"The problem of estimating a surface from a set of discrete measurements lying along straight lines is considered. This situation arises when one attempts to determine the S-Band antenna gain pattern for the Space Shuttle, from measurements taken at several ground stations. The results of previous investigators concerned with the performance of surface approximation techniques for the present application, are extended in this study by examining the case where the data samples are corrupted by measurement noise. Results have been obtained using least-squares approximation with bicubic B-spline basis functions, and for an interpolation algorithm in conjunction with a spatial smoothing filter. Because of the nature of the data acquisition and the impracticality of the least-squares algorithm when many sample points are used, the application of the Kalman filter to the surface estimation problem is discussed, although no numerical results were obtained using this approach. It is shown that a direct application of Kalman filter theory yields a filter algorithm which would be extremely difficult to implement. Based on the applications of reduced-order, suboptimal filters to image processing, a suboptimal approximation to the Kalman filter, applied to the surface estimation problem, is considered. The use of a decentralized estimation approach to this problem is briefly examined"--Abstract, page ii.
Tranter, William H.
Stuller, John A.
Pagano, Sylvester J., 1924-2006
Electrical and Computer Engineering
M.S. in Electrical Engineering
University of Missouri--Rolla
xix, 212 pages
© 1984 Paul Matthew Niall Feehan, All rights reserved.
Thesis - Open Access
Print OCLC #
Link to Catalog Record
Feehan, Paul Matthew Niall, "An investigation of methods of surface estimation with application to the interpolation of antenna patterns" (1984). Masters Theses. 192.