B5G: Intelligent Coexistence Model for Edge Network

While researchers are focusing on the fifth-generation (5G) cellular network, the network operators and standard bodies are discussing specifications for beyond fifth-generation (B5G) and 6G. Attributes of B5G include edge intelligence which involves artificial intelligence or machine learning (ML) in the architecture. Network edge servers, or base stations (BS), use edge computing to make time-critical decisions, especially in IoT devices while the data are being transmitted into the cloud. As the dynamic spectrum sharing continues in B5G, BSs implements the coexistence between Wi-Fi and cellular network. These exciting advances require energy efficiency to be considered as network operators pay the majority of the expenses to energy consumption. In this paper, different prediction models on traffic behaviour are computed to determine the lowest root mean square error. The best prediction model is used in the wake-up policy to consider the communication and computing times of the BS needed to return in service. Furthermore, a wake-up policy for the BS is introduced to maintain Quality of Service (QoS) while minimizing energy consumption. Particularly, a wake-up time threshold is set so that if the duration of the traffic prediction time does not cross this threshold, the decision will not be in favour to put it into sleep mode. This ensures that the QoS of the user is not compromised, as this threshold removes the unnecessary wasted time for BS to go to sleep and wake-up.