Can cognitive radio technology help solve some difficult spectrum management issues by creating "virtual guardbands"?

The fundamental goal of spectrum management is to maximize use of the spectrum. This generally is taken to mean use without "harmful interference". The first problem in avoiding harmful interference has classically been to avoid having two signals on the same frequency that can cause interference to each other at locations where interference-free use is desired. This has been the classical issue in spectrum management. However, cochannel interference is not the only interference mechanism in practical radio systems. Adjacent channel interference, receiver desensitization, and receiver-generated intermodulation are also potential interference mechanisms and have been the underlying issues in several contentious spectrum controversies in the last decade. Realistic receivers have "near/far problems" when the desired signal is much weaker than an adjacent channel signal. These problems most seriously affects mobile receivers since cost, size, and power constraints in mobile units limit the filter performance that is practical to keep adjacent channel signals out of the receiver amplifiers and mixers where strong undesired signals can result in interference.