We’re just now getting 5G... What is 6G about?


6G will be a new framework for cellular communication, sensing, and machine-learning-based functions and services that requires a fundamental re-design of 5G along several key dimensions. 6G will strike a much-improved tradeoff between extremely high data rates, coverage and low latency guarantees as compared to 5G. The key technology enablers for 6G will include sensing for situational awareness and deeply embedded machine learning to deliver ever-evolving data-driven communication algorithms. 6G will embrace novel communication and sensing co-design principles for unprecedented cognition regarding channel conditions, spectral and spatial occupancy, and device localization including inside buildings. Such detailed awareness will allow the network to exploit novel contextualizing capabilities over time to improve network performance and enable anticipatory applications.  Machine learning techniques will be fundamental to 6G, from physical layer encoding/decoding techniques up through all layers of the network stack, including software for network configuration performance management and fault identification. 6G will utilize a greatly increased quantity and variety of spectral bands, with novel topologies providing unprecedented infrastructure density and coverage. Key metrics for 6G will be low power consumption and reliable coverage—as opposed to ever-increasing peak and average data rate—as well as localization accuracy.  


What is 6G@UT about?


6G@UT is a new research initiative within the Wireless Networking and Communications Group (WNCG) at UT Austin. Launched in mid-2021, 6G@UT brings together leading researchers at the intersection of wireless, networks, sensing, and machine learning with the goals of making new discoveries that enable 6G, training and educating future thought leaders in the cellular industry, and working closely with industry to innovate at the intersection of theory and practice.


The four key research directions for 6G@UT are (i) Deeply embedded machine learning from the physical layer up through to the application; (ii) Pervasive sensing for comprehensive situational awareness; (iii) Enabling new spectrum and new topologies and (iv) Flexible resource sharing, including of spectrum and computation resources.  We expect much of the innovation to lie at the intersection of these areas; e.g., the application of machine learning tools in conjunction with comprehensive sensing to enable efficient resource and spectrum sharing or highly-directional beamforming. We envision an increasingly open and software-defined cellular network, building on the ORAN paradigm, that provides a platform for continuous and more rapid innovation compared to 5G, particularly for ML-related technologies. 


6G@UT Research


6G will be a new framework beyond 5G that enabling superior contours of reliability, data rate and latency to a huge class of use cases and applications. Contrary to prevailing wisdom, we do not see 6G as being driven by increased peak data rates, although those will continue to improve. We also do not foresee a single dominant use case for 6G, such as mobile robots or low-latency IoT (e.g. mobile medicine, driverless cars, industrial automation), although those use cases will increase in importance relative to today.


Instead, we expect key technical metrics for 6G will be the reliability of high rate coverage, energy efficiency, and localization accuracy and environmental awareness. Of equal importance will be cost efficiency and deployability. As such, 6G will aim not to just continue to push data rates higher, but in the view of large bandwidths, instead emphasize a much-improved tradeoff between high data rates, coverage and latency guarantees as compared to 5G.


The key technology enablers for 6G will include sensing for situational awareness and deeply embedded machine learning to deliver ever-evolving data-driven communication algorithms. 6G will embrace novel communication and sensing co-design principles for unprecedented cognition regarding channel conditions, spectral and spatial occupancy, and device localization including inside buildings. Such detailed awareness will allow the network to exploit novel contextualizing capabilities over time to improve network performance and enable anticipatory applications. Machine learning techniques will be fundamental to 6G, from physical layer encoding/decoding techniques up through all layers of the network stack, including software for network configuration performance management and fault identification. 6G will utilize a greatly increased quantity and variety of spectral bands, with novel topologies providing unprecedented infrastructure density and coverage.


In this context, we categorize the four key novel directions for 6G as:


  1. Deeply embedded machine learning: from the physical layer through the application
  2. Pervasive sensing
  3. New spectrum and new network topologies
  4. Network Slicing and Sharing

Please click on the above topics to learn about our ongoing research on them.