Dr. Nikita Barabanov
Department of Mathematics
Antenna array signal recognition
We consider the problem of signal recognition (direction of arrival and power) using antenna array with equidistant receivers. The famous methods for super resolution will be described and discussed. Then we formulate an adequate concept of signal-to-noise ratio, which incorporates the number of observations and the number of receivers. A new super resolution method will be introduced and analyzed. The method allows dealing with highly correlated signals, and with sufficiently close signals. Examples show that it works better than all other methods.
We also will discuss briefly some algorithms based on ideas of Kohonen neural networks.