Imec announced that a team of scientists from its research groups at the Universities of Antwerp and Ghent, together with researchers from Rutgers University, will participate in the finals of the U.S. Defense Advanced Research Projects Agency (DARPA) Spectrum Collaboration Challenge on Wednesday, Oct. 23 at the Mobile World Congress 2019 (MWC19) Los Angeles for a chance to win the $2 million first prize. Team "SCATTER" finished in the top 10 after two rounds of competition—receiving $750,000 in each round—and will compete against nine other teams to find a solution to the problem of clogged mobile networks, which are the result of public consumption of increasing amounts of mobile data and a static allocation of spectrum.

Online activity is ubiquitous to modern life and business. Thanks to the expansion of and upgrades to mobile networks consumers, enterprises and others can web surf and download at ever faster speeds and mobile internet traffic is accelerating at an unprecedented rate. However, the mobile spectrum is subject to physical limits: as soon as the RF through which data travels to our smartphones, laptops or tablets are at capacity, the transmission of additional data is delayed or data cannot get through.

A weak Wi-Fi connection at the busy coffee shop around the corner, for example, is due to the limited capacity of the mobile spectrum. The RF used by that Wi-Fi network is overcrowded during busy times, resulting in users being thrown off the network, or data moving very slowly. And not only Wi-Fi connectivity is suffering from the spectrum bottle neck, but also many other more critical or professional services, such as crisis communication during rescue operations or wireless controlled processes in manufacturing (Industry 4.0). The mobile spectrum is doomed to become increasingly congested unless we find a way to make smarter use of that limited spectrum.

The solution that team SCATTER brings to the finals is constantly scanning the available spectrum and intelligently analyzing which moment is best for carrying out a task, depending on the spectrum needs of the other teams and the state of the network. Artificial Intelligence is used to learn how other teams use the spectrum and what is the best way to respond to it.

DARPA’s Spectrum Collaboration Challenge

In recognition of the spectrum problem, DARPA launched the Spectrum Collaboration Challenge to encourage research teams worldwide to look for solutions. Historically, DARPA's challenges have always led to concrete results, and this is the ultimate goal this year as well.

Since the invention of radio, frequencies have been allocated on an exclusive basis. Much of the allocated spectrum is barely used, while others (e.g., the bands assigned to Wi-Fi) are completely overloaded. It is precisely this approach that is hampering the efficient use of the mobile spectrum in the longer term since many parts of the spectrum are currently not used. With its Spectrum Collaboration Challenge, DARPA is looking towards "intelligent spectrum sharing," with various actors continuously and dynamically sharing the mobile spectrum.

The DARPA Challenge is all about cooperation: all competing teams have access to a common frequency band. They have to share it with each other, and at the same time manage to use it efficiently when carrying out tasks such as forwarding heavy video files or supporting time-sensitive drone communication. For example, if one team needs more spectrum at a given time to transmit multiple high-quality heavy video streams, the teams need to collaborate to find a balance, which may involve focusing on delivering files that need less bandwidth. The eventual winner of the DARPA Spectrum Collaboration Challenge is the team that completes its own tasks in the best possible way and at the same time, in the most flexible way, supporting the other teams in their search for bandwidth.