Date of Award

Spring 1993

Document Type


Degree Name

Doctor of Philosophy in Electrical Engineering - (Ph.D.)


Electrical and Computer Engineering

First Advisor

Yeheskel Bar-Ness

Second Advisor

Nirwan Ansari

Third Advisor

Alexander Haimovich

Fourth Advisor

Zoran Siveski

Fifth Advisor

Michael Blair Porter

Sixth Advisor

Jack H. Winters


Adaptive antenna arrays are widely used in many advanced radar, sonar, and communication systems because of their effectiveness in cancelling intentional or unintentional interferers. A uniformly spaced linear array, referred to as a Uniform Regular Array (URA), is the usual structure used for interference cancellation. The Minimum Redundancy Array (MRA) structure proposed in this work is a special kind of thinned array whose application was limited in the past to direction finding. MRAs with the same number of array elements can resolve directions of much more closely spaced signals than URAs.

The URA structure is customarily utilized for interference cancellation, and the Minimum Noise Variance (MNV) criterion is a common performance measure for deriving optimum weights, provided that the desired signal is absent during adaptation. The MNV criterion is to minimize the combined sum of the interference and background noise power.

Another approach to interference cancellation using the URA structure is the eigencanceling method. This method, which is based on the eigenstructure of the spatial autocorrelation matrix, when compared to the conventional beamforming method, has the following advantages: 1) deeper interference cancellation 2) independence of the interfers' power, and 3) faster optimum weight convergence. In this work, both the conventional beamforming and eigencanceling methods were applied to the MRA structure and investigated analytically. Performance of the MRAs were studied and compared to that of the URAs.

For uncorrelated interferers, the cancellation depth of the MRA in the main beam region was almost the same as that of the URA with the same aperture and many more elements. When the eigencanceling technique was applied, it was found that the convergence rate of the MRA was about four times faster than that of the URA.

This work also contains other topics, such as the relation between the eigenspaces of the MRA structure and its corresponding URA. Preliminary results on planar MRA structures are also included. For an array application with a large aperture requirement in terms of the number of array elements, the MRA proved to be a much better choice than the URA in achieving interference cancellation.