Bayesian Greedy Receiver for Pulse Position Modulation without an Error Floor under Thermal Noise

We investigate quantum receiver architectures for demodulation $M$-ary Pulse Position Modulation under thermal noise and photon-starved conditions. Building on the greed receiver framework, we analyze it using jointly optimized displacement-squeezing and the Dolinar receiver. We further introduce a novel slicing Bayesian greedy receiver, which partitions each PPM time slot into multiple optical slices and update the full posterior distribution over codeword hypothesis at each slice using Bayes' rule. Under the presence of noise, our slicing greedy receiver outperforms the standard greedy receiver and appears to avoid a noise floor given sufficient slicing resolution.