Point Source Localization with a Planar Optical Phased Array Compressive Sensor
Julian A. Brown,Steven J. Spector,Michael Moebius,Lucas Benney,Daniel Vresilovic,Brian Dolle,Alexandra Z. Greenbaum,Alex Huang,Christopher Poulton,Michael R. Watts,Robin Dawson,Benjamin F. Lane,J. P. Laine,Kerri Cahoy,Hannah A. Clevenson
DOI: https://doi.org/10.1117/12.2594214
2021-01-01
Abstract:Compressive sensing has been used to demonstrate scene reconstruction and source localization in a wide variety of devices. To date, optical compressive sensors have not been able to achieve significant volume reduction relative to conventional optics of equivalent angular resolution. Here, we adapt silicon-photonic optical phased array technology to demonstrate, to our knowledge, the first application of compressive imaging in a photonic-integrated device. Our novel sensor consists of an 8 x 8 grid of grating couplers with a spacing of 100 mu m. Path-matched waveguides route to a single multimode interferometer (MMI), which mixes and randomizes the signals into 64 outputs to be used for compressed sensing. Our device is fully passive, having no need for phase shifters, as measurement matrix calibration makes the measurements robust to phase errors. For testing, we use an Amplified Spontaneous Emission (ASE) source with a bandwidth of 40 nm, centered at 1545 nm. We demonstrate simultaneous multi-point (2 sources demonstrated in this work) brightness recovery and localization with better than 10 arcsecond precision in a sub-millimeter thick form-factor. We achieve a single source recovery rate higher than 99.9% using 10 of the 64 outputs, and a 90% recovery rate with only 6 outputs, 10 times fewer than the 64 needed for conventional imaging. This planar optical phased array compressive sensor is well-suited for imaging sparse scenes in applications constrained by form factor, volume, or high-cost detectors, with the potential to revolutionize endoscopy, beam locators, and LIDAR.