PaperlayerView exact revision on arXiv

Source-linked AI summary

The Quantum Illumination Story

Jeffrey H. Shapiro

arXiv:1910.12277v4quant-ph

TL;DR

The paper asks whether quantum techniques can improve radar sensing and traces quantum illumination, which combines a retained idler with returned light to detect a target. It concludes that QI can improve target detection on an entanglement-breaking channel under an energy constraint, while noting important limits on when that advantage applies.

  • Problem

    The paper asks whether quantum techniques can improve radar sensing, where loss and noise can destroy fragile quantum properties.

    Source: INTRODUCTION · p. 1

  • Method

    Quantum illumination sends entangled signal-idler pairs and decides target presence by measuring the retained idler together with returned light.

    Source: LLOYD’S QUANTUM ILLUMINATION · p. 1

  • Results

    QI is the first bosonic system shown to improve performance on an entanglement-breaking channel under a transmitter energy constraint.

    Source: CONCLUDING REMARKS · p. 11

  • Takeaways & Limitations

    The authors argue that entanglement should not be dismissed in lossy, noisy microwave-radar settings and that this lesson may lead to other useful quantum systems.

    Source: CONCLUDING REMARKS · p. 11

    QI’s OPA receiver offers no performance gain for Rayleigh-fading targets, although the cited advantage assumes a return with known amplitude and phase.

    Source: QI RECEIVERS AND EXPERIMENTS · p. 7

Abstract

from arXiv · show

Superposition and entanglement, the quintessential characteristics of quantum physics, have been shown to provide communication, computation, and sensing capabilities that go beyond what classical physics will permit. It is natural, therefore, to explore their application to radar, despite the fact that decoherence---caused by the loss and noise encountered in radar sensing---destroys these fragile quantum properties. This paper tells the story of "quantum illumination", an entanglement-based approach to quantum radar, from its inception to its current understanding. Remarkably, despite loss and noise that destroy its initial entanglement, quantum illumination does offer a target-detection performance improvement over a classical radar of the same transmitted energy. A realistic assessment of that improvement's utility, however, shows that its value is severely limited. Nevertheless, the fact that entanglement can be of value on an entanglement-breaking channel---the meta-lesson of the quantum illumination story---should spur continued research on quantum radar.

INTRODUCTION

The introduction frames superposition and entanglement as quantum phenomena with growing applications in computing, communication, and sensing. It asks whether quantum techniques can improve radar and identifies quantum illumination as the paper’s focus.

INTRODUCTION

  • Quantum computers, quantum key distribution, and squeezed-state light in LIGO illustrate the movement of quantum techniques into computing, communication, and sensing.The passage also notes that large-scale quantum computers could break Internet public-key infrastructure, while quantum key distribution may counter that threat.Source: INTRODUCTION · p. 1
  • The introduction asks whether quantum techniques can bring performance gains to radar sensing.Source: INTRODUCTION · p. 1
  • The paper follows quantum illumination, an entanglement-based avenue of quantum radar research, from its inception to the present.Source: INTRODUCTION · p. 1

LLOYD’S QUANTUM ILLUMINATION

Lloyd’s quantum illumination compares single-photon radar with an entangled signal-idler approach for detecting a weak target in strong background noise. Its predicted advantages depend on operating regime and idealized assumptions, and later work compared it with classical radar.

LLOYD’S QUANTUM ILLUMINATION

  • Lloyd coined “quantum illumination” for entanglement-based detection of a weakly reflecting target in background noise that can greatly exceed the return.His work built on Sacchi’s earlier studies of quantum operation discrimination,.Source: LLOYD’S QUANTUM ILLUMINATION · p. 1
  • The SP receiver decides from returned light, whereas QI decides using both the returned light and a retained idler entangled with each signal pulse.Both scenarios use N high-time-bandwidth-product single-photon signal pulses to test for a weakly reflecting target against always-present background light.Source: LLOYD’S QUANTUM ILLUMINATION · p. 1, Figure 1 · p. 2
  • Lloyd’s analysis assumes low signal transmissivity, low background brightness, and MNB ≪ 1, so at most one photon returns per signal pulse.Source: LLOYD’S QUANTUM ILLUMINATION · p. 1
  • QI’s good regime extends to κ ≫ NB/M, while SP’s requires κ ≫ NB, giving QI an advantage over a broader parameter region.In their good regimes, the two scenarios have the same error-probability expression.Source: LLOYD’S QUANTUM ILLUMINATION · pp. 1–2
  • In the bad regime, QI has a factor-of-M greater error-probability exponent than SP, even though background noise destroys the initial entanglement.At optical frequencies, the passage gives a 60 dB higher effective signal-to-noise ratio for Lloyd’s bad-regime QI at M = 106.Source: LLOYD’S QUANTUM ILLUMINATION · p. 2
  • The predicted enhancement rests on idealized assumptions: on-demand high-time-bandwidth entangled pairs, lossless idler storage, and a perfectly realized optimum receiver.These assumptions motivated follow-on research despite the idealized analysis.Source: LLOYD’S QUANTUM ILLUMINATION · p. 2
  • Shapiro and Lloyd found coherent-state radar matches Lloyd’s QI in its good regime and performs better in its bad regime.Source: LLOYD’S QUANTUM ILLUMINATION · p. 3
  • Tan et al. showed Gaussian-state QI outperforms all classical radars of the same transmitted energy.Source: LLOYD’S QUANTUM ILLUMINATION · p. 3

TAN ET AL.’S QUANTUM ILLUMINATION

Tan et al.’s quantum-illumination radar uses low-brightness entangled signal-idler pulses and compares target-detection error bounds against coherent-state radar. Its advantage is supported in principle, but depends on operating conditions and is not accessible with conventional optical receivers.

TAN ET AL.’S QUANTUM ILLUMINATION

  • The CS radar uses a coherent-state pulse and decides from returned light, while QI uses SPDC-generated entangled signal-idler pulses and observes both the return and retained idler.Both use M = TW ≫1 and NS ≪1; the target is weakly reflecting and lies in always-present, high-brightness background light.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 3, Figure 2 · p. 4
  • Tan et al.’s bad-regime QI radar improves the error-probability exponent by a factor of four (6 dB) over equal-energy CS radar, regardless of M ≥106.This is smaller than Lloyd’s M-fold bad-regime exponent improvement.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 3
  • At sufficiently high M, Fig. 3’s QI error-probability upper bound falls below the CS lower bound, supporting in-principle improvement over all classical radars at equal transmitted energy.The QI bound is exponentially tight, whereas the Bhattacharyya lower bound for CS is always loose.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 3, Figure 3 · p. 4
  • With NS ≪1, the target-present return-idler cross-correlation is the QI detection signature and greatly exceeds the classical-state limit.When NS ≫1, the signature is only slightly better than the classical limit; QI’s preferred regime has low signal brightness and high background brightness.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 5
  • Homodyne, heterodyne, direct detection, and photon-coincidence counting do not provide a QI advantage over CS in the stated high-background setting.Heterodyne adds noise when measuring both quadratures, while background light makes idler coincidences occur whether the target is present or absent.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 5
  • Guha and Erkmen offered an optical parametric amplifier receiver as a partial solution to measuring QI’s target-presence signature.The phase-sensitive return-idler correlation cannot be measured in second-order interference.Source: TAN ET AL.’S QUANTUM ILLUMINATION · p. 5

QI RECEIVERS AND EXPERIMENTS

Receiver analyses distinguish ideal QI performance from receiver constraints: OPA offers a partial advantage, while an ideal feed-forward SFG receiver reaches a 6 dB effective-SNR improvement, with practical utility bounded by demanding assumptions and target fading.

QI RECEIVERS AND EXPERIMENTS

  • Low-gain OPA reception uses photon counting, with phase conjugation making QI’s phase-sensitive cross-correlation observable in the idler output.The receiver’s optimum gain is close to unity because the returned background is bright relative to the retained idler.Source: QI RECEIVERS AND EXPERIMENTS · p. 6
  • 3 dB error-probability exponent improvement over Tan et al.’s same-energy coherent-state system is available to OPA reception in QI’s usual operating regime.This assumes a weak signal, bright background, and low target transmissivity.Source: QI RECEIVERS AND EXPERIMENTS · p. 6
  • The tabletop OPA experiment achieved a 20% SNR improvement, equivalent to a 20% error-probability exponent advantage, over a same-energy coherent-state system.A separate SPDC photon-coincidence experiment did not exploit entanglement, while another operated where QI offers no advantage over coherent-state radar.Source: QI RECEIVERS AND EXPERIMENTS · p. 6
  • OPA’s mode-pair measurements are LOCC and suboptimal for mixed-state hypothesis testing, while coherent SFG across mode pairs offers a route beyond that limit.The proposed SFG receiver assumes 100% photon-pair conversion efficiency, beyond current nonlinear-optics capability.Source: QI RECEIVERS AND EXPERIMENTS · p. 6
  • 6 dB effective-SNR improvement over same-energy coherent-state operation is the quantum-ROC advantage of the ideal feed-forward SFG receiver.Its receiver operating characteristic plots detection probability against false-alarm probability.Source: QI RECEIVERS AND EXPERIMENTS · p. 7
  • For Rayleigh-fading targets, OPA offers no performance gain; SFG still outperforms classical operation, but its subexponential advantage is more vulnerable to nonidealities.The usual performance advantages assume known target-return amplitude and phase, unlike many lidar returns.Source: QI RECEIVERS AND EXPERIMENTS · p. 7

CONCEPT AND REALITIES

Quantum illumination drew radar interest after a microwave proposal brought its predicted advantage into a naturally high-background band, but practical limits sharply constrain that advantage. Idler-storage loss, time-bandwidth demands, and resolution-bin restrictions all narrow its operating range.

CONCEPT AND REALITIES

  • At optical wavelengths, the high-background regime in which Tan et al. found an advantage is unrealistic, while naturally high background at microwave frequencies helped renew radar interest.The expected optical-frequency advantage would accrue only with bright-light jamming.Source: CONCEPT AND REALITIES · p. 7
  • Barzanjeh et al.’s microwave system entangles a transmitted microwave signal with an optical idler, then upconverts the return for a phase-conjugate joint measurement.An ideal phase-conjugate receiver offers a 3 dB error-probability-exponent advantage, but system nonidealities can erode it.Source: CONCEPT AND REALITIES · p. 7, Figure 4 · p. 8
  • 6 dB of idler loss eliminates the FF-SFG or SFG advantage, while 3 dB eliminates the OPA or PC advantage.Barzanjeh et al. noted that optical-fiber delay-line storage precludes an advantage for targets more than 11.25 km away.Source: CONCEPT AND REALITIES · p. 8
  • Microwave QI has difficulty achieving the high time-bandwidth product needed for useful microwave pulse durations, and its advantage degrades as transmitter power rises at fixed pulse duration and bandwidth.A 1 µs pulse has a 10^6 time-bandwidth product at 1 µm but only 10^2 at 1 cm under the stated fractional bandwidths.Source: CONCEPT AND REALITIES · p. 8
  • QI interrogates one resolution bin at a time; splitting the idler across bins costs 10 log10(KB) dB per bin, and target extent or motion within a pulse can further reduce performance.At optical QI’s 1 THz bandwidth, a target range extent below 1 mm is needed for κm ≈1.Source: CONCEPT AND REALITIES · p. 8
  • For equally likely target absence or presence, the cited transmitter-state result gives strong evidence that a 6 dB QI error-probability-exponent advantage over a same-energy coherent-state transmitter is the maximum possible.The cited work identifies the SPDC state as optimizing the ROC in Tan et al.’s scenario.Source: CONCEPT AND REALITIES · pp. 8–9
  • Initial microwave experiments reported gains over their chosen classical comparisons, but their reception techniques preclude outperforming an optimized classically-correlated-noise radar.The paper calls those setups quantum-correlated-noise radars, distinguishing them from Tan et al.’s OPA-receiver QI radar.Source: CONCEPT AND REALITIES · p. 9

CORRELATED-NOISE RADARS: QUANTUM VERSUS CLASSICAL

The comparison tests quantum-correlated-noise radar against a classical correlated-noise radar with matched signal brightness and detection capabilities. Their performance converges under quantum-limited detection at high idler brightness, and the classical radar can do better with nonideal detectors.

CORRELATED-NOISE RADARS: QUANTUM VERSUS CLASSICAL

  • The QCN radar uses an SPDC source and pre-amplified heterodyne detection, while the CCN radar splits high-brightness classical noise into a low-brightness signal and a bright retained idler.SPDC ties signal and idler brightnesses, whereas the CCN splitter can match the QCN signal brightness while keeping NI ≫1.Source: CORRELATED-NOISE RADARS: QUANTUM… · p. 9
  • With quantum-limited detectors, QCN and CCN have identical conditional statistics as NI →∞, so their performance is expected to be virtually the same for NI ≫1.This conclusion applies to deterministic known target parameters and to random parameters with a known joint distribution.Source: CORRELATED-NOISE RADARS: QUANTUM… · p. 10
  • For nonideal detectors, CCN outperforms QCN when NI > NF (NS + 1)/(NF −1), because its transformed idler modes then have lower noise.The other second moments of the two radars are identical in this comparison.Source: CORRELATED-NOISE RADARS: QUANTUM… · p. 10
  • The analysis concludes that QCN cannot outperform all classical radars with the same transmitted energy and identical detector capabilities.The comparison models target-present and target-absent returns with conditional Gaussian statistics for the mode pairs.Source: CORRELATED-NOISE RADARS: QUANTUM… · p. 10

ROC COMPARISONS

In the stated best-case comparison, QI-OPA outperforms the coherent-state homodyne, correlated-noise, and heterodyne radars, while QCN and CCN are indistinguishable. The ROC comparison uses the same target, background, and transmitted energy across the five radar configurations.

ROC COMPARISONS

  • The five radars compare detection of a κ = 0.01 target in an NB = 20 background using Ntot = 2 × 10^4 photons on average.QCN and CCN use NS = 0.01, M = 2 × 10^6, NI = 10^3, and NF = 1; the coherent-state systems transmit Ntot and use heterodyne or homodyne detection.Source: ROC COMPARISONS · p. 10, Figure 5 · p. 11
  • QI-OPA outperforms CS-Hom, CCN, and CS-Het, while QCN and CCN ROCs are indistinguishable in the comparison.CS-Hom and CCN are also indistinguishable; all systems use quantum-limited detection with a known, deterministic target phase and transmissivity.Source: ROC COMPARISONS · p. 10, Figure 5 · p. 11
  • QCN requires a dilution refrigerator for its Josephson junction parametric amplifier, whereas CCN could use a less expensive room-temperature microwave noise generator.The QCN equipment requirement is stated in the ROC discussion; the CCN source is presented as a possible alternative.Source: ROC COMPARISONS · pp. 10–11
  • The ROCs plot log10(PM) against log10(PF) for PF and PM ≤ 0.5, with QCN, CCN, and QI-OPA curves approximated and the coherent-state ROCs computed exactly.The approximation is reported as accurate for M ≫1, especially for Gaussian signals in Gaussian noise.Source: ROC COMPARISONS · p. 10, Figure 5 · p. 11

CONCLUDING REMARKS

Quantum illumination is not an immediate practical boon to radar, but it establishes that an energy-constrained bosonic system can gain performance on an entanglement-breaking channel. The authors argue this result warrants continued attention to entanglement in lossy, noisy settings.

CONCLUDING REMARKS

  • QI is the first reported bosonic system to improve performance on an entanglement-breaking channel under a transmitter energy constraint.The authors caution that QI target detection is not an immediate boon to the radar community, while arguing that its meta-lesson may lead to useful quantum systems.Source: CONCLUDING REMARKS · p. 11

Appendix A: Entangled State for Lloyd’s QI System

Lloyd’s QI analysis assumes a high time-bandwidth product and restricts each pulse pair to vacuum and single-photon states. Its signal is a single-photon superposition across M temporal modes.

Appendix A: Entangled State for Lloyd’s QI System

  • For M = TW ≫ 1, Lloyd’s signal and idler temporal modes are specified using pulse-pair index n and mode index m.The repetition period satisfies Tr > T, and M is an odd integer.Source: Appendix A · p. 11
  • Lloyd’s QI analysis restricts the temporal-mode state space to vacuum and single-photon states.The nth signal-idler pulse pair is in a quantum state within this space.Source: Appendix A · p. 11
  • Lloyd’s single-photon signal is equally likely to occupy any of its M temporal modes.The single photon is in a superposition across those modes.Source: Appendix A · p. 11

Appendix B: Entangled State for Tan et al.’s QI System

Tan et al.’s QI system uses a high time-bandwidth product and does not impose Lloyd’s vacuum-and-single-photon restriction. Its signal-idler modes have a joint state that includes photon-number states.

Appendix B: Entangled State for Tan et al.’s QI System

  • For M = TW ≫ 1, Tan et al.’s signal-idler pulse pair has a specified temporal-mode structure.The mode index satisfies |m| ≤ (M−1)/2, with M an odd integer.Source: Appendix B · p. 12
  • Tan et al. do not restrict their temporal modes to the span of vacuum and single-photon states.This differs from Lloyd’s QI analysis.Source: Appendix B · p. 12
  • Tan et al.’s joint state is given for each signal-idler mode pair, whose photon-number states are denoted by |n⟩Sm and |n⟩Im.The passage contrasts this state with the coherent state of Tan et al.’s transmitter.Source: Appendix B · p. 12

Appendix C: Phase-Insensitive and Phase-Sensitive Cross Correlations

Although loss and noise leave the returned and retained idler light in a classical state, the analysis uses semiclassical photodetection theory. The section distinguishes two cross-correlations and explains why ordinary interferometric photon counting does not reveal the target signature.

Appendix C: Phase-Insensitive and Phase-Sensitive Cross Correlations

  • Loss and noise destroy the initial signal-idler entanglement, leaving returned and retained idler light in a classical state for semiclassical photodetection analysis.Source: Appendix C · p. 12
  • The phase-insensitive cross-correlation is unchanged by applying a phase shift θ to both modes, whereas the phase-sensitive correlation’s phase shifts by 2θ.Source: Appendix C · p. 12
  • Photon counting after combining returned and idler modes on a 50–50 beam splitter provides no target-presence signature because the phase-insensitive correlation is zero under either hypothesis.Source: Appendix C · p. 12
  • Tan et al.’s QI system therefore needs a way to measure the phase-sensitive cross-correlation signature.Source: Appendix C · p. 12

Appendix D: Near-Equivalence of the Quantum-Limited CS-Hom and CCN Radars

The CCN radar starts with independent signal-idler pairs drawn from classically random coherent-state mixtures. In the stated high-idler-brightness limit, heterodyne detection and maximal-ratio combining yield a ROC near-equivalent to the quantum-limited CS-Hom radar’s under specified conditions.

Appendix D: Near-Equivalence of the Quantum-Limited CS-Hom and CCN Radars

  • The CCN source produces M independent, identically distributed signal-idler pairs in classically random mixtures of coherent states.Source: Appendix D · p. 12
  • In the limit NI → ∞, quantum-limited heterodyne detection of the idler modes measures their coherent-state eigenvalues without noise.The post-heterodyne processing uses an optimum maximal-ratio combiner.Source: Appendix D · p. 12
  • For NI ≫ 1 and M ≫ 1, the CCN radar’s unconditional ROC matches the CS-Hom radar’s when NB ≫ 1.Source: Appendix D · p. 12
  • The near-equivalence also applies with noise figures NF > 1 when NF ≪ NB.Source: Appendix D · p. 12
Loading 1910.12277v4…