Breadboarding the European Moon Rover System: discussion and results of the analogue field test campaign

Alba Guerra; Almudena Moreno; Angus Cameron; Armin Wedler; Augusto G\'omez Egu\'iluz; Bernhard rebele; Carmen Cama\~nes; Cristina Luna; Jennifer Reynolds; Jorge Barrientos-D\'iez

REVIEW 3 major objections 4 minor 14 references

Breadboarding the European Moon Rover System: discussion and results of the analogue field test campaign

T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The European Moon Rover System breadboard validates a modular mobility design by passing all required locomotion modes on lunar regolith simulant, with energy and wheel-deflection data to back it.

desk verdict Useful first quantitative mobility and energy data for ESA's EMRS breadboard, but the wheel-deflection conclusion contradicts the paper's own caveat and needs an explicit fix. read the letter →

arxiv 2411.13978 v1 pith:ITG4NMI7 submitted 2024-11-21 cs.RO astro-ph.EPastro-ph.IM

Cristina Luna , Augusto Gómez Eguíluz , Jorge Barrientos-Díez , Almudena Moreno , Alba Guerra , Manuel Esquer , Marina L. Seoane , Steven Kay

show 8 more authors

Angus Cameron Carmen Camañes Philipp Haas Vassilios Papantoniou Armin Wedler Bernhard Rebele Jennifer Reynolds Markus Landgraf

This is my paper · ORCID

classification cs.ROastro-ph.EPastro-ph.IM

keywords lunarrovermodularmobilitybreadboardlocomotionmodescostoftransportwheeldeflectionregolithsimulantfieldtestcampaign

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper reports a field test campaign of a 1:2-scale breadboard of the European Moon Rover System, a modular lunar rover whose four independently steered wheels let one chassis serve different missions. The authors argue that the tests validate the modular mobility concept: the rover executed Ackermann, skid, crab, and point-turn locomotion, with point turns becoming more energy-efficient than skid steering as rotation angle grows. Measured nominal wheel deflection during obstacle traversal was 3.5 to 5 percent of wheel volume, and cost-of-transport values on slopes up to 25 degrees and in excavation mode indicate the rover can handle diverse payloads and tasks. If the results scale to the flight design, a single versatile rover platform could cover polar prospecting, astrophysical observatory deployment, in-situ resource utilization, and geological survey missions.

What carries the argument

The central object is the EMRS breadboard: a 1:2-scale, 84 kg prototype with four independently steerable wheel modules, aluminium suspension arms, and embedded drive and steering motors that measure current draw. It embodies the modular mobility concept, allowing locomotion mode (Ackermann, skid, crab, point turn) and payload configuration to be swapped for different missions. The quantitative machinery is the cost of transport, $\epsilon = P/(m g v)$, which collapses energy consumption into a nondimensional number for comparing terrains and modes; wheel deflection is estimated by fitting a 3D wheel model to calibrated camera images and computing the volume of the deflected segment. These measurements — motion-capture trajectories, odometry-efficiency ratios, deflection curves, and cost-of-transport tables — carry the argument that the modular design meets its requirements.

What would settle it

Run the flight-scale rover (or a dynamically scaled version with matched wheel stiffness) across the same obstacle and slope profiles in a low-gravity or lunar-gravity analogue, and measure wheel deflection and the skid-versus-point-turn energy crossover. If the nominal deflection leaves the 3.5 to 5 percent band, or the crossover angle moves by more than measurement error, the campaign's validation of the modular mobility concept for lunar conditions would be falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the successful development and testing of the EMRS breadboard validates the modular mobility concept for a multipurpose lunar rover. Using commercial off-the-shelf and space components, the breadboard met its design requirements and demonstrated locomotion in four modes — Ackermann, skid, crab, and point turn — on a slope-configurable sandbox filled with lunar regolith simulant. The data show that point-turn steering becomes more energy-efficient than skid steering as the yaw angle grows, that wheel deflection from obstacle traversal stabilises around 3.5 to 5 percent of wheel volume, and that the rover's cost of transport remains acceptable across flat, sloped (up to 25 degrees), and excavation scenarios. The authors conclude that the modular design lets the choice of locomotion mode be tailored to mission needs, supporting the EMRS concept for polar, astrophysical, in-situ resource utilisation, and geological missions.

Load-bearing premise

The whole validation rests on assuming that a 1:2-scale test rover running in a sandbox of lunar regolith simulant at Earth gravity, with two wheels that could not be set to identical stiffness, behaves like the full flight rover on the Moon.

Editorial extensions

If this is right

A single modular rover platform could be reconfigured for four mission types — polar prospecting, lunar observatory deployment, in-situ resource utilisation, and geological survey — without redesigning the chassis.
Point-turn steering should be chosen for large rotations; skid steering loses about a quarter of wheel motion to slip on low-compaction regolith and is best reserved for small manoeuvres or fault tolerance.
The 3.5–5% wheel deflection range gives suspension designers a quantitative target for stiffness tuning and a starting estimate of sinkage and traction losses.
Cost-of-transport values between about 0.55 and 1.4 across flat, sloped, and excavation tasks indicate the rover's motors are sized for loaded or difficult terrain, not for empty flat running.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The reported ~25% odometry loss for skid steering implies that any rover relying on wheel odometry in low-compaction regolith will need external localisation; the paper does not address navigation, but the slip data make that need concrete.
The 3.5–5% deflection band was measured on two wheels with non-identical stiffness; a systematic stiffness sweep, or a match to the analytical sinkage model the paper cites, would turn the band into a predictive curve for flight design.
If the point-turn/skid energy crossover holds at flight scale and under lunar gravity, then small mission rovers that rotate frequently could justify individual wheel steering purely on energy grounds, even with its added mass and failure modes.
The cost-of-transport table suggests an efficiency sweet spot for loaded or sloped operation; an explicit search over velocity, payload, and slope could find the operating point that minimises $\epsilon$ for each mission class.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

This paper is a straightforward conference-style test report from the EMRS Pre-Phase A campaign. What's actually new is the data: cost-of-transport values on a regolith simulant across slopes, the energy comparison between skid and point-turn steering, and wheel deflection curves from obstacle traversal at 0.6 m/s. These are the first quantitative numbers for this breadboard, and they're useful for anyone sizing rover power systems or deciding whether individual steering is worth the mass and complexity. The steering comparison is the strongest part: the crossover where point-turn becomes cheaper than skid at larger yaw angles, and the 25% odometry loss in skid, are practical results.

The paper does not ship data, code, or repeated trials. That's expected for a short conference paper, but it means the numbers should be treated as preliminary, not design-level certainties. The larger issue is the wheel deflection claim. Section V-C explicitly says there is not enough evidence for a nominal deflection across all wheels because identical stiffness settings were not feasible, then the next sentence asserts the nominal range is 3.5–5%. Section VI repeats that range as if established, and the conclusions lean on it. That's a direct internal contradiction. It could be resolved by reframing the range as a description of the two measured wheels in this configuration, not a rover-level property. As written, it overclaims.

Two smaller things. The overlap with the authors' earlier IAC paper [9] is not stated; the reader is left guessing which results are new here. And the CoT table shows nominal flat-ground CoT increasing with speed (0.646 at 3 cm/s vs 1.10 at 6 cm/s), which looks counterintuitive at first glance—likely a motor-efficiency effect, but the paper doesn't explain it.

The central engineering utility holds up. The qualitative conclusions—modularity is viable, point-turn beats skid beyond a yaw threshold, CoT is reasonable—are consistent with the data. The wheel-deflection overclaim is fixable in revision and doesn't sink the rest.

Who is this for? Planetary rover engineers, especially anyone working on EMRS follow-on or ESA lunar logistics. It deserves a serious referee, but only with the deflection contradiction addressed. I'd send it to review without hesitation.

Referee Report

3 major / 4 minor

Summary. This paper reports results from an analogue field test campaign of the European Moon Rover System (EMRS) breadboard, a 1:2-scale modular rover with four individually steerable wheels, tested on a slope-configurable sandbox with lunar regolith simulant at DLR. The manuscript describes trajectory and orientation data for Ackermann, skid, crab, and point turn locomotion modes (Section V-A), compares the energy consumption of skid steering versus point turn as a function of yaw angle (Section V-B), analyzes wheel deflection during a single obstacle traversal using camera images and a fitted 3D wheel model (Section V-C), and computes cost of transport at several slopes and speeds using the Gabrielli–von Kármán formula (Section V-D). The paper concludes that the breadboard validates the modular mobility concept, that the rover surpassed requirements for the tested locomotion modes, and that nominal wheel deflection is about 3.5–5%. The manuscript is an empirical test report rather than a theoretical derivation, and its main value lies in the presentation of direct measurements from a representative lunar-regolith environment.

Significance. If the reported results are reliable, the paper provides useful experimental evidence for ESA's EMRS pre-phase A study, particularly the comparison of skid steering and point turn energy usage and the demonstration that a modular, four-wheel-steered rover can execute multiple locomotion modes on lunar regolith simulant. The cost-of-transport values, despite the absence of uncertainty quantification, are plausible and grounded in an established metric. The wheel deflection analysis is less convincing because the paper itself acknowledges that stiffness settings could not be matched between wheels, yet it still generalizes to a nominal 3.5–5% rover-level value. Overall, the work is valuable as an engineering test report, but its central validation claims need to be scoped more carefully with respect to the limited number of trials and acknowledged experimental limitations.

major comments (3)

[Section V-C] The wheel deflection conclusion is internally inconsistent. The paragraph states that 'there is not enough evidence to state a nominal wheel deflection value for all wheels' because 'ensuring identical stiffness settings for both wheels was not feasible,' but the next sentence asserts that 'experimental data have shown the nominal wheel deflection for the EMRS rover is within the range of ~3.5 and ~5%.' Since the two measured wheels had different, uncalibrated stiffness settings and only one obstacle traversal was analyzed, the 3.5–5% range cannot be presented as a rover-level property. This claim is repeated in Section VI and used to support the validation narrative. The authors should either remove the generalized nominal value, or report it strictly as the observed range for wheels A and B under the specific test conditions, with an explicit statement that it is not a rover-level specification.
[Sections V-B, V-C, V-D] The quantitative results lack uncertainty quantification and repeated trials. Figure 3 reports energy consumption per degree of yaw for skid steering versus point turn, Figure 4 reports angular speed efficiency as a single time series, and Table II lists cost-of-transport values for individual tests, yet no error bars, confidence intervals, or number of repetitions are given. As a result, the crossover point in Fig. 3 and the differences between CoT values at different slopes cannot be statistically assessed. Because the paper's central claim—that the breadboard 'validates' the modular concept—depends on these quantitative comparisons, the authors should add information about the number of runs, measurement uncertainty, and variability, or explicitly weaken the claims to qualitative observations.
[Section V-C] The wheel deflection estimation method relies on an unvalidated assumption and a single scenario. The text states that the deflected part of the outboard perimeter is estimated 'by assuming equal deflection on both sides of the wheel' and then fitting a convex hull to the projected perimeters. No validation of this geometric assumption is provided, and the deflection is computed from only two wheels during one obstacle traversal at 0.6 m/s with no payload and with one wheel momentarily airborne. This means the observed peak-impact deflection and the nominal range cannot be generalized to other speeds, payloads, or wheel stiffness configurations. The authors should present this as a case study and specify the conditions under which the deflection data are valid.

minor comments (4)

[Section V-D] The sentence describing the speed measurement is ambiguous: 'The speed used is the one obtained directly from the motor encoders, after analysing the slippage with the delta between this speed and the speed calculated by the motion capture system.' It is unclear whether the encoder speed was corrected for slip or only compared with motion-capture speed; please clarify the exact procedure used to obtain the speeds in Table II.
[Section VI] There is a typo: 'The test were conducted' should be 'The tests were conducted'.
[Section V-C] Minor wording issue: 'wheel-A is still flat on the obstacle' is confusing; it likely means the wheel is resting on top of the obstacle. Consider rephrasing to 'wheel-A is still in contact with the obstacle' or similar.
[References] Reference [2] cites a Zenodo preprint without a clear author or publication venue; for a state-of-the-art section, a more substantive reference on Chandrayaan-3 would be preferable.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is an empirical test report whose quantitative results (trajectories, power, wheel deflection, cost of transport) are measured directly and compared against an external formula; the main self-citations are design-context references, not inputs that predetermine the test outcomes.

full rationale

The paper does not present a derivation whose conclusion is equivalent to its inputs. The cost-of-transport results use the standard Gabrielli–von Kármán formula (Eq. 1) with directly measured power, mass, gravity, and speed, so the table entries are observational rather than fitted predictions. The wheel-deflection percentages are obtained from image-based pose fitting and manual annotation of two wheels, and although the paper overgeneralizes from non-identical stiffness settings (Section V-C), that is an empirical-support problem, not a circularity: the 3.5–5% range is not defined in terms of the validation claim nor fitted from the claim. Locomotion-mode validation rests on motion-capture ground truth and odometry, which are independent measurements. Self-citations [7]–[9] are used to describe the modular design concept, mission contexts, and the scaling approach; they do not by themselves supply the experimental results or the cost-of-transport numbers. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. The internal inconsistency between the Section V-C caveat and the repetition of the 3.5–5% nominal deflection in Sections V-C and VI is best treated as a correctness or robustness concern rather than a circular derivation. Therefore the appropriate circularity finding is no significant circularity, score 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claims rest on measurements and standard engineering formulas, not on derived laws. The important unstated premises are the representativeness of the analogue terrain, the simplified wheel-deflection model, and the generalization from two wheels to the whole rover.

assumptions (4)

domain assumption The formula for Cost of Transport, epsilon = P/(mgv), is a valid and sufficient metric for comparing locomotion energy efficiency across the tested conditions.
Invoked in Section V-D as the evaluation metric; it is a standard definition in biomechanics and vehicle engineering, cited to Gabrielli and von Karman.
domain assumption The DLR Planetary Exploration Lab sandbox with lunar regolith simulant is representative enough of lunar surface terrain for the qualitative conclusions about locomotion modes and wheel deflection.
The entire test campaign is conducted in this analogue environment; the paper does not validate the simulant's mechanical properties against actual lunar regolith.
ad hoc to paper Wheel deflection volume can be reconstructed by assuming equal deflection on both sides of the wheel and fitting a convex hull to the projected perimeters.
Section V-C states the outboard perimeter deflection is estimated 'by assuming equal deflection on both sides of the wheel' and the deflected volume is the convex hull of perimeter points; this is a simplifying modeling assumption specific to this analysis.
ad hoc to paper The measured behavior of wheels A and B generalizes to the EMRS rover's nominal wheel deflection despite acknowledged differences in stiffness settings.
Section V-C admits 'there is not enough evidence to state a nominal wheel deflection value for all wheels', yet the conclusion asserts a 3.5-5% nominal range; the paper relies on this generalization while also flagging its weakness.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Breadboarding the European Moon Rover System: discussion and results of the analogue field test campaign." pith.science (2026). https://pith.science/paper/ITG4NMI7

@misc{pith2026241113978,
  author       = {Pith},
  title        = {Pith review of: Breadboarding the European Moon Rover System: discussion and results of the analogue field test campaign},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ITG4NMI7}},
  note         = {Machine review of arXiv:2411.13978}
}

read the original abstract

This document compiles results obtained from the test campaign of the European Moon Rover System (EMRS) project. The test campaign, conducted at the Planetary Exploration Lab of DLR in Wessling, aimed to understand the scope of the EMRS breadboard design, its strengths, and the benefits of the modular design. The discussion of test results is based on rover traversal analyses, robustness assessments, wheel deflection analyses, and the overall transportation cost of the rover. This not only enables the comparison of locomotion modes on lunar regolith but also facilitates critical decision-making in the design of future lunar missions.

Figures

Figures reproduced from arXiv: 2411.13978 by the authors.

**Figure 2.** Trajectories and orientations of the EMRS Breadboard [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Energy consumption per degree of rover yaw captured [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Angular velocity ratio between the obtained with [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]

**Figure 5.** Figure 5: Wheel deflection estimation for each of the two [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]

**Figure 6.** Figure 6: Deflection of wheels A and B on the left and right, [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[9]

Modularity for lunar exploration: European Moon Rover System Pre-Phase A Design and Field Test Campaign Results

C. Luna et al., “Modularity for lunar exploration: European moon rover system pre-phase a design and field test cam- paign results,” in 74th International Astronautical Congress (IAC), Baku, Azerbaijan, Oct. 2023. DOI: 10 . 48550 / arXiv.2311.03098

work page Pith review arXiv doi:10.48550/arxiv.2311.03098 2023
[1]

Skid mobility system to travel on plan- etary surface for a small lander,

H. EGUCHI et al., “Skid mobility system to travel on plan- etary surface for a small lander,” AEROSPACE TECHNOL- OGY JAPAN, THE JAPAN SOCIETY FOR AERONAUTICAL AND SPACE SCIENCES, vol. 18, no. 0, pp. 189–198, 2019. DOI: 10.2322/astj.jsass-d-18-00035

work page doi:10.2322/astj.jsass-d-18-00035 2019
[2]

India’s chandrayaan missions: Exploring the moon for science and resources,

D. George, “India’s chandrayaan missions: Exploring the moon for science and resources,” 2023. DOI: 10.5281/ ZENODO.8285558

work page 2023
[3]

Luna-25: The first polar mission to the moon,

I. G. Mitrofanov et al., “Luna-25: The first polar mission to the moon,” Solar System Research, vol. 55, no. 6, pp. 485– 495, Nov. 2021. DOI: 10.1134/s0038094621060095

work page doi:10.1134/s0038094621060095 2021
[4]

Moonshot: Affordable, simple, flight hardware for the artemis-1 mission and beyond,

T. G. Hammond et al., “Moonshot: Affordable, simple, flight hardware for the artemis-1 mission and beyond,” Frontiers in Space Technologies, vol. 1, Nov. 2020. DOI: 10.3389/ frspt.2020.593523

work page arXiv 2020
[5]

Scientific objectives and payload configu- ration of the chang’e-7 mission,

W. Chi et al. , “Scientific objectives and payload configu- ration of the chang’e-7 mission,” National Science Review , Dec. 2023. DOI: 10.1093/nsr/nwad329

work page doi:10.1093/nsr/nwad329 2023
[6]

Ispace modular rover design,

J. Walker et al., “Ispace modular rover design,” Oct. 2019

work page 2019
[7]

The European Moon Rover System: a modular multipurpose rover for future complex lunar missions

C. Luna et al., “The european moon rover system: A mod- ular multipurpose rover for future complex lunar missions,” in 17th Symposium on Advanced Space Technologies In Robotics And Automation (ASTRA 2023) , Leiden, Nether- lands, Oct. 2023. DOI: 10.48550/arXiv.2311.03136

work page Pith review arXiv doi:10.48550/arxiv.2311.03136 2023

Show all 14 references

[8]

Designing a modular multipurpose rover for future lunar missions,

C. Luna et al., “Designing a modular multipurpose rover for future lunar missions,” in 11th European Lunar Symposium, Padova, Italy, Jun. 2023

2023
[10]

Chen, Viper: The rover and its onboard toolkit , 2020

R. Chen, Viper: The rover and its onboard toolkit , 2020

2020
[11]

A 2-year locomotive exploration and scien- tific investigation of the lunar farside by the yutu-2 rover,

L. Ding et al., “A 2-year locomotive exploration and scien- tific investigation of the lunar farside by the yutu-2 rover,” Science Robotics, vol. 7, no. 62, Jan. 2022. DOI: 10.1126/ scirobotics.abj6660

2022
[12]

What price speed? specific power re- quired for propulsion of vehicles,

G. Gabrielli et al. , “What price speed? specific power re- quired for propulsion of vehicles,” Mechanical Engineering, vol. 72, no. 10, pp. 775–781, 1950

1950
[13]

What price of speed? a critical revision through constructal optimization of transport modes,

M. Trancossi, “What price of speed? a critical revision through constructal optimization of transport modes,” Inter- national Journal of Energy and Environmental Engineering, vol. 7, no. 4, pp. 425–448, Feb. 2015. DOI: 10 . 1007 / s40095-015-0160-6

2015
[14]

Modeling of flexible metal wheel for pres- surized lunar rover and traction performance prediction,

J. Zhu et al. , “Modeling of flexible metal wheel for pres- surized lunar rover and traction performance prediction,” Journal of Field Robotics , vol. 40, no. 8, pp. 2030–2041, Aug. 2023. DOI: 10.1002/rob.22239

2023 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [9]

Modularity for lunar exploration: European Moon Rover System Pre-Phase A Design and Field Test Campaign Results

C. Luna et al., “Modularity for lunar exploration: European moon rover system pre-phase a design and field test cam- paign results,” in 74th International Astronautical Congress (IAC), Baku, Azerbaijan, Oct. 2023. DOI: 10 . 48550 / arXiv.2311.03098

work page Pith review arXiv doi:10.48550/arxiv.2311.03098 2023

[2] [1]

Skid mobility system to travel on plan- etary surface for a small lander,

H. EGUCHI et al., “Skid mobility system to travel on plan- etary surface for a small lander,” AEROSPACE TECHNOL- OGY JAPAN, THE JAPAN SOCIETY FOR AERONAUTICAL AND SPACE SCIENCES, vol. 18, no. 0, pp. 189–198, 2019. DOI: 10.2322/astj.jsass-d-18-00035

work page doi:10.2322/astj.jsass-d-18-00035 2019

[3] [2]

India’s chandrayaan missions: Exploring the moon for science and resources,

D. George, “India’s chandrayaan missions: Exploring the moon for science and resources,” 2023. DOI: 10.5281/ ZENODO.8285558

work page 2023

[4] [3]

Luna-25: The first polar mission to the moon,

I. G. Mitrofanov et al., “Luna-25: The first polar mission to the moon,” Solar System Research, vol. 55, no. 6, pp. 485– 495, Nov. 2021. DOI: 10.1134/s0038094621060095

work page doi:10.1134/s0038094621060095 2021

[5] [4]

Moonshot: Affordable, simple, flight hardware for the artemis-1 mission and beyond,

T. G. Hammond et al., “Moonshot: Affordable, simple, flight hardware for the artemis-1 mission and beyond,” Frontiers in Space Technologies, vol. 1, Nov. 2020. DOI: 10.3389/ frspt.2020.593523

work page arXiv 2020

[6] [5]

Scientific objectives and payload configu- ration of the chang’e-7 mission,

W. Chi et al. , “Scientific objectives and payload configu- ration of the chang’e-7 mission,” National Science Review , Dec. 2023. DOI: 10.1093/nsr/nwad329

work page doi:10.1093/nsr/nwad329 2023

[7] [6]

Ispace modular rover design,

J. Walker et al., “Ispace modular rover design,” Oct. 2019

work page 2019

[8] [7]

The European Moon Rover System: a modular multipurpose rover for future complex lunar missions

C. Luna et al., “The european moon rover system: A mod- ular multipurpose rover for future complex lunar missions,” in 17th Symposium on Advanced Space Technologies In Robotics And Automation (ASTRA 2023) , Leiden, Nether- lands, Oct. 2023. DOI: 10.48550/arXiv.2311.03136

work page Pith review arXiv doi:10.48550/arxiv.2311.03136 2023