Manhasset, N.Y. In an effort to provide reliable and highly available battlefield communications at low system cost, the Defense Advanced Research Projects Agency has launched a program based on the premise that low-cost handsets or nodes can be effective if combined with intelligent, adaptive networks.
"The idea is to try to shift the emphasis from very high-capacity, high-cost radios to a more inexpensive radio, and then make up for that by shifting the burden to the network," said Preston Marshall, director of Darpa's XG communications program. The XG short for next-generation program's mandate is to develop the technologies needed for the Department of Defense to dynamically access all available spectrum on an as-needed basis. The new program, called the Wireless Network after Next (WNaN), will leverage the XG work to bring down the cost of radio nodes by lowering their performance requirements.
According to Marshall, spurious-free dynamic range and linearity, two crucial performance determinants for radios, are hard to achieve in software-defined radios, which almost by definition have a wide operating bandwidth. In the drive to overcome those performance limiters, the cost of the radio rises rapidly. "Maybe we've gotten as good as we can afford to get," said Marshall, referring to the cost/performance trade-off.
Instead, using work done under XG, "we can make a radio that can find holes in the spectrum and can then move around the spurs [and other interference]," he said. "We can work around [radio] limitations instead of trying to solve them." The end result, in principle, would be a radio node that costs less than $500, can be deployed rapidly in an ad hoc fashion in trees, poles or vehicles, for example and can operate in the range of 900 MHz to 6 GHz.
Such a node, Marshall said, would have four integrated radios, to allow for the use of multiple-input, multiple-output technologies, multiple bands or repeater-like techniques. Darpa in September sent out a call for proposals for technologies and system concepts that would enable such a network/node combination. The WNaN program also recently held an industry day, which drew not only typical military contractors but also companies more entrenched in the consumer, low-cost world, said Marshall. "The military contractors know the military's needs, but in very low volume," he said, whereas the consumer/commercial companies "know how to build things at very low cost."
WNaN (see www.darpa.mil/ato/solicit/WNaN/index.htm) has already received proposals from both military and commercial interests, but program officials would not identify them. Marshall also pointed to an analog spectral-processing effort that recently got under way to complement the WNaN initiative. It will focus on the development of inexpensive and widely tunable filters (see www.darpa.mil/mto/solicitations/open.html#baa06-08).