AbstractsEngineering

Broadband beamforming plays an important role in many applications involving antenna array signal processing such as communications, radar, sonar, imaging, geophysical exploration, astrophysical exploration and biomedical diagnostics. The traditional approaches to broadband beamforming use tapped delay line filter architectures, where the weights of the filters are chosen by optimization algorithms. Unfortunately, these approaches have some limitations. Firstly, large matrices need to be manipulated to estimate the filter coefficients and the necessary steering delays are cumbersome to implement, with implementation errors causing loss of system performance. Secondly, when some imperfections occur in the array, the resulting mismatch in array responses causes degradation in the beamformer performance. Therefore, the need for robust beamforming arises in many practical applications. Additionally, the placement of nulls in the array pattern is also a vital issue for cancelling slow moving jammers and known unwanted directional sources. Moreover, the traditional techniques have difficulties in controlling side lobe level regardless of the main lobe position in the array pattern. This thesis presents algorithms for broadband beamforming that overcomes these limitations. The result is a computationally efficient beamformer that does not require presteering delays. A unique feature of these new algorithms is inherent robustness to array imperfections, giving the user some degree of flexibility in choosing array parameters. The thesis also discusses techniques for broadband pattern synthesis. These techniques are designed to produce isolated point and broad sector spatial nulls in the presence of array imperfections. Moreover, a robust Chebyshev broadband antenna array pattern synthesis technique is proposed to achieve a specified side lobe level with interference cancellation capabilities. The thesis also presents broadband arrays having circular and concentric circular antenna element geometries, where these arrays are able to scan a beam through entire space with little change in either the beamwidth or the side lobe level. The performance of this proposed processor is compared with the performance of existing broadband processors and shows that the limitations of steering delays, computational cost, robustness, equal side lobe level, null placement and array geometries can be overcome by the use of the robust broadband methods presented in this thesis.