DARPA sees artificial intelligence’s machine learning enabling on-the-fly sharing of spectrum at “machine” timescales. It is announcing a 3-year competition, starting in 2017, to identify the best collaborative – not competitive – noble ways to manage spectrum beyond current band assignments and static allocation schemes.
Making Spectrum Use More Efficient: On one side, society’s use of devices with wireless capabilities continues to grow, and with the imminent arrival of 5G IoT more products, from vacuum cleaners, to automobiles to drones, need access to wireless capability. It will take far more efficient and nimble use of finite spectrum resources to meet this ever increasing demands. In light of spectrum becoming more crowded, and with a presidential order on the table to free up 500 MHz of spectrum assigned to the military for commercial use by 2020, the DoD has prioritized this activity. Making more efficient use of spectrum has become one of US DoD priorities, and its Defense Advanced Research Projects Agency (DARPA) unveils the world’s first collaborative machine-learning effort.
Spectrum Collaboration Challenge for RF Communications: This signals the start of the work at DARPA which announces that is prepared to work with partners developing systems that collaboratively adapt in real time to changes in congested electromagnetic spectrum. To do this, it says the plan is to infuse radios with advanced machine-learning capabilities to collectively develop strategies for optimizing the use of wireless spectrum that is not possible Today due to the approach of pre-allocating exclusive access to designated frequencies. There are two participation options for seeking one of the up to 30 available slots in this SC2 challenge.
The COLOSSEUM Testbed: DARPA is building a large test bed for researchers to remotely conduct large-scale experiments with intelligent radio systems in realistic RF conditions and use cases that reflect normal situations such as a crowded commuter station, a busy downtown intersection and such.
A 3-Year Competition: Just amounted a few days ago the terms of a 3-year SC2 competition as well as the $2 million prize. Competitors operate software defined radios seeking to collaborate with other radios operating simultaneously in ways that optimize spectrum efficiency across an entire communications network. The program is after machine-learning tools that provide this capability. At its core, the program aims to redefine conventional spectrum management using emerging tools like programmable software-defined radios augmented with machine learning tools. “We want to radically accelerate the development of machine-learning technologies and strategies that will allow on-the-fly sharing of spectrum at machine timescales,” Paul Tilghman, program manager for the spectrum challenge, noted in a statement.
Implications for the Mobile Industry Are Far- Reaching: The 3-year program has just launched. Wait to hear news from us about the competition. Overall, the project will be closely watched by the mobile communications industry, which will be paying attention to the spectrum sharing proposals coming out of the program that are suitable for commercial use, and complement existing ones proposed by industry groups and technical communities such as the Multefire Alliance.
Resources: DARPA Spectrum Challenge to Focus on Collaboration, Collaborative Intelligent Radio Networks, DARPA Spectrum Collaboration Call, SC2 Architectural Needs, MulteFire Alliance, US Spectrum Allocation Challenge
