Wireless Innovation Forum Top Ten Most Wanted Innovations

Innovation #1: Dynamic Spectrum Management 

1.1 Executive Summary

The realization of a fully Dynamic Spectrum Management (DSM) system powered by cognitive network technology and supported by collaborative intelligent radios[1] is considered a future vision for DSM. In the next decade, it is anticipated that the spectrum management ecosystem will experience a paradigm shift by relying extensively on spectrum knowledge (which will leverage the application of data science (e.g. machine learning and data mining) to spectrum and related data) to perform proactive dynamic spectrum management in order to optimize spectrum utilization and foresee spectrum demand. Ultimately, a DSM system will be capable of adapting its prediction and decision-making through machine learning, in order to capture changes and adapt to the RF environment and related spectrum usage parameters.

DSM design paradigms must evolve to demonstrate elements of the paradigm shift by integrating spectrum knowledge as a core function in spectrum management; possibly leading to new and innovative spectrum management strategies and enhancing existing spectrum management strategies. One can envision a long term outcome to spectrum management where a combination of different types of services could dynamically share large portions of pooled spectrum, and further down the road the whole spectrum, in order to maximize spectrum usage efficiency.


[1] Paul Tilghman, DARPA, Invited Talk: Spectrum sharing through collaborative autonomy, Sep 25, 2017, VTC Fall 2017 Toronto

1.2 Application

Maximizing spectrum usage efficiency based on knowledge will benefit a broad range of user communities including:

  • Network operators can apply a data-driven proactive approach to spectrum access and management, where spectrum shortage and oversupply across different networks can be predicted and managed.
  • The military, first responders, and public safety communities can enable quick and meaningful spectrum access based on data-driven decisions, through the provision of relevant spectrum knowledge. 
  • Regulators can automatically track and act upon unauthorized spectrum usage (automated compliance function) through the use of spectrum knowledge.

New DSM functions such as flexible licensing, including real-time spectrum monetization with real-time auctions and dynamic spectrum assignment of the licenses, could also be considered as part of the DSM evolution going forward and would benefit regulators and brokers. 

1.3 Description

The dynamic spectrum management is enabled by several core technologies:

  • Advancements in cognitive/intelligent RF radios that allow them to learn and adapt the operational aspects of the radio to their current propagation environment. Cognitive Radios (CR) will operate over multiple frequency bands over a wide frequency range. Lower costs will enable CRs to become ubiquitous. 
  • Advanced spectrum sensing will be needed to quickly and accurately classify known and unknown sensed RF signals (users) in targeted bands and identify transmission opportunities over a very wide spectrum pool that may host a large number of different wireless services and higher priority services.
  • Adaptive antennas that can scan the surrounding environment and adapt directivity and beam patterns to target transmissions.
  • Data fusion methods and big-data-analytics for helping with machine learning, prediction and decision-making with appropriate training methods (e.g. Reinforcement Learning). 
  • Adaptive receivers that can better manage interference and track changes in the RF environment.

Taken together these technologies can enable wireless devices and networks to coordinate access to spectrum among themselves by sharing information on the spectrum environment, prioritizing new spectrum demands (e.g. emergency access), and protecting the priority access of legacy systems (e.g. military radar). 

Ultimately, dynamic spectrum management could include local information processing (e.g. machine learning algorithms residing on the cognitive radios), as well as cloud management, to create an ecosystem where the cognitive radios could coexist seamlessly without a need for human intervention / licensing.