Channel-noise tracking for sub-shot-noise-limited receivers with neural networks
February 24, 2021 - Matt DiMario
(a) A sender (Alice) and receiver (Bob) attempt to communicate across a noisy channel. Bob implements a state discrimination measurement based on photon counting to decode the information sent by Alice, and uses a neural network to track and correct for channel noise. (b) Example of the error probability as a function of time when applying phase and amplitude noise to the input states. The noise tracking algorithm based on a neural network estimator (blue) achieves equivalent performance to a strategy based on a far more computationally expensive Bayesian estimator (orange).
Non-Gaussian receivers for optical communication with coherent states can achieve measurement sensitivities beyond the limits of conventional detection, given by the quantum-noise limit (QNL). However, the amount of information that can be reliably transmitted substantially degrades if there is noise in the communication channel, unless the receiver is able to efficiently track and compensate for such noise. Furthermore, current approaches to solve these issues of noise tracking in conventional measurement receivers do not apply to non-Gaussian strategies. We investigate the use of a deep neural network as a computationally efficient estimator of multiple channel noise parameters, which enables a reliable method of noise tracking for non-Gaussian receivers. Using numerical simulations, we find that this noise tracking method allows the non-Gaussian receiver to maintain its benefit over the QNL in the presence of channel noise. We also compare this neural network based method to a far more computationally expensive Bayesian estimator, which may be impossible to realize at the necessary bandwidths for optical communication. In addition, the noise tracking method based on neural networks can easily include other types of noise to ensure sub-QNL performance in channels with many noise sources.