Dual-Comb LIDAR: Fundamentals & Advances

• Dual-comb LIDAR is a ranging method that interferes two optical frequency combs—one as a probe and one as a local oscillator—to map optical delay into radio-frequency signals.
• It achieves high precision and rapid update rates by controlling a small repetition-rate offset, enabling techniques like heterodyne phase-slope ranging and two-photon cross-correlation.
• Applications include long-range target tracking, industrial 3D profiling, and atmospheric sensing, with design trade-offs between update rate, precision, and ambiguity range.

Searching arXiv for papers on dual-comb LiDAR and related implementations. Dual-comb LIDAR is a ranging modality in which two optical frequency combs with slightly different repetition rates are interfered so that optical delay is mapped into a radio-frequency interferogram, enabling absolute distance measurement, high update rates, and precision that can approach interferometric regimes. In the implementations reported across fiber, solid-state, electro-optic, and microresonator platforms, one comb typically serves as a probe or signal and the other as a local oscillator, while the small repetition-rate offset $\Delta f_{\rm rep}$ sets the interferogram repetition period and, depending on architecture, the update rate, aliasing behavior, and non-ambiguity range. Reported variants include heterodyne phase-slope ranging, dead-zone-free linear optical sampling, dual-comb FMCW, differential-absorption lidar, and two-photon cross-correlation schemes that are carrier-phase-insensitive and avoid near-gigasample/s digitization (Camenzind et al., 2024, Trocha et al., 2017, Lukashchuk et al., 2021, Rosas et al., 2024, Wright et al., 2021, Forman et al., 8 Mar 2026).

1. Fundamental operating principle

In dual-comb ranging, two optical frequency combs have repetition rates $f_{\rm rep,1}=f_{\rm rep}$ and $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$, and their interference on a photodetector produces a multi-heterodyne radio-frequency comb whose lines are separated by $\Delta f_{\rm rep}$. A representative expression for the photocurrent is

$$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$

with each RF comb line carrying phase information related to optical delay (Camenzind et al., 2024). In probe/LO language, if one comb interrogates a target and the return is heterodyned against the second comb, the resulting interferogram or RF beat phases encode the round-trip path length.

Several equivalent ranging formalisms appear in the literature. In time-of-flight form,

$d=\frac{c\,\Delta t}{2},$

with $\Delta t$ the round-trip delay (Camenzind et al., 2024). In phase-based dual-comb ranging, a phase shift $\Delta\phi$ at a beat frequency can be converted into distance through

$d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$

as used in single-tone or phase-unwrapped formulations (Soghomonyan et al., 13 Apr 2025). In multi-line synthetic-wavelength interferometry, one fits the phase variation across comb index to recover distance from the slope, exploiting the fact that the phase of line $m$ varies linearly with $f_{\rm rep,1}=f_{\rm rep}$0 and with path length (Trocha et al., 2017, Martin et al., 2022).

The same general principle admits distinct detection modalities. Conventional heterodyne implementations detect RF combs or interferograms directly on fast photodiodes and usually process them with FFTs, Hilbert transforms, phase fitting, or center-of-mass delay extraction (Camenzind et al., 2024, Soghomonyan et al., 13 Apr 2025). Two-photon dual-comb LIDAR instead uses a nonlinear detector so that overlap of sub-picosecond probe and LO pulses generates electrical cross-correlation pulses. Because the two-photon response depends on pulse overlap rather than optical carrier phase, this detection is described as carrier-phase-insensitive and is not constrained by the conventional RF aliasing condition (Wright et al., 2021, Forman et al., 8 Mar 2026).

A recurring architectural distinction is between dual-comb systems built from two separate combs and single-cavity dual-comb lasers. The latter generate two combs in a common cavity, often by polarization multiplexing or bidirectional propagation, so environmental perturbations act as common-mode fluctuations. This suppresses differential timing noise and enhances mutual coherence, which is especially valuable for long interferogram acquisition and dead-zone-free ranging (Zhu et al., 1 Sep 2025, Camenzind et al., 2024, Cuevas, 4 Aug 2025).

2. Distance ambiguity, update rate, and precision trade-offs

A central design parameter is the non-ambiguity range. In formulations based directly on the repetition-rate offset, the maximum unambiguous distance is commonly written as

$f_{\rm rep,1}=f_{\rm rep}$1

or equivalently as $f_{\rm rep,1}=f_{\rm rep}$2 in axial-profiling notation (Zhu et al., 1 Sep 2025, Soghomonyan et al., 13 Apr 2025). The implication is immediate: reducing $f_{\rm rep,1}=f_{\rm rep}$3 increases non-ambiguity range but slows the interferogram repetition rate. Zhu et al. reported that varying $f_{\rm rep,1}=f_{\rm rep}$4 from $f_{\rm rep,1}=f_{\rm rep}$5 down to $f_{\rm rep,1}=f_{\rm rep}$6 changes $f_{\rm rep,1}=f_{\rm rep}$7 from $f_{\rm rep,1}=f_{\rm rep}$8 to $f_{\rm rep,1}=f_{\rm rep}$9 $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$0 (Zhu et al., 1 Sep 2025).

In other architectures, ambiguity is tied instead to the comb repetition rate or the synthetic wavelength associated with RF line spacing. The analysis of repetition-rate-limited dual-comb ranging gives

$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$1

where increasing $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$2 improves precision but shortens the non-ambiguity range (Martin et al., 2022). In integrated dissipative-Kerr-soliton systems, the synthetic wavelength is $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$3, and the ambiguity interval is reported as $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$4 for $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$5 (Trocha et al., 2017). In FMCW dual-soliton microcomb ranging, the chirp period rather than the RF offset sets the unambiguous range through $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$6 for $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$7 (Lukashchuk et al., 2021).

Update rate is usually equal to $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$8 or its inverse relation through one interferogram per $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$9. Camenzind et al. used $\Delta f_{\rm rep}$0, corresponding to a new distance and phase update every $\Delta f_{\rm rep}$1 (Camenzind et al., 2024). In micromachining metrology, $\Delta f_{\rm rep}$2 gives an update time of $\Delta f_{\rm rep}$3 (Soghomonyan et al., 13 Apr 2025). Integrated DKS dual-comb LIDAR reached a minimum acquisition time $\Delta f_{\rm rep}$4 and thus a maximum ranging rate of $\Delta f_{\rm rep}$5 (Trocha et al., 2017). Two-photon continuous-streaming systems reported sustained streaming at $\Delta f_{\rm rep}$6 and burst rates up to $\Delta f_{\rm rep}$7 (Forman et al., 8 Mar 2026).

Precision depends on architecture, bandwidth, SNR, and averaging. Reported figures span several regimes. In free-space dead-zone-free dual-comb ranging, the single-shot time-of-flight precision is around $\Delta f_{\rm rep}$8 on a cooperative target at more than $\Delta f_{\rm rep}$9, and interferometric use of phase improves the single-shot precision to $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$0 (Camenzind et al., 2024). Microresonator soliton ranging achieved $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$1 in a single $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$2 acquisition and $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$3 after $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$4 averaging (Trocha et al., 2017). The electro-optic analysis of system limits reported $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$5 one-shot precision with $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$6 FFTs and a best Allan deviation of $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$7 after averaging to $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$8 (Martin et al., 2022). In two-photon dual-comb imaging at a $$I(t)\propto \Re\bigl\{E_1(t)\,E_2^*(t)\bigr\} \;\approx\;\sum_{m}B_{m}\,\exp\!\bigl[i\,(2\pi\,m\,\Delta f_{\rm rep}\,t + \phi_{m})\bigr],$$9 stand-off, precisions averaged to $d=\frac{c\,\Delta t}{2},$0 after $d=\frac{c\,\Delta t}{2},$1, with accuracies of $d=\frac{c\,\Delta t}{2},$2 to $d=\frac{c\,\Delta t}{2},$3 (Nelmes et al., 13 Mar 2026).

This recurring precision–range trade-off was formalized by a performance factor in which the ratio of precision to non-ambiguity range is independent of repetition rate for a fixed comb envelope and SNR. The relation

$d=\frac{c\,\Delta t}{2},$4

captures that trade-off explicitly (Martin et al., 2022). This suggests that improvements in comb flattening, optical bandwidth, and detection SNR can be as consequential as tuning repetition rates.

3. Source architectures and mutual coherence strategies

Dual-comb LIDAR has been implemented with single-cavity solid-state lasers, polarization-multiplexed and bidirectional fiber lasers, electro-optic combs, and microresonator soliton combs. Each source class imposes a different balance among coherence, tunability, repetition rate, RF bandwidth, and system complexity.

Single-cavity lasers are frequently used to exploit common-mode noise suppression. In the Yb:CaF$d=\frac{c\,\Delta t}{2},$5 free-running dual-comb laser used for dead-zone-free moving-target tracking, the carrier wavelength is $d=\frac{c\,\Delta t}{2},$6, the pulse repetition rate is $d=\frac{c\,\Delta t}{2},$7, and the repetition-rate difference is $d=\frac{c\,\Delta t}{2},$8 (Camenzind et al., 2024). In a polarization-multiplexed Er-doped fiber ring cavity, two orthogonal states of polarization circulate with slightly different group velocities, giving a fundamental repetition rate of $d=\frac{c\,\Delta t}{2},$9 and $\Delta t$0, with RF beat-note SNR $\Delta t$1, linewidth $\Delta t$2, and drift $\Delta t$3 over long operation (Cuevas, 4 Aug 2025). These reported drifts support sub-millimetre ranging in metre-scale ambiguity ranges and in-principle ambiguity lengths of hundreds of kilometres (Cuevas, 4 Aug 2025).

The bidirectional Lyot-filtered fiber laser of Zhu et al. represents a specific advance in repetition-rate-difference control. It uses a $\Delta t$4 long PM fiber ring, bidirectionally pumped at $\Delta t$5, with a thermally controlled bidirectional Lyot filter formed by a $\Delta t$6 PMF segment spliced at $\Delta t$7 at each end (Zhu et al., 1 Sep 2025). The slow and fast polarization-axis combs exhibit thermal sensitivities of $\Delta t$8 and $\Delta t$9, so the differential tuning coefficient is $\Delta\phi$0, corresponding to $\Delta\phi$1 (Zhu et al., 1 Sep 2025). With a heater resolution of $\Delta\phi$2, the absolute uncertainty in $\Delta\phi$3 is $\Delta\phi$4, described as about $\Delta\phi$5 better than a mechanical delay line (Zhu et al., 1 Sep 2025). The reported tuning range is from $\Delta\phi$6 down to $\Delta\phi$7 as temperature varies from $\Delta\phi$8 to $\Delta\phi$9 (Zhu et al., 1 Sep 2025).

Microresonator-based systems occupy the opposite extreme of repetition rate. Dual DKS combs in Si$d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$0N$d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$1 microresonators operated at $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$2 and $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$3 with $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$4 and an optical bandwidth of $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$5 over about $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$6 lines (Trocha et al., 2017). Crystalline MgF$d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$7 multi-resonator stacks produced soliton combs with repetition rates near $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$8 and a difference of $d=\frac{c\,\Delta\phi}{2\pi\,\Delta f_{r}},$9, yielding an RF comb span of about $m$0 (Pavlov et al., 2016). These high-repetition-rate combs support very high update rates and compact integration, but their ambiguity intervals are correspondingly shorter unless compensated by other techniques.

Electro-optic and chirped-comb implementations emphasize deterministic spectral structure and waveform agility. In dual-comb DIAL for greenhouse-gas and wind sensing, a single narrow-linewidth CW laser at $m$1 feeds phase modulators to generate two combs with tooth spacings $m$2 and $m$3, i.e. $m$4, using only three teeth per comb to maximize power per tooth (Rosas et al., 2024). In swept dual-soliton microcomb FMCW LiDAR, a single pump is chirped with a triangular waveform of period $m$5 and excursion $m$6–$m$7, generating synchronously chirped combs at $m$8 and $m$9 or, in a denser-comb configuration, around $f_{\rm rep,1}=f_{\rm rep}$00 and $f_{\rm rep,1}=f_{\rm rep}$01 (Lukashchuk et al., 2021).

4. Measurement topologies and signal processing

Dual-comb LIDAR encompasses several distinct measurement topologies. In linear optical sampling or direct heterodyne ranging, the signal comb is split between a reference and a target path and the returns are interfered with the LO comb. Camenzind et al. used a free-space transceiver with a Wollaston prism that combines orthogonally polarized combs into a common path, plus two reference reflections $f_{\rm rep,1}=f_{\rm rep}$02 and $f_{\rm rep,1}=f_{\rm rep}$03 separated by a fixed delay $f_{\rm rep,1}=f_{\rm rep}$04 so that at least one reference interferogram does not overlap the target interferogram (Camenzind et al., 2024). Both signal combs probe the target and interchange signal/LO roles in parallel, allowing Vernier non-ambiguity-range extension (Camenzind et al., 2024).

Their real-time processing chain is GPU-accelerated. In each $f_{\rm rep,1}=f_{\rm rep}$05 window, interferograms sampled at $f_{\rm rep,1}=f_{\rm rep}$06 are transferred to the GPU, frequency-shifted toward DC, converted to complex envelopes via Hilbert transform, searched for the two references and target pulse, and analyzed for center-of-mass delays and phase delays. Overlap cases are detected and corrected by subtracting the clean reference, and absolute time-of-flight and interferometric phase are unwrapped and tracked in real time. A sustained processing rate up to $f_{\rm rep,1}=f_{\rm rep}$07 was reported, with the Hilbert transform taking about $f_{\rm rep,1}=f_{\rm rep}$08 on an RTX A5000 (Camenzind et al., 2024).

In industrial in-situ 3D profiling, the dual-comb beam is combined coaxially with a $f_{\rm rep,1}=f_{\rm rep}$09 picosecond micromachining beam by a polarizing beam splitter, scanned through a 2D galvanometric system, and focused onto the workpiece with an $f_{\rm rep,1}=f_{\rm rep}$10 f-theta lens to a $f_{\rm rep,1}=f_{\rm rep}$11 spot (Soghomonyan et al., 13 Apr 2025). The back-scattered comb light is collected through the same objective and detected by a $f_{\rm rep,1}=f_{\rm rep}$12 photodiode, digitized at $f_{\rm rep,1}=f_{\rm rep}$13, then processed by FFT or Hilbert-transform methods (Soghomonyan et al., 13 Apr 2025). A typical scan uses a $f_{\rm rep,1}=f_{\rm rep}$14 grid with $f_{\rm rep,1}=f_{\rm rep}$15 lateral spacing and five measurements per point, corresponding to $f_{\rm rep,1}=f_{\rm rep}$16 lateral positions and a total scan time of about $f_{\rm rep,1}=f_{\rm rep}$17 (Soghomonyan et al., 13 Apr 2025).

Dual-comb FMCW departs from static multi-heterodyne ranging by sweeping both combs synchronously. For comb mode index $f_{\rm rep,1}=f_{\rm rep}$18, the up- and down-ramp beat frequencies are

$f_{\rm rep,1}=f_{\rm rep}$19

from which both range and radial velocity are obtained per comb line (Lukashchuk et al., 2021). The signal chain includes an optical hybrid, balanced photodiodes producing I and Q outputs, and digitization at $f_{\rm rep,1}=f_{\rm rep}$20 with $f_{\rm rep,1}=f_{\rm rep}$21 bandwidth, followed by STFT and Gaussian peak fitting (Lukashchuk et al., 2021). This architecture supports parallel ranging and velocimetry across up to $f_{\rm rep,1}=f_{\rm rep}$22 spectrally dispersed optical channels and megapixel-line rates (Lukashchuk et al., 2021).

Two-photon systems use markedly simpler electronics. Probe and LO pulses overlap in a two-photon absorption detector, producing a cross-correlation pulse train at $f_{\rm rep,1}=f_{\rm rep}$23 rather than GHz fringe data (Wright et al., 2021, Forman et al., 8 Mar 2026). The front-end can consist of a TPA photodiode, a $f_{\rm rep,1}=f_{\rm rep}$24 current amplifier, a constant-fraction discriminator, and digital time stamping with TDCs or a microcontroller (Forman et al., 8 Mar 2026). One implementation used two TI-TDC7200 time-to-digital converters with $f_{\rm rep,1}=f_{\rm rep}$25 resolution and a Teensy 4.0 microcontroller streaming $f_{\rm rep,1}=f_{\rm rep}$26 bits per sample over USB at up to $f_{\rm rep,1}=f_{\rm rep}$27, for a data burden below $f_{\rm rep,1}=f_{\rm rep}$28 (Forman et al., 8 Mar 2026). This is the principal reason two-photon dual-comb ranging can operate continuously without near-gigasample/s acquisition.

5. Demonstrated application domains

The most direct application is absolute distance metrology and target tracking. The free-space dead-zone-free architecture tracked a cooperative target moved over $f_{\rm rep,1}=f_{\rm rep}$29 and compared dual-comb results with a He–Ne reference interferometer. The residuals were below $f_{\rm rep,1}=f_{\rm rep}$30 over the tested range, while phase-enabled operation provided sub-$f_{\rm rep,1}=f_{\rm rep}$31 single-shot precision (Camenzind et al., 2024). The same work reports real-time tracking of moving targets, with sufficiently stable free-running operation to use interferometric phase without $f_{\rm rep,1}=f_{\rm rep}$32 stabilization (Camenzind et al., 2024).

High-speed compact ranging is a major theme in microresonator systems. Integrated DKS dual-comb LIDAR demonstrated acquisition rates up to $f_{\rm rep,1}=f_{\rm rep}$33, Allan deviations down to $f_{\rm rep,1}=f_{\rm rep}$34 at $f_{\rm rep,1}=f_{\rm rep}$35, and moving-target measurements on air-gun projectiles flying at $f_{\rm rep,1}=f_{\rm rep}$36 and on a rotating disk at edge speed $f_{\rm rep,1}=f_{\rm rep}$37 (Trocha et al., 2017). The reported lateral sampling on the rotating disk was about $f_{\rm rep,1}=f_{\rm rep}$38, and the projectile profile agreed with swept-source OCT of the recovered bullet (Trocha et al., 2017). This established dual-comb LIDAR as compatible with ultrafast motion tracking rather than only static metrology.

Parallel coherent ranging and velocimetry have been demonstrated with swept dual-soliton microcombs. The reported system delivered line-scan pixel rates of $f_{\rm rep,1}=f_{\rm rep}$39 with $f_{\rm rep,1}=f_{\rm rep}$40 channels in a $f_{\rm rep,1}=f_{\rm rep}$41 comb and $f_{\rm rep,1}=f_{\rm rep}$42 with roughly $f_{\rm rep,1}=f_{\rm rep}$43 channels in a $f_{\rm rep,1}=f_{\rm rep}$44 comb (Lukashchuk et al., 2021). Range resolution was $f_{\rm rep,1}=f_{\rm rep}$45–$f_{\rm rep,1}=f_{\rm rep}$46 in the $f_{\rm rep,1}=f_{\rm rep}$47 case and $f_{\rm rep,1}=f_{\rm rep}$48–$f_{\rm rep,1}=f_{\rm rep}$49 in the $f_{\rm rep,1}=f_{\rm rep}$50 case, with a common unambiguous range of $f_{\rm rep,1}=f_{\rm rep}$51 and demonstrated velocity imaging of a $f_{\rm rep,1}=f_{\rm rep}$52 flywheel at $f_{\rm rep,1}=f_{\rm rep}$53 (Lukashchuk et al., 2021). This positions dual-comb methods within coherent imaging rather than only point ranging.

A different application domain is in-situ advanced manufacturing. Coaxial dual-comb LiDAR integrated into a laser micromachining station enabled 3D profiling with sub-micron axial precision without moving the workpiece (Soghomonyan et al., 13 Apr 2025). Measured single-point standard deviation on rough, non-cooperative surfaces was $f_{\rm rep,1}=f_{\rm rep}$54 with no averaging at $f_{\rm rep,1}=f_{\rm rep}$55 and $f_{\rm rep,1}=f_{\rm rep}$56 acquisition, improving to about $f_{\rm rep,1}=f_{\rm rep}$57 with $f_{\rm rep,1}=f_{\rm rep}$58 averages and about $f_{\rm rep,1}=f_{\rm rep}$59 with $f_{\rm rep,1}=f_{\rm rep}$60 averages (Soghomonyan et al., 13 Apr 2025). The authors explicitly frame this as in-situ nondestructive testing and process evaluation during micromachining (Soghomonyan et al., 13 Apr 2025).

Atmospheric and remote-sensing uses appear in dual-comb DIAL. An electro-optic multi-heterodyne differential absorption lidar at $f_{\rm rep,1}=f_{\rm rep}$61 measured atmospheric CO$f_{\rm rep,1}=f_{\rm rep}$62 over a $f_{\rm rep,1}=f_{\rm rep}$63 optical path and simultaneously extracted radial wind speed from aerosol backscatter (Rosas et al., 2024). The path-average CO$f_{\rm rep,1}=f_{\rm rep}$64 retrieval had a precision of about $f_{\rm rep,1}=f_{\rm rep}$65, while wind-speed measurements showed standard deviation typically below $f_{\rm rep,1}=f_{\rm rep}$66 in most range gates, with pulse-limited range resolution $f_{\rm rep,1}=f_{\rm rep}$67 and PRF-limited unambiguous range $f_{\rm rep,1}=f_{\rm rep}$68 (Rosas et al., 2024). This broadens the term “dual-comb LIDAR” beyond geometric ranging to multi-frequency absorption and Doppler sensing.

Two-photon dual-comb imaging extends the method to rough or discontinuous surfaces where coherent phase can be compromised by speckle. At a $f_{\rm rep,1}=f_{\rm rep}$69 stand-off, imaging of an aluminum test object produced point-cloud data with ranging accuracies of $f_{\rm rep,1}=f_{\rm rep}$70 to $f_{\rm rep,1}=f_{\rm rep}$71 and precisions averaging to $f_{\rm rep,1}=f_{\rm rep}$72 after $f_{\rm rep,1}=f_{\rm rep}$73 (Nelmes et al., 13 Mar 2026). Continuous-streaming two-photon metrology with free-running $f_{\rm rep,1}=f_{\rm rep}$74 Er,Yb:glass lasers further demonstrated capture of a four-minute audio track from the displacement of a loudspeaker-mounted mirror, with nearly $f_{\rm rep,1}=f_{\rm rep}$75 precision in $f_{\rm rep,1}=f_{\rm rep}$76 averaging (Forman et al., 8 Mar 2026).

6. Limitations, misconceptions, and current directions

A common misconception is that dual-comb LIDAR is uniformly free of ambiguity. In practice, ambiguity depends on the specific observable. In repetition-rate-offset timing architectures, smaller $f_{\rm rep,1}=f_{\rm rep}$77 extends non-ambiguity range but reduces update rate (Zhu et al., 1 Sep 2025, Camenzind et al., 2024). In repetition-rate-limited phase-slope systems, increasing $f_{\rm rep,1}=f_{\rm rep}$78 improves precision but reduces $f_{\rm rep,1}=f_{\rm rep}$79 (Martin et al., 2022). FMCW and synthetic-wavelength systems distribute ambiguity differently, often into chirp duration or synthetic wavelength (Lukashchuk et al., 2021, Trocha et al., 2017). Thus “absolute” ranging does not imply arbitrarily large ambiguity-free operation without architectural concessions.

Another misconception is that dual-comb systems inherently resolve multiple static targets inside one pixel. An explicit limitation study showed the impossibility to resolve different targets in a particular heterodyne phase-slope architecture when two static reflectors contribute to the same photodiode signal without Doppler separation (Martin et al., 2022). The measured RF-beat phase becomes a nonlinear mixture, yielding either a false weighted distance or loss of linearity severe enough that no distance can be reported (Martin et al., 2022). This is a fundamental caution for cluttered scenes and suggests the need for time gating, Doppler multiplexing, or alternative separation mechanisms.

Mechanical and thermal tuning of $f_{\rm rep,1}=f_{\rm rep}$80 illustrate another active design trade-off. Mechanical delay lines can provide tuning rates around $f_{\rm rep,1}=f_{\rm rep}$81 but suffer open-loop errors of hundreds of hertz and mechanical resonances (Zhu et al., 1 Sep 2025). Thermal bidirectional Lyot filtering reduces the differential-comb-line control uncertainty to $f_{\rm rep,1}=f_{\rm rep}$82 without moving parts, but thermal tuning is naturally slower. This suggests that future systems may combine coarse fast tuning with fine thermal stabilization, although that specific hybrid strategy is not explicitly demonstrated in the cited work.

Environmental robustness remains a major issue for deployment. Outdoor long-range ranging is limited by atmospheric refractive-index fluctuations, motivating multi-point meteorology or dispersion-based correction (Camenzind et al., 2024). In automotive or harsh environments, rain, fog, snow, dust, and water droplets on the window produce spurious echoes and artifacts; proposed mitigations include a synchronized spinning shield, moving-average and gated detection, and sensor fusion with radar and cameras (Cuevas, 4 Aug 2025). In micromachining environments, optical isolation is needed to protect detectors during high-power machining pulses, while speckle on rough metallic surfaces can dominate the height uncertainty, with measured speckle errors around $f_{\rm rep,1}=f_{\rm rep}$83 on machined sections (Soghomonyan et al., 13 Apr 2025).

Current directions are correspondingly diverse. One route emphasizes ever faster, more integrated comb sources, including photonic integrated circuits, nanophotonic phased arrays, and ASIC/FPGA DSP for chip-scale solid-state LIDAR engines (Trocha et al., 2017, Lukashchuk et al., 2021). A second route emphasizes robust free-running single-cavity sources with long operation times and low drift for industrial metrology (Camenzind et al., 2024, Cuevas, 4 Aug 2025, Zhu et al., 1 Sep 2025). A third route uses nonlinear or simplified detection, particularly two-photon cross-correlation, to eliminate high-rate digitization and relax $f_{\rm rep,1}=f_{\rm rep}$84 stabilization requirements (Wright et al., 2021, Forman et al., 8 Mar 2026, Nelmes et al., 13 Mar 2026). A plausible implication is that the field is separating into application-specific regimes: ultrafast integrated ranging, fieldable coherent metrology, and low-data-burden precision sensing.

7. Representative performance landscape

The following examples summarize the range of operating regimes reported for dual-comb LIDAR and related dual-comb ranging systems.

System	Key reported parameters	Representative outcome
Free-running single-cavity solid-state dual-comb	$f_{\rm rep,1}=f_{\rm rep}$85, $f_{\rm rep,1}=f_{\rm rep}$86, $f_{\rm rep,1}=f_{\rm rep}$87	$f_{\rm rep,1}=f_{\rm rep}$88 single-shot ToF precision; $f_{\rm rep,1}=f_{\rm rep}$89 phase precision; residuals below $f_{\rm rep,1}=f_{\rm rep}$90 over $f_{\rm rep,1}=f_{\rm rep}$91 (Camenzind et al., 2024)
Thermally tuned bidirectional Lyot-filter fiber dual-comb	$f_{\rm rep,1}=f_{\rm rep}$92, control accuracy $f_{\rm rep,1}=f_{\rm rep}$93, $f_{\rm rep,1}=f_{\rm rep}$94 from $f_{\rm rep,1}=f_{\rm rep}$95 to $f_{\rm rep,1}=f_{\rm rep}$96	Non-ambiguous distance extended from $f_{\rm rep,1}=f_{\rm rep}$97 to $f_{\rm rep,1}=f_{\rm rep}$98 (Zhu et al., 1 Sep 2025)
Integrated DKS dual-comb LIDAR	$f_{\rm rep,1}=f_{\rm rep}$99, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$00 lines, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$01	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$02 ranging rate; $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$03 Allan deviation at $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$04 (Trocha et al., 2017)
Dual-soliton microcomb FMCW	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$05 to $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$06 channels, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$07, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$08–$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$09	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$10 to $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$11 line rates with parallel range and velocity (Lukashchuk et al., 2021)
In-situ micromachining dual-comb LiDAR	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$12, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$13 bandwidth, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$14	Sub-micron axial precision in coaxial 3D profiling; $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$15 for a $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$16 scan with five measurements per point (Soghomonyan et al., 13 Apr 2025)
Two-photon dual-comb LiDAR imaging	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$17, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$18, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$19	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$20–$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$21 accuracy and $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$22 precision after $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$23 at $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$24 stand-off (Nelmes et al., 13 Mar 2026)
Continuous-streaming two-photon dual-comb	$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$25, $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$26–$f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$27	Sustained $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$28 streaming, nearly $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$29 precision in $f_{\rm rep,2}=f_{\rm rep}+\Delta f_{\rm rep}$30 (Forman et al., 8 Mar 2026)

Taken together, these results define dual-comb LIDAR as a family of ranging and remote-sensing techniques rather than a single instrument design. The common core is the controlled mapping of optical delay into a low-frequency observable using two combs with slightly different repetition rates. The main differentiators are the method used to preserve mutual coherence, the way ambiguity is managed, whether detection is coherent or nonlinear, and the application-specific balance among precision, update rate, data burden, and environmental robustness (Camenzind et al., 2024, Zhu et al., 1 Sep 2025, Trocha et al., 2017, Wright et al., 2021, Martin et al., 2022).