Grant: Discovery Grants Program - Individual

The sixth generation (6G) communication networks are expected to support unprecedentedly diverse services for vertical industries with highly increased data traffic volume. Both the scale and complexity of 6G networks will increase significantly by comparison with 5G networks, to support seamless global coverage and various extreme use cases. Innovative solutions are required for network management automation and for intelligent service support to end users. Artificial intelligence (AI) will be a native component in 6G networks, and the future communication systems will be required to effectively support AI in its essence. Network virtualization is a new and key technology in 5G systems, which softwarizes network control functions and decouples them from the physical network, thus providing a foundation for network automation. The traditional model-based network optimization is limited to small scale systems and can become inapplicable for 6G networks. Data-driven solutions based on machine learning and big data provide a promising alternative approach to 6G network management. We will focus on integrated model-based and data-driven solutions for performance enhancement, to take advantage of technical insights obtained in the past few decades from mathematical modeling and optimization. We will also explore meta learning models to address non-stationary network dynamics in the lifecycle management of machine learning models. We will investigate and develop digital twin techniques to facilitate integration of model-based and data-driven solutions, to emulate the network evolution, and to achieve real-time network management automation based on machine learning models trained by the emulated network operation trajectory data. This research program includes developing: 1) a holistic network virtualization framework for a space-air-ground integrated network, which utilizes AI techniques and data analytics for intelligent network management and user service support; 2) joint computation, communication, and sensing solutions that exploit cooperation between devices and edge servers for intelligent Internet-of-Things applications; and 3) network management automation solutions based on meta learning and digital twins for vehicular applications with non-stationary system dynamics. This research program will provide insights and generate new knowledge for smart, agile, and efficient service provisioning and network management. Research on AI for networking and on networking for AI has been gaining momentum in the global telecommunication community. This program will contribute to technology development for 6G network intelligence and automation. Four PhD students and two postdoctoral fellows will participate in this program. They will establish solid theoretical foundation and develop research skills in telecommunications and advanced AI techniques, emerging as expert additions to the Canadian high technology industry labor force.