Distinguishing between optical coherent states with imperfect detection

Several proposed techniques for distinguishing between optical coherent states are analyzed under a physically realistic model of photodetection. Quantum error probabilities are derived for the Kennedy receiver, the Dolinar receiver, and the unitary rotation scheme proposed by Sasaki and Hirota for subunity detector efficiency. Monte Carlo simulations are performed to assess the effects of detector dark counts, dead time, signal processing bandwidth, and phase noise in the communication channel. The feedback strategy employed by the Dolinar receiver is found to achieve the Helstrom bound for subunity detection efficiency and to provide robustness to these other detector imperfections making it more attractive for laboratory implementation than previously believed.