Fast self-testing quantum random number generator based on homodyne detection

Abstract

Self-testing and semi-device independent protocols are becoming the preferred choice for quantum technologies, being able to certify their quantum nature with few assumptions and simple experimental implementations. In particular, for quantum random number generators, the possibility of monitoring, in real time, the entropy of the source only by measuring the input/output statistics is a characteristic that no other classical system could provide. The cost of this possibility is not necessarily increased complexity and reduced performance. Indeed, here we show that with a simple optical setup consisting of commercially available components, a high bit generation rate can be achieved. We manage to certify 145.5 MHz of quantum random bit generation rate.

Quantum technologies in the telecommunications industry

“…the light source) are not explicitly modelled and the security is derived from the challenge-response (input/output) data analysis. There are different proposals in the literature [56,57] using assumptions as bounded energy or bounded dimensions, among others. In brief, QRNGs will either be discrete variable or continuous variable with respect to the physical carrier employed and use VPM, CRP, or mixed VPM/CRP methods for the entropy estimation process.…”

Semi-device-independent randomness from d -outcome continuous-variable detection

“…By homodyning, the quantum state is measured from samples obtained from projected Wigner functions, whereas heterodyne detection directly samples phase space coordinates from the Husimi Q-function [26,27]. For what regards Semi-DI QRNG protocols, both heterodyne and homodyne detection have been employed at the receiver side, as shown in [16] and [15], respectively. In these works, the (potentially) infinite outcomes of the CV measurement are grouped into two disjoint sets, corresponding to a binary outcome.…”

“…Semi-DI protocols are an intermediate approach between DI and trusted-device schemes, which offer an optimal trade-off between generation rate, security, and easeof-implementation [4]. Depending on the protocol needs, assumptions can vary; for very secure protocols, there are fewer assumptions on the device, i.e., single assumption on the overlap or energy of the prepared states [11][12][13][14][15][16]. Depending on the protocol needs, assumptions can vary; they can be related to the dimension of Hilbert space [17,18], they may require trusted measurement in the case of source-DI protocols [19][20][21] or they may assume a trusted source, an in measurement-DI protocols [22,23].…”

“…Depending on the protocol needs, assumptions can vary; they can be related to the dimension of Hilbert space [17,18], they may require trusted measurement in the case of source-DI protocols [19][20][21] or they may assume a trusted source, an in measurement-DI protocols [22,23]. Recently a new class of protocols has been proposed, where both source and measurement are untrusted and only a single assumption on the overlap or energy of the prepared states is required [11][12][13][14][15][16]. These protocols can provide an increased security, since they reduce the number of assumptions on the devices.…”

“…In this work, we investigate the impact of increasing the number of outcomes of the measurement apparatus given a binary-input semi-DI QRNG [15,16]. The protocol builds upon the prepare-and-measure scheme, with a single assumption on the maximum energy of the prepared states that implies a lower bound on the state's overlap.…”