Linear Mean Square Interpolation of Missing Samples
Cheryl Jaffe · 2006
The problem of missing samples is described in the context of a beamforming operation. Tapering a complete, uniformly spaced sequence is shown to suppress beam sidelobes, but the taper fails to suppress sidelobes when uniformity of sample spacing is destroyed by missing samples. A linear mean square estimator (LMSE) is employed to interpolate the missing samples, thereby regaining sidelobe suppression afforded by the taper. The algorithm is described, and results are compared to several common interpolation techniques. The number and configuration of missing samples that can be simultaneously reconstructed in this manner is discussed as part of a broader discussion of the robustness of the algorithm