Abstract

Using beamforming in a wireless communication channel increases its capacity. A major challenge in deploying beamforming in commercial wireless communications is the high implementation cost. Microwave beamforming (MBF) is a low complexity alternative due to using only one radio frequency (RF) down-conversion and analog to digital converter (ADC). However, traditional MBF structures employ a phase shifter and a gain control block per antenna element which are expensive and bulky. Moreover, signal processing for MBF structures is more difficult due to lack of an antenna array signal vector in the processor. This thesis investigates both the microwave hardware and the signal processing issues associated with the MBF structures. It introduces a new low-complexity beamforming implementation technique, the microwave sampling beamformer (MSBF). In addition, a new perturbation technique based on array signal estimation is devised for application of advanced signal processing to the MBF structures. Throughout this research different RF implementation and signal processing issues are jointly taken into account. In the proposed MSBF structure, phase shift and amplitude attenuation are controlled through switching of the antenna array signals using control pulses with adjusted time delay and pulse width. Both the phase shift and amplitude attenuation are controlled over the full range in one simple block. The proposed structure is validated in the frequency and time domains with microstrip and wire antenna arrays. Different MSBF issues such as switch design, finite duration pulse train, and image replica rejection are investigated. A four-element prototype of the MSBF structure composed of the microwave and control hardware is constructed and tested. The beamforming structure is designed to have eight bits phase shifter resolution and 12 bits amplitude resolution. The whole prototype is examined for beam steering, side-lobe level (SLL) control, and null-forming. In the new perturbation technique the antenna array signal vector is estimated based on the array signal temporal correlation. The temporal correlation is provided by a fast perturbation rate or receiver bandwidth reduction. Single-port adaptive unconstrained least mean square (ULMS) beamforming and multiple signal classification (MUSIC) DOA algorithms are investigated and they outperform the corresponding multi-port algorithms depending on the channel scenario.