AN UNBIASED SOURCE LOCATION ALGORITHM AND PERFORMANCE SURFACE OF ZERO TRACKING ALGORITHMS

In this thesis, a new zero tracking algorithm to estimate the locations of directional sources in an unbiased manner is presented. The zeros of the system, which correspond to the sources' locations, are updated recursively taking into account the effect of receiver noise. A formula relating the variances of the sources' locations to the number of samples used has been obtained, allowing the accuracy of the estimated source locations to be controlled. Since these estimates are unbiased, the new algorithm has a much better performance while locating weak sources in difficult signal environment when compared with conventional methods such as the LMS and biased null steering algorithms. In addition, a study on the stationary points and transient behaviour of using zero tracking based techniques in a power inversion array is presented. As a function of complex response zeros, the output power performance surface has been found to possess no local minimum, although saddle points may exist at positions where some of the zeros a.re identical. Hence, steepest descent based zero tracking algorithms will always converge to the global minimum, even though the zeros being adjusted may wander close to the saddle points for an appreciable amount of time. With only a single asymptotic time constant, the use of zero tracking has been found to have a transient convergence which is generally much faster and more uniform with respond to the starting positions and environment. Together with the advantages that the zeros are available directly and that the tracking behaviour is determined by the asymptotic time constant only, zero tracking algorithms are attractive viable alternatives for algorithms whose complexity is proportional to the size of the adaptive array.

Author dc:creator

• BALABHASKAR G