The rapid evolution of the wireless communication industry has been met with increasing demand for data-driven services and new paradigms to improve connectivity between people, and everything. Beyond, data provision, future generation networks are required to provide a platform for the integration of communication, sensing, and intelligence. This integration is propelled by the increasing efficiency of wireless communication networks, sensing technologies, and the growth in data analytics as demonstrated in the field of artificial intelligence and machine learning. To achieve this, after the reliability, latency, data-rates, and massive machine type communications support in 5G, 6G and future generation wireless networks (FGWNs) are required to facilitate computation-oriented networks, Agile based enhanced Mobile Broadband communications, and intelligence driven ultra-reliable low latency communication (iURLLC).
Computation oriented networks both support the provision of intelligence towards self-aware or self-organizing networks and facilitates the aggregation of intelligence from multiple devices (or user equipment, UE) via privacy aware aggregation techniques such as federated learning. Intelligence towards self-organizing networks will allow the network to optimize and repair itself. By integrating sensing and data collection, the self-organizing networks will have the ability to adapt to environment, devices, application requirement, and hardware. By taking advantage of softwarisation, the network AI will have the ability to tune itself to suit application requirements, thereby maintaining improved quality of service (QoS) for the applications. For example, the network will differentiate dataflows from healthcare devices from those from industrial IoTs and, in event of emergency health issue (as observed by the network through data), the network will give high priority to the data from the healthcare device. Federated learning, where knowledge held by the network is distilled from multiple UEs, allows decentralized computation towards the formation of more secure intelligent models.
Agile based enhanced Mobile broadband (eMBB) communications provides agility and adaptability for eMBB introduced in 5G networks. eMBB covers data driven use cases that require high data rates across a wide coverage area. Although in 5G, eMBB delivered higher data rates, enhanced connectivity, and higher user mobility, for 6G and FGWNs, the network topology adapts using agile methods to manage link congestion, in both physical and social contexts. This is carried out by dynamically allocating radio transmission resource blocks on a per user basis in both downlink and uplink transmission separately with a goal to fulfil the QoS targets of data radio bearers of the served UEs. On the FGWN infrastructure, in higher layers, this is implemented by the inclusion of applications and service awareness, multi-cell or multi-node coordination of eMBB delivery actions, and multi-link or multi-node connectivity options. While in lower layers, a flexible radio structure, inclusion of options for different PHY numerologies, and enhanced MIMO and mMIMO support are used to deliver agile based eMBB.
Although the transmission delay achieves the 1ms target in 5G, FGWNs will be required to achieve significantly improved targets for both transmission and stochastic delays. This implies that, in addition to the end-to-end delay at the radio access network (RAN), FGWN will also address bottlenecks in the upper communication layers such as queueing, processing, and access delays. For iURLLC or event defined URLLC, 6G and FGWNs will be required to facilitate URLLC in extreme or highly critical events where the device densities, spectrum availability, infrastructure state, and traffic patterns changes with respect to both space and time. This will be needed to fulfil emerging vehicular or similar industry 5.0 applications which are expected to come with new performance metrics that have not been considered in 5G. Such performance metrics will include spectrum efficiency, context awareness, throughput, energy efficiency, delay jitters, network availability, and security. To achieve this, intelligent models will be embedded in URLLC designs that will use machine learning algorithms, information theory, and knowledge in wireless communication to design.
Computation oriented networks, agile based eMBB, and iURLLC will serve as drivers towards the integration of communication, sensing, and intelligence for FGWNs. In addition to these, emerging technologies such as intelligent reflecting surfaces, massive MIMO, holographic radio, terahertz communication, blockchain for trust and safety, advanced channel coding modulation, photonics defined radio, nano -networks and visible light communications will serve as key enablers to achieve the requirements in FGWNs. Most of these technologies -some evolutionary and others revolutionary, are still in their initial explorations stages and hold the capacity to serve the various application needs expected from industry and society.
Future generation wireless networks will be required to provide the bedrock technologies that will drive various expected applications including but not limited to:
- The use of digital twins which will provide state information in real-time of replicas of physical devices or processes. This will provide data for descriptive -data driven monitoring, diagnostic -AI fault detection, predictive -use past data for future estimations, and prescriptive analytics where outputs of predictions are used to prescribe options to address expected future events.
- Truly immersive extended realities with the emergence of high-fidelity holograms and network-enabled robot for multi-sensory object manipulation which will enable seamless interaction between the physical world and the virtual world. This will deliver solutions in teleoperation, tele medicine, tele surgery, telepresence, and interactions within and without the metaverse.
- Connected Intelligence and Internet of everything (IoE) where everything comes online including people, processes, data, things -plants, vehicles, robots, sensors, devices. IoE provides access to real time connectivity to and from everything. Connected intelligence involves the connection of intelligent things, and services to an intelligent network.
Wireless communications keep evolving to meet needs in industry and in society. This article has shown the fundamental features of future generation wireless networks, how they are achieved, and current areas of research that are necessary to meet the demand requirements of the networks. Although this technology opens up a lot of interesting applications that have been highlighted, there are concerns with privacy and security and the need for privacy preserving intelligent models -such as federated learning, and consensus models -such as blockchain, to ensure that these exciting technologies are safely delivered.
