Autonomous Optical Alignment of Satellite-Based Entanglement Sources using Reinforcement Learning

TL;DR

Global-scale quantum communication via satellites hinges on robust, autonomous generation of high-quality entanglement, which is challenged by misalignment under dynamic orbital conditions. The authors propose two recalibration strategies for a PPLN-based SPDC source: a heuristic alignment (HA) method and a reinforcement learning (RL) approach, evaluated with a time-aware, modified ROC-AUC metric. Using Soft Actor-Critic within an explicit MDP formulation, the RL agent outperforms HA, achieving $AUC_{RL}^{max}=0.9119$ vs $AUC_{HA}^{max}=0.7042$ and near-perfect alignment around $10$ minutes compared to ~30 minutes for HA. The work delivers a concrete, onboard-compatible calibration framework with detailed state/action/reward definitions and demonstrates a practical path to autonomous satellite quantum calibration, enhancing the viability of scalable space-based quantum networks.

Abstract

Quantum entanglement distributed via satellites enable global-scale quantum communication. However, onboard sources are susceptible to misalignment due to dynamical orbital conditions. Here, we present two recalibration techniques for efficient generation of high quality entanglement using a periodically poled lithium niobate (PPLN)-based spontaneous parametric down-conversion (SPDC) source with minimum intervention. The first is a heuristic algorithm (HA) which mimics the manual alignment process in a laboratory. The second is based on reinforcement learning (RL). Our simulation demonstrates superior performance of RL with AUC=0.9119 compared to HA's 0.7042 in the modified ROC analysis (60 min threshold). RL achieves perfect alignment in 10 min as opposed to HA's 30 min. Both the methods operate within feasible satellite constraints, offering scalable automation for complex quantum communication scenarios.

Figures (8)

• Figure 1: Modified ROC curves showing the accuracy metric defined in Eq. \ref{['eq:A']} as a function of calibration time.
• Figure 2: End-to-end optical coupling between two fibers. The input fiber is adjusted axially ($Z$) and transversely ($XY$) relative to a fixed output fiber to maximize coupling efficiency.
• Figure 3: The dependence of OPO on temperature is obtained from the spatial and spectral characteristics of the signal and the idler photon pairs. The plots represent the wavelength, opening angle and wavevector mismatch of these photons as a function of temperature respectively. For temperatures above $25^{\circ}C$, the wavelengths of the signal and the idler photons are farther apart and their opening angle is $0$. The logarithm of the wavevector mismatch indicates that efficiency of SPDC at higher temperatures reduces.
• Figure 4: The dependence of SPDC on temperature variation is exhibited by the mean, mode, standard deviation and the full width at half maximum of the spectral distribution of the photons. The brightness estimation of the source demonstrates better performance of the source approximately above $25^{\circ}c$. Below it, SPDC vanishes.
• Figure 5: Real part of the biphoton wavefunction at room temperature obtained from numerical simulations of the SPDC process. The imaginary part is zero due to the symmetric phase-matching conditions and the absence of pump spectral phase. Because the pump is assumed to be monochromatic, the biphoton state is effectively 1D and its amplitude depends on the effective phase-matching function of the PPLN crystal.