ANALYSIS OF AN ADAPTIVE ANTENNA ARRAY WITH INTERMEDIATE-FREQUENCY WEIGHTING PARTIALLY IMPLEMENTED BY DIGITAL PROCESSING (LMS-ALGORITHM)

This dissertation is concerned with the analysis of an adaptive antenna array based on the LMS (Least Mean Square) algorithm with weighting functions at a non-zero frequency. The ultimate goal is to develop a method of generating these weighting functions by digital means.; It is shown that for optimal performance in interference rejection, a filtering operation is necessary due to the non-zero frequency weighting. With the constraints required to accommodate wideband signals, a preprocessing filter for proper digitization is also necessary. Both filtering operations are shown to be primary sources of instability in the feedback processing loops.; This investigation provides a time-domain analysis of the effect of the necessary filters on stability. The differential equations describing this I.F. array are derived. Time-delays related to the filters are introduced in these equations, and their impact on stability and performance is shown.; Stability conditions that alleviate the effects of the required filters are derived. These conditions are given as time-compensation methods accounting primarily for the group delays of the filters. Computer simulation results verifying the effectiveness of the stabilization procedure are provided.; Finally, a digital weight processor which allows the generation of the inphase and quadrature components of the weighting functions is proposed. The primary purpose of this design is the elimination of DC offsets and feedthrough usually associated with a reconversion to baseband during demodulation. The digital weight generation is based on the bandpass-sampling technique and on the interleaving of two sample-and-hold circuits, resulting in a sampling rate twice the carrier frequency. The inphase and quadrature components are combined digitally by means of switching logic. Analog filtering then produces the weights at an intermediate frequency.