The wide proliferation of wireless services and applications with increasing bandwidth needs is rapidly creating a spectrum shortage. However, the problem is caused primarily by inefficient legacy spectrum allocation and utilization policies, so that even when some applications suffer from lack of bandwidth, there is idle capacity in the band they are using or other bands. This project develops and demonstrates a revolutionary approach for addressing spectrum scarcity and unlocking hidden communication capacity thereby increasing the reach and utility of wireless connectivity. The non-traditional communication technique studied in this research effort detects transmission opportunities that occur when incumbent primary users enjoy signal to noise ratio values that are higher than the minimum value required to maintain their quality of service. It then judicially exploits these opportunities while preserving the current quality of service of the primary users.
The project develops novel change detection methods that fuse goodness of fit tests and density estimation and similarity assessment using information theoretic methods to study network traffic and designs innovative distributed goodness of fit tests and density estimation and similarity assessment techniques implemented over an asynchronous communications network. It compares the performance of the distributed goodness of fit tests and density estimation and similarity assessment techniques with those of a new class of distributed cumulative sum (cumsum) change detection methods that it constructs generalizing cumsum methods that have been proposed in the literature in other application domains. Finally, it demonstrates experimentally the benefit of using these approaches to unlock hitherto hidden communication capacity and quantify the increase in local communication capacity created by using the new schemes.
