October 18, 2017
Countless devices are connected through wireless medium and all of those phones, sensors and smart home networks are putting increasing pressure on the radio frequency (RF) spectrum. Right now, the spectrum is divided into exclusively licensed bands, thus creating enormous inefficiency in the overall system. An unused band by any carrier can not accommodate users from other networks, a wasted opportunity for improving the overall efficiency. While researchers have been pushing for a more altruistic wireless system where spectrum sharing is encouraged, progresses are slow and confined to theory on paper. The Defense Advanced Research Projects Agency (DARPA) is looking at how to accelerate the paradigm shift from the century old licensing process to a more efficient system where spectrum sharing is incentivized.
“Basically, instead of that rigid framework of each has its own territory – you are allowed to share spectrum,” said Professor Biao Chen.
DARPA set up a competition unlike any other. The Syracuse University graduate students and undergraduates on the Air Orange team are working to develop a software radio network that can efficiently transmit data wirelessly – not by pushing other data aside – but helping machines learn how to share the road.
“If you put a radio right here, without telling it anything, it can work with any other radios in the area so everyone – even a WiFi access point or a garage door opener can share the spectrum in an efficient way,” said master’s student Eric Jiang. “If we develop a radio in such a way that it is mindful to other radios, it shouldn’t be a big problem if other radios do the same thing. It is sort of this utopia idea.”
Air Orange is one of just 30 teams selected for Phase One of the Spectrum Collaboration Challenge. On December 13th, their software will be put into an environment called the coliseum that emulates data and different transmission environments. To advance, their system has to be able to perform well – not only by delivering their own data in a timely manner, but to do so in a manner that help other teams to achieve the same. The biggest challenge is that they have to do so completely unaware of who other teams are and how their radio network operates.
“We are forbidden to talk to any of the other teams that participate in the collaboration so it is kind of like a blind test. Just throw a radio in there, into the environment and the goal is, I don’t know any of the other radios in this environment but I will be able to figure out a way to effectively and efficiently collaborate with them to share the limited spectrum,” said Jiang.
It is a different way to look at competition. The winning teams will not crush their opponents – they will be the most collaborative
“It’s not you win, other people lose. It’s more – you win and other people win so we both benefit from it,” said Xin Zhang ’18.
“It doesn’t make sense to hurt other teams, you are essentially hurting yourself. The goal is to figure out a solution to share this common resource,” said Jiang.
The interdisciplinary Air Orange team is made up of electrical engineering and computer science students. They have been testing their system in a series of scrimmages. Chen and the team meet daily to discuss improvements to their software.
“Every scrimmage has a deadline of midnight Sunday, which basically the weekend before is going to be crazy,” said Chen.
The final competition to be held in 2019 and the wining team will receive a grand prize of $2 million.