PORTLAND, Ore. Virginia Tech hopes to smarten up experimental cognitive radios so that ad hoc communications networks can adapt to aid in disaster relief, battlefield communications, consumer Wi-Fi and other cognitive radio applications.
By sharing a distributed knowledge base, Virginia Tech's "cognitive engine" will serve as the
communication system's "brain" by sensing unused bandwidth, avoiding interference, adapting to changing circumstances and optimizing network performance. At the same time, the engine will help maintain the autonomy of individual cognitive-radio nodes.
Virginia Tech researchers have applied for a patent on their cognitive engine, which will also work with existing hardware, said Charles Bostian, director of Virginia Tech's Center for Wireless Telecommunications. "We are going to see if we can use vacant TV channels for Wi-Fi-like services without interfering with other emergency services that are already operating there."
Cognitive engines allow radios to share a distributed knowledge base that parcels out individual and collective reasoning tasks to network nodes as a way to automate adaptation and learning. Cognitive radios have surpassed software-defined radio as the focus of radio design because they give radios the ability to decide which bands to use based on availability, location and experience.
Cognitive radios may not autoconfigure by themselves, but under the supervision of a cognitive engine a network can be optimized.
Bostian said Virginia Tech's cognitive engine tests
will be performed within a shielded environment to prevent any interference with emergency radios that are already using vacant TV spectrum. If successful, the cognitive engine will be able to coordinate the frequency assignments and communication protocols of ad hoc networks so existing radios do not interfere with existing emergency broadcasts.
The project's second goal is to learn how to integrate cognitive radios with existing radios and provide metrics for improvements needed for mixed networks. The cognitive engine will then be able to optimize a
network's performance, depending on the mix of radios available, while providing developers with techniques for extending the use of existing radio spectrum.
In 2003, the Federal Communications Commission issued a request for proposals to add artificial intelligence to cognitive radios. The upgrade would allow the new radios to sense dead zones, interference and usage patterns as well as adapt to changing real-time circumstances.