modelled and being compared to the traditional methods. The results show that deep learning is able to outperform traditional methods, even when the signals are unequalized. Hence, this underscores the potential of deep learning to enhance the communication systems. |
|---|