On the power of moving quantum sensors: fully flexible and noise-resilient sensing

Paul Aigner, Wolfgang Dür · 2025

Read On the power of moving quantum sensors: fully flexible and noise-resilient sensing by Paul Aigner, Wolfgang Dür (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-15876.txt.

Introduction

Paul Aigner and Wolfgang Dur show that a single moving quantum sensor can outperform a fixed sensor network in a 2025 paper. Quantum sensors measure fields with precision beyond classical limits, and networks of them are usually assumed stronger. Motion, though, gives one sensor access to many positions over time.

With control of its path or its internal state, one moving sensor can pick out any linear functional of a spatially correlated scalar field. Examples include a gradient or a Fourier coefficient. At the same time it can cancel all noise signals with orthogonal spatial correlation. Fixed entangled networks can suppress only as many noise signals as they have positions. Moving sensors also beat the static scaling limit for quantum Fisher information.

Catalog record

CREATORPaul Aigner, Wolfgang Dür
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-15876

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On the power of moving quantum sensors: fully flexible and noise-resilient sensing Paul Aigner, Wolfgang Dür Abstract We show that a single moving quantum sensor provides complete access to spatially correlated scalar fields. We demonstrate that with either trajectory or internal state control, one can selectively measure any linear functional, e.g. a gradient or a spatial Fourier series coefficient, while successfully eliminating all noise signals with orthogonal spatial correlation. This even exceeds the capabilities of a sensor network consisting of multiple entangled, yet spatially fixed, quantum sensors, where the number of suppressed noise signals is limited by the number of sensor positions. We show that one can achieve an improved scaling of the quantum Fisher information for moving sensors beyond the static fundamental limit of T 2 T^{2} . Introduction.— Quantum metrology exploits quantum mechanical effects to estimate parameters with precision beyond classical limits 14 ; 25 ; 9 . In settings where the imprinting Hamiltonian is time independent, the optimal precision is determined by the quantum Fisher information (QFI) which has the familiar T^{2} time scaling limit per probe, with entanglement offering Heisenberg scaling N^{2} in the sensor number N when noise is absent 14 ; 25 . Recently, progress has been made on using distributed probes across networks to sense spatially varying fields 19 ; 12 ; 27 , and on mitigating decoherence with techniques ranging from decoherence-free subspaces (DFS) 29 ; 15 to quantum error correction (QEC) and fast control 18 ; 11 ; 10 ; 30 . However, most current approaches assume that sensors remain in fixed positions and, hence, their output is determined only by the field values at those stationary points. This perspective implicitly constrains both which functionals of a field can be accessed and how many independent noise contributions can be cancelled 29 ; 27 . Here, in contrast, we consider mobile quantum sensors whose position and internal dynamics can be shaped by designing a trajectory and quantum control, respectively. We demonstrate that the imprinted phase becomes an engineered linear functional of the field along the path, which enables selective readout of arbitrary linear functionals (e.g., gradients or a spatial Fourier component) and the cancelation of any number of fluctuating signals whose spatial correlations are linearly independent of the signal. Those are capabilities that strictly exceed those of static networks with the same number of physical qubits. Mobility also fundamentally changes the metrological time scaling. Motion makes the imprinting Hamiltonian time dependent, where with adaptive control one can surpass the static T^{2} scaling limit of the QFI 24 . We show this mechanism with a minimal moving sensor example and demonstrate that a moving sensor can have an increased QFI growth rate. In addition, we compare mobile sensing against the best static network strategies for dealing with fluctuating noise signals. Even in regimes where networks can cancel noise, a single N -qubit sensor that sequentially visits the N positions of a fixed position network can achieve a strictly larger QFI than the sensor network. Finally, we extend our constructions to fields with explicit time dependence, showing how adaptable velocity schedules and quantum control decouple spatial weighting from temporal reshaping, enabling noise cancelation. Moving quantum systems where the quantum information carriers can be physically transported or rearranged over short distances in a controlled way are natively available on several quantum computing platforms. In certain trapped ion processors, one can shuttle ions between zones in segmented traps 26 ; 23 ; 2 . Furthermore, neutral atom computers similarly allow atoms to be repositioned using optical tweezers, making large-scale rearrangements possible 7 ; 5 ; 31 ; 34 ; 3 . Such systems can also be used as quantum sensors, and hence our approach not only promises theoretical advantages but could be readily implemented in existing platforms. Other conceivable examples of moving sensors include quantum sensors mounted on satellites, vehicles, or drones. While for the former trajectories are largely fixed, the latter offer also trajectory control. Taken together, these results establish motion as a powerful and largely under-explored resource for quantum metrology. By designing trajectories and quantum control, one gains full access to linear functionals of spatially (and temporally) correlated signals, achieves robust noise cancelation beyond static limits, and enables QFI time scaling inaccessible to stationary sensors. Background.— We briefly summarize the main theoretical concepts and figures of merit for single-parameter quantum metrology in the Fisher regime, i.e. the sensing regime of many m\gg 1 measurements 16 . Although most of our results are valid beyond the Fisher regime, we adopt it here for the clarity of presentation and use the QFI as the primary figure of merit. In single-parameter quantum metrology, the objective is to estimate an unknown parameter \omega encoded into a quantum state through an imprinting unitary operation U(\omega)=\exp(-it G \omega), where G is the generator of the parameter-dependent evolution. The precision of any unbiased estimator \hat{\omega} is fundamentally limited by the quantum Cramér–Rao bound, which relates the mean squared error to the QFI as \mathrm{Var}(\hat{\omega})\geq(m{I}_{\omega}^{(Q)}[\rho(\omega)])^{-1} 16 . The QFI quantifies the sensitivity of a quantum state \rho(\omega) to changes in \omega , and is defined for pure states |\psi_{\omega}\rangle=U(\omega)|\psi_{0}\rangle by 6 I^{(Q)}_{\omega}=4t^{2}\left(\langle\psi_{0}|G^{2}|\psi_{0}\rangle-\langle\psi_{0}|G|\psi_{0}\rangle^{2}\right)=4t^{2}(\Delta G)^{2}. (1) For N probe states and interrogation time T one finds that the QFI scaling is limited by T^{2}N^{2} 6 ; 13 . (a) (b) Figure 1: (a) Sketch of a mobile quantum sensor, moving through signal fields, illustrated by blue concentric circles, and noise fields, illustrated by jagged orange blobs. The quantum sensor follows a path, illustrated by the gray dashed line, where it has at different times t_{i} different local velocities, illustrated with black arrows. (b) Plot of the signal generating function f(\boldsymbol{\gamma}(t))=t^{4} , noise generating functions g_{i}(\boldsymbol{\gamma}(t))\in\{t^{0},t^{1},t^{2},t^{3}\} , and corresponding optimal control function \mathrm{sgn}[f(\boldsymbol{\gamma})-p^{*}](t) , with p^{*} being the best L^{1}[0,1] approximation of f(\boldsymbol{\gamma}) out of \mathrm{span}_{\mathbbm{R}}\{t^{0},t^{1},t^{2},t^{3}\} . The control function \mathrm{sgn}[f(\boldsymbol{\gamma})-p^{*}] cancels the contributions of the noise signals g_{i}({\boldsymbol{\gamma}}) , while maximizing the residual sensitivity to the signal function f(\boldsymbol{\gamma}) . Setting and general methodology.— We consider the estimation of a signal described by the Hamiltonian of the form H_{\text{signal}}=f(\boldsymbol{x}_{\text{sensor}}(t))G, where G is a hermitian generator with symmetric spectrum, \boldsymbol{x}_{\text{sensor}}(t) is the position of the sensor at time t , and we want to estimate some property of the position dependent scalar field f:\mathbbm{R}^{d}\rightarrow\mathbbm{R} . Assuming our sensor position follows a path \boldsymbol{\gamma}:[0,T]\rightarrow\mathbbm{R}^{d} , we obtain the imprinting unitary U_{\Phi}=\exp(-i\Phi G),\ \Phi=\int_{0}^{T}f(\boldsymbol{\gamma}(t))dt. (2) We show that by appropriately controlling the sensor trajectory and phase imprinting via sign-control, any linear functional of the field f(\boldsymbol{\gamma}) can be realized up to scaling. Our methodology is state independent and therefore applicable in both Fisher and Bayesian estimation settings. Fixing the geometric path \gamma([0,T]) , we first vary only the velocity schedule by introducing a smooth, strictly increasing reparametrization h:[0,T]\rightarrow[0,T] , yielding \Phi_{h}=\!\int_{0}^{T}f(\boldsymbol{\gamma}(h(t)))dt=\!\int_{0}^{T}f(\boldsymbol{\gamma}(\tau))w(\tau)d\tau, (3) where w(\tau)\equiv 1/h^{\prime}(\tau)\geq 0 satisfies \int_{0}^{T}w(\tau)d\tau=T . By the Riesz representation theorem 28 , as already indicated in 27 , appropriate choices of a velocity schedule allow realization of any non-negative linear functional of f(\boldsymbol{\gamma}) . Introducing additional sign control s(t)\in\{0,1\} gives \Phi_{w,s}=\int_{0}^{T}f(\boldsymbol{\gamma}(t))w(t)(-1)^{s(t)}dt, (4) thus enabling the imprinting of arbitrary linear functionals. Similarly, with a fixed velocity schedule but bounded control function c(t)\!\in\![-1,1] , the

Read the complete text