Intelligent Cognitive Communication System Based on Multimodal Deep Learning
1School of Information Science and Engineering Shandong University , China
2Beijing University of Posts and Telecommunications, Beijing, China
3Manchester Metropolitan University, UK
4Shandong Management University, Jinan, China
Intelligent Cognitive Communication System Based on Multimodal Deep Learning
Description
The sixth generation (6G) communication system is the era of the Internet-of-Everything (IoE), guiding the development of human civilization into a new stage. In 6G non-cooperative communication systems, intelligent cognitive communication plays an important role as a key component of semantic communication. Cognitive communication aims to utilize artificial intelligence to automatically identify a series of parameters, including modulation scheme, coding style, channel, etc. The significant progress of artificial intelligence in deep learning methods has made intelligent cognitive communication possible at the engineering level. The essence of deep learning-based intelligent cognitive communication is to establish a potential mapping of signals and parameters, which involves a series of research directions including multimodal fusion, signal representation, model structure, hyper-parameters setting, optimization, generalization, reinforcement learning, etc.
In practical wireless communication environments, signals can be easily affected by adverse factors such as multipath propagation, interference, and noise, posing formidable challenges to the cognitive systems. In the multipath propagation environment, the received signal copies may have different amplitudes, phases, and time delays due to multipath propagation, resulting in signal fading and distortion at the receiver. The time delay causes the signal to spread out in the time domain, leading to misaligned symbol boundaries between the received signals and the transmitted signal. Circuit elements in wireless communication systems generate phase noise during operation. The Doppler effect caused by the high-speed movement of drones causes the frequency of the received signal to differ from the frequency of the transmitted signal, thus causing a frequency offset. In recent years, deep learning has been gradually developed for intelligent wireless communications and has shown great potential. Unlike traditional methods that rely on manually designed features and domain expertise, deep learning models are able to automatically learn feature representations of signals, and thus can be more easily generalized to a variety of application scenarios without the requirements for extensive feature engineering for specific situations. It is an important research direction in the future.
This Special Issue aims to provide a platform for researchers to discuss opportunities, problems, challenges, and possible solutions for modulation, classification, channel modelling, and parameter estimation in deep learning-based intelligent cognitive communication systems. This special issue will provide opportunities to present theoretical proofs, technical strategies, and empirical evidence to propose new tools, methods, models, frameworks, and techniques to design solutions to a range of challenges during the application process of intelligent semantic communication. We welcome both original research and review articles.
Potential topics include but are not limited to the following:
- Parameter estimation in intelligent communication systems
- Dynamic spectrum access in cognitive radio
- Joint optimization of multiple modules in intelligent cognitive communication
- Semantic encoding and decoding based on deep neural networks
- Extraction and reconstruction of semantic information with graph neural network
- Deep learning for automatic modulation classification/recognition
- Deep learning in designing reconfigurable intelligent surfaces
- Communication channel modelling with deep learning driven by big data
- Lightweight deep learning model structure for real-time reasoning at the receiver
- Regularization and generalization in deep learning for intelligent communication across scenarios
- Full stage intelligent processing and fusion of multisource signals driven by deep learning
- Parallel data processing method in intelligent MIMO systems
- Self-organizing form and scheme of neural network model structures
- The theory and method for guiding hyper-parameter setting in deep neural networks