The six Thinknet 6G focus areas

We anticipate that artificial intelligence (AI) and machine learning (ML) will be used in the design and optimization of the wireless interface of self-optimizing transmitters and receivers and for new forms of context-aware decision making.

AI and ML technologies have made significant advances in recent years, particularly in image classification and computer vision, and are being used in systems as diverse as social media and autonomous driving. 6G researchers are also investigating how AI/ML can be applied to wireless systems and networks. Plans for the current generation of mobile communications, 5G, envisage the use of AI in at least three different ways:

• Replacing some layer 1 and 2 algorithms (e.g. channel estimation, preamble detection, equalization and user scheduling) with new AI-based algorithms that are either more powerful, less complex, or both.
• In deployment optimization, e.g. to configure an optimal subset of beams for a given coverage area, taking into account current cell traffic patterns. AI/ML techniques are used to configure and reconfigure the network on the fly, optimizing the network without human intervention.
• More accurate localization of end devices based on machine learning

In addition to these three advances in RAN, AI/ML is also being used for comprehensive end-to-end network automation and orchestration across multiple heterogeneous networks and layers. As we move from 5G to 6G, the use of AI will no longer be just an extension, but a fundamental element of the design and optimization of the air interface.

As 6G is expected to use the frequency range from 114 GHz to 300 GHz, new spectrum technologies will need to be developed in addition to the existing spectrum holdings in FR1 and dFRA. These technologies include new radio frequency integrated circuits (RFIC) with integrated antenna arrays and phase shifters, new concepts for hybrid beamforming in combination with MIMO and new receiver architectures.

One of the biggest challenges in using these "sub-THz" high bands is the realization of devices with high output power at reasonable costs. Signals with short wavelengths are easily blocked, diffraction by objects is limited, and signal absorption by water (including humidity) is significant. These challenges are partially offset by non-line-of-sight (NLOS) coverage made possible by reflections from buildings and walls in dense urban and indoor environments. These challenges and peculiarities of high-band propagation will require a significant amount of research and development before deployment is feasible.

We expect sub-terahertz bands from 114 GHz to 300 GHz to become available for backhaul networks. In addition, these bands can free up frequencies in the mmWave bands through close point-to-point communication. Other possible applications include short-range communication between display and computing devices or rack-to-rack communication in data centers.

In addition to the new technologies for the sub-terahertz bands, cost-effective MIMO techniques will enable better utilization of the mmWave/FR2 and cmWave bands through massive, reusable MIMO with increased network density.

Research into improved spectrum utilization in the lower frequency bands is important as spectrum is scarce and needs to be optimized. We expect 6G operators and service providers to use AI/ML for dynamic spectrum access in time, frequency and space.

6G networks will not only be used as communication networks, but also for detecting and locating passive objects. Many of the 6G use cases require precise positioning and localization, e.g. in industrial automation or autonomous driving. Although the real-time kinematic global navigation satellite system (RTK GNSS) can enable high-precision localization, it requires good satellite visibility, which is not available indoors. Currently, indoor localization is based on ultra-wideband (UWB) or Bluetooth Low Energy (BLE), for which additional access points and devices need to be installed. These systems run in parallel with the regular communication system and therefore incur additional infrastructure and operational costs that could be avoided if the regular communication system also enabled precise localization.

5G includes features to improve localization, with a focus on industrial automation environments. As we move towards 6G, we expect the communication network to take on not only localization but also additional sensing tasks. The 5G solutions will achieve centimeter-level accuracy through AI/ML-based channel mapping as well as data fusion from RF, camera and other sensors on robots.

The 6G system design will not only be optimized for communication, but will also include special capabilities for sensing. For example, waveforms capable of detecting chirp signals can be multiplexed with waveforms optimized for communication. The wider signal bandwidth in the sub-terahertz and terahertz bands can increase the precision of sensing, possibly to within a millimeter. This level of precision enables new applications such as defect detection in manufacturing or cancer diagnosis.

The combination of precise multimodal sensing with the cognitive technologies enabled by 6G will enable advances in the analysis of behavioral patterns and human preferences, creating a "sixth sense" that anticipates users' needs, enabling more intuitive interaction between the digital and physical worlds.

The 6G network will be able to automatically and dynamically reconfigure itself to adapt to changing requirements, demand and environments.

Some of these automatic adaptation techniques have already been mentioned above in the discussion on AI/ML. The RAN network will be able to automatically adapt to current requirements, e.g. through:

• Improved layer-1 and layer-2 behavior for channel estimation, preamble detection, equalization and user scheduling
• Optimized beam configuration to configure the current optimal subsets of beams for a given coverage area, taking into account current cell traffic patterns and user behavior

However, dynamic adaptation in the radio access network is only a small part of the dynamic adaptation envisaged for the 6G network. To enable comprehensive end-to-end automation and organization in real time across multiple heterogeneous networks and operators, significant changes to the core network architecture are required.

Sub-networks

Previous generations of mobile networks focused on extending the voice and data network to individual mobile endpoints, mainly people with cell phones in their hands. 5G is the first generation to extend the focus to the industrial environment, introducing new architectural developments such as support for time-sensitive networks (TSN) as a bridging function. 6G will extend 5G to provide full deterministic, wireline reliability for a variety of connectivity scenarios ranging from static, isolated devices to interconnected, locally interacting devices to fast-moving swarms connecting to each other and the network.

To ensure both high temporal and high spatial reliability and determinism, the 6G network will be divided into several semi-autonomous sub-networks. The sub-networks must provide at least the most critical services in the sub-network, even if connectivity with the wider network is poor or non-existent. To ensure reliability, the network and/or sub-networks will utilize multiple paths and opportunistic device-to-device connections, resulting in cell-less architectures.

Highly specialized slicing

In addition to the above requirements for sub-networks and multi-connection scenarios, we expect the 6G network to be divided into multiple slices and virtual networks. Slices may be highly specialized and could run separate software stacks for differentiated flow handling in each slice. The current trend towards virtualization in the RAN, especially at higher layers, will lead to the development of micro-RAN services that can then be flexibly assembled into slice-specific RAN implementations. For example, a slice for video services could include video optimization microservices, while a slice for low throughput IoT devices could provide connectionless access.

We expect flexible, slice-specific functions across multiple devices (gateways, relays, cell sites, remote edges, edges, regional clouds, ...) and across multiple hardware platforms, depending on the requirements of each slice. New concepts and innovations are needed to manage and orchestrate such highly specialized slices.

RAN/core convergence

In 5G, we are already seeing two trends that will impact the 6G architecture. In 5G, many of the "edge" functions in the RAN are moving to the core, as parts of the Centralized Unit (CU) are virtualized and then implemented in the edge or metro cloud. At the same time, some of the core 5G functions will be more decentralized and virtualized and then moved to the regional or metro clouds or even to the edge clouds for low latency services.

We expect these trends to continue, with more RAN functions moving to the core and more core functions moving to the edge. Within the 6G timeframe, we expect to see a convergence of functions in a few functional blocks that implement both a RAN and a core, leading to a "coreless" RAN.

A wireless mobile network is by its very nature a highly dynamic network. Devices and their users come and go, they move from cell to cell at different speeds, and their data rate requirements rise and fall as they watch a video or make a call. Adding complex subnetting and network slicing running on hardware and software blocks with different functions across multiple operators around the world requires a level of consensus and automated orchestration that goes far beyond current 5G capabilities.

Performance requirements for 6G networks can include sub-1 ms latency and extreme reliability ("nine 9s").

The key techniques to achieve low latency and high reliability in 5G are mini-slots and guarantee-free channel access for low latency, and multi-connected links with multiple access points, carriers and packet duplication for reliability. However, even 1 ms latency over the air is insufficient for some 6G use cases, e.g. as a replacement for traditional wired connections for industrial communication (e.g. for EtherCAT). These use cases require much lower wireless latencies, which are in the order of 100 µs for Gbit/S data rates. And the desired reliability can reach nine 9s for some use cases in industrial automation.

6G is designed to meet these requirements by utilizing the wider bandwidth of the mmWave spectrum to meet latency and data rate specifications. Reliability can be improved by simultaneous transmission over multiple paths involving multiple hops. Cooperative relaying over device-to-device links can create multiple separate paths from the network to a particular end device. Finally, predictive beam management based on AI/ML predictions can reduce uncertainty in connection quality.

Use cases in IoT and industrial IoT bring another form of extreme networking into the discussion, in the form of ultra-low or zero-power end devices. There are several use cases that require low-power or zero-power sensors with extremely long operating times that are distributed over a large geographical area. For example, for the ongoing monitoring and inspection of bridges, wireless sensors installed during the construction phase should operate for approximately 100 years without human intervention. Solutions for such extreme situations must include low-power communication, extremely low idle currents, energy harvesting and accessible energy storage.

Some of the use cases envisioned for 6G have critically high security requirements. As the physical and biological worlds merge, currently available data protection solutions are woefully inadequate. As multimodal sensor technology will capture almost everything a person does and the entire environment they are in, users must be able to restrict what data they share and with whom it is shared. This requires clear and easy to understand ways for users to express their data sharing preferences and for users to retain complete control over their personal data.

There are a variety of new signal processing techniques to ensure security and privacy in mixed reality worlds, and these mechanisms will be an essential part of the 6G network.

When the reliability of wired networks is replaced by a wireless network in industrial applications, new security and privacy requirements arise for this wireless network. Interference is a new threat to industrial networks that must be combated. Attackers may try to disrupt industrial networks from outside the plant, so physical security alone is not enough. Disruption can seriously affect industrial operations, especially in time-critical networks. 5G and 6G networks will be designed to eliminate such risks.

The network architecture envisaged for 6G will see applications such as Car2x communications, robots and personal body sensors split into separate sub-networks. This requires a change in the approval strategy to shift approval from the network level to the sub-network level. For example, in a body area network (BAN), the devices in the subnet belong to that subnet, so the authorization and management of the devices must be done at the trust boundary of the subnet. The entire 6G network will connect several subnets, so an additional authorization level may be required at the network level. In addition, the subnets may belong to mutually untrusted entities, which requires not only a clear separation between the subnets, but also between the network and its subnets. It is crucial that the subnets act as independent networks that have their own authorization authority and are responsible for managing their own subnet assets. The dynamic behavior of assets in a cellular network (e.g. constantly joining and leaving the network) makes it a challenge to maintain the privacy and anonymity of sub-networks.

Trust in the network is critical to the success of 6G.