Enabling Beyond-5G Industrial IoT through Intelligent Networking in Unlicensed Spectrum
Lead Research Organisation:
University of Kent
Department Name: Sch of Engineering & Digital Arts
Abstract
EPSRC : George Frangulea : EP/T518141/1
The use of 5G operating in unlicensed spectrum (NR-U) as an enabler for industrial Internet of Things (IIoT)/Industry 4.0 services (e.g., smart power grids, factory automation and connected hospital etc.), is a research area currently generating significant interest from both academia and industry (with vendors such as Nokia and Qualcomm and operators such as Orange actively engaged in this area). The purpose of the proposed project is to enhance reliability and spectrum utilisation and minimize latency in NR-U, by using intelligence informed by physical layer messages and data gathered through spectrum scanning. This type of intelligence will be achieved through the use of reinforcement learning (RL-a type of machine learning algorithm). The RL algorithm will leverage beamforming (a process by which a radio beam is formed in a particular direction) techniques within the frequency channel access procedure to enhance spatial-reuse (multiple devices communicating concurrently, using the same time/frequency resources), and overcome problems such as hidden and exposed nodes within the network. The overarching target of the system will be to ensure that stringent latency and reliability requirements of IIoT/Industry 4.0 services are met, to ensure the seamless coexistence with other wireless technologies and enable the seamless coexistence of services at different priority levels, within the same network (a critical step forward in meeting Industry 4.0 / IIoT quality-of-service requirements).
The use of 5G operating in unlicensed spectrum (NR-U) as an enabler for industrial Internet of Things (IIoT)/Industry 4.0 services (e.g., smart power grids, factory automation and connected hospital etc.), is a research area currently generating significant interest from both academia and industry (with vendors such as Nokia and Qualcomm and operators such as Orange actively engaged in this area). The purpose of the proposed project is to enhance reliability and spectrum utilisation and minimize latency in NR-U, by using intelligence informed by physical layer messages and data gathered through spectrum scanning. This type of intelligence will be achieved through the use of reinforcement learning (RL-a type of machine learning algorithm). The RL algorithm will leverage beamforming (a process by which a radio beam is formed in a particular direction) techniques within the frequency channel access procedure to enhance spatial-reuse (multiple devices communicating concurrently, using the same time/frequency resources), and overcome problems such as hidden and exposed nodes within the network. The overarching target of the system will be to ensure that stringent latency and reliability requirements of IIoT/Industry 4.0 services are met, to ensure the seamless coexistence with other wireless technologies and enable the seamless coexistence of services at different priority levels, within the same network (a critical step forward in meeting Industry 4.0 / IIoT quality-of-service requirements).
People |
ORCID iD |
| Philippos Assimakopoulos (Principal Investigator) |