Partially constrained adaptive beamforming for super-resolution at Low SNR
Erik David Hornberger, Shannon D. Blunt, Thomas P. Higgins · 2015
The reiterative super-resolution (RISR) algorithm was previously developed to enable adaptive beamforming with as few as one time snapshot, is robust to temporally correlated signals, and accounts for array calibration errors. Here a gain-constrained version (denoted GC-RISR) is derived followed by a partially-constrained version (PC-RISR). It is shown that an interesting trait of the latter is spatial super-resolution at SNR values lower than is typical for adaptive beamforming techniques as a trade-off for requiring more iterations to converge. The PCRISR formulation is controlled by a selectable parameter that serves a role similar to that of an adaptive step-size which balances between convergence speed and accuracy.