Accession Number:



Machine Learning Aided Efficient and Robust Algorithms for Spectrum Knowledge Acquisition in Wideband Autonomous Cognitive Radios

Personal Author(s):

Corporate Author:

University of New Mexico Albuquerque United States

Report Date:



The objective of this project was to conduct research that will advance the wideband autonomous cognitive radio WACR technology. These are radios that have the ability to sense state of the radio frequency RF spectrum and the network and self-optimize its operating mode in response to this sensed state. First, a formal framework was developed for robust spectrum knowledge acquisition in a wideband autonomous cognitive radio. The performance of this framework was evaluated based on simulations. Next, a machine learning based sub-band selection algorithm for WACRs was developed based on reinforcement learning and its performance was analyzed through a combination of analysis and simulations. Finally, motivated by certain application scenarios of interest, a new definition for the state of the spectrum of interest to a WACR was developed. Currently, this new approach is being used for developing practical cognitive communications protocols with considerably less computational complexity than previous alternatives. Future work will include design, implementation and analysis of cognitive communications protocols suited for space and satellite communications using this new state definition.

Descriptive Note:

Technical Report,27 Apr 2015,27 Jul 2016



Communities Of Interest:

Modernization Areas:

Distribution Statement:

Approved For Public Release;

File Size: