Performing homodyne detection at one port of squeezed-state light interferometer and then binarzing measurement data are important to achieve super-resolving and super-sensitive phase measurements. Here we propose a new data-processing technique by d
ividing the measurement quadrature into three bins (equivalent to a multi-outcome measurement), which leads to a higher improvement in the phase resolution and the phase sensitivity under realistic experimental condition. Furthermore, we develop a new phase-estimation protocol based on a combination of the inversion estimators of each outcome and show that the estimator can saturate the Cramer-Rao lower bound, similar to asymptotically unbiased maximum likelihood estimator.
