Hybrid optical and electronic signal processing for ultra-wideband RF antenna arrays

We extend two existing narrowband beamforming techniques for ultra-wideband applications: minimum mean square error (MMSE) beamforming and MMSE beamforming subject to the constraint on complete ing of interfering signals. Assuming that the signal at each antenna element is processed by a linear time invariant (LTI) filter, we derive the optimal filters for each beamformer. We then focus on the scenarios in which the number of antenna elements L is much larger than the number of users K in the system, e.g. multi-element antenna arrays on satellites. To improve on the expensive baseline implementation in which one electronic LTI filter is used for each antenna element, for a total of L filters, we propose the use of simple optical signal processing to reduce the cost of electronic hardware. We show that, by performing conventional beamforming in optics using fiber delay lines as a preprocessing step, the number of electronic filters required for both MMSE and constrained MMSE beamforming is reduced to K. The key observation is that, for L>K, the number of filters depends on the number of users K and not on the number of antenna elements L.