ASAP-MO:Advanced Situational Awareness and Perception for Mission-critical Operations

Edith Brotherton; Effie Daum; Fran\c{c}ois Pomerleau; Jean-Michel Fortin; Nicolas Samson; Olivier Gamache; Veronica Vannini; William Dubois

REVIEW 3 major objections 5 minor 10 references

ASAP-MO:Advanced Situational Awareness and Perception for Mission-critical Operations

T0 review · 3 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A single operator teleoperated two different ground robots from one base station and merged their maps into a single global map for a mock nuclear inspection.

desk verdict A solid, honest field report on a two-robot inspection mission; the integration is real, but the 'coordinated' claim overstates a mostly sequential workflow with a manual relocalization step. read the letter →

arxiv 2505.01547 v2 pith:DYDMEVWC submitted 2025-05-02 cs.RO

Veronica Vannini , William Dubois , Olivier Gamache , Jean-Michel Fortin , Nicolas Samson , Effie Daum , François Pomerleau , Edith Brotherton This is my paper · ORCID

classification cs.RO

keywords multi-robotteleoperationheterogeneousrobotfleetunifiedmappingindoor-outdoornavigationnuclearinspectionsimulationradiationdetectionfieldroboticsSLAM

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 field report tries to establish that recent robotics tools are mature enough for one operator or a small team, sitting at a single base station, to run a multi-stage indoor-outdoor inspection with a heterogeneous fleet: a large ground robot maps the building exterior, then a smaller tracked robot with a camera and arm enters to search for simulated radiation sources. The paper's evidence is a set of winter field trials plus a final demonstration in which two ground robots teleoperated from one station produced a single consistent map. A sympathetic reading is that the feasibility claim rests on the integration, not on any single new algorithm: shared hierarchical SLAM, a follow-the-gap navigation assist, and a camera-to-map light-intensity projection that plays the role of a directional Geiger counter. If true, it matters because it suggests practical missions can be run without a tether and without an operator physically near the hazard.

What carries the argument

Central object is the shared map and the relocalization handoff. The outdoor map is transferred wirelessly to the second robot; because that robot's starting location differs, the operator enters an approximate pose in the graphical interface, and iterative closest point matching locks the second robot into the first map, so subsequent indoor scans extend the same global reference frame. The perception layer for radiation uses a fixed-exposure grayscale camera: the average grayscale value is computed, a threshold (44 percent, or 112 in an 8-bit image) marks an area as radioactive, and the readings are projected onto lidar points within 2, 3, or 4 meters to color the map.

What would settle it

Run the same two-robot mission with the second robot's starting pose withheld and see whether the indoor map still aligns with the outdoor map automatically; a large misalignment or a failed relocalization would falsify the claimed feasibility of coordinated multi-robot mapping. Alternatively, have both robots move simultaneously while only one map is displayed and check whether the operator can maintain a consistent global reference.

Watch

Extended reading notes

Core claim

The central discovery is that a unified global map can be maintained across two heterogeneous ground robots teleoperated from one remote base station, in a scenario simulating a nuclear decommissioning inspection. The large outdoor robot first maps the outdoor area and locates an entrance; its map is transferred wirelessly to the smaller tracked robot, which relocalizes itself using an operator-entered approximate pose and then maps the interior while its arm-mounted camera detects bright lights as proxies for radioactive material. The paper claims this demonstrates feasibility of coordinated multi-robot inspection, acknowledging several operational constraints: 915 MHz mesh radio links limited bandwidth, only teleoperation commands went over the data distribution service while the operator watched the screens through a remote desktop, and the camera-to-lidar projection fails when the camera and lidar fields of view do not overlap.

Load-bearing premise

The unified-map coordination claim depends on the operator manually entering an approximate starting position for the second robot on the first robot's map; if that manual handoff is not an acceptable step in real mission-critical operations, the demonstration does not show coordinated multi-robot operation.

Editorial extensions

If this is right

A single base station with low-frequency mesh radios can sustain teleoperation commands and map sharing across robots at roughly 100 meters range, including through walls, without a fiber tether.
Camera-derived sensor values can be attached to map points in real time, giving operators a color-coded hazard layer during teleoperation.
Indoor-outdoor mapping with a common global reference can be achieved sequentially when an operator supplies an approximate starting pose for each subsequent robot.
Switching control among heterogeneous robots from one station works, but the operator must verify all configurations at each handoff, and low-bandwidth links force protocol choices such as commands over a lightweight service and screen sharing for video.
Operator feedback from defense stakeholders suggests that untethered teleoperation addresses a real operational limitation, and that adding lidar and inertial feedback improves awareness of slopes and soft obstacles.

Reading between the lines

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

The feasibility claim is sequential, not simultaneous: the field report does not test two robots moving at once or automatic map merging, so a fair extension would be to run both robots concurrently and measure whether the single-map result persists.
The 44 percent threshold and the 2, 3, and 4 meter distance bands are calibrated to this particular camera and light setup; a testable extension is to calibrate against an actual Geiger counter and see whether the projected distribution matches real radiation readings.
The unified map depends on a manual approximate relocalization step; automating that step with global localization or loop closure would be a natural next test and would reduce operator workload.
The varied winter, forest, urban, and indoor conditions hint at environmental robustness, but the reported data are too sparse to quantify success rates; a future experiment could report per-condition success rates over repeated runs.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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

Quick read: this is a field report, and a solid one. The team integrated a Warthog and a smaller HD2 with a Kinova arm, ran them from a single base station over a 915 MHz mesh, and produced a shared map used to locate simulated radiation sources. The operation lasted over 5.6 hours across several sites, including winter conditions. That is real field evidence, not a simulation.

The most valuable part is the lessons learned. The lidar-camera field-of-view mismatch in Figure 7 is a concrete, reproducible observation that will save other teams time. The comparison with the stakeholders' tethered system is also genuinely informative; they showed a wireless solution at comparable range, which matters for real operations. The radiation proxy is simple and honestly labeled: the 44% threshold is stated as particular to the camera-light pair, and the distance bins are hand-set. No circular reasoning here, and the self-citations are to established methods used as tools.

The soft spots are about wording and evidence, not about the core integration. The word 'coordinated' in the abstract and conclusion overstates what is shown. The mission is sequential: Warthog maps outside, then the HD2 is driven inside after the operator manually inputs an approximate pose in the GUI to relocalize it on the Warthog map. That is a legitimate operational step in many field systems, and the paper describes it honestly in Section II. But the abstract presents the unified map as a demonstration of coordinated multi-robot missions without noting that the coordination was mostly sequential and depended on an operator seed. There is no metric for relocalization accuracy or map alignment error, so we can't tell how robust that step was. Adding one caveat sentence and one alignment metric would materially strengthen the claim. Missing code/data is a minor issue for a field report.

The stress-test note worries that the manual handoff invalidates the feasibility claim. I do not agree. A single operator running two heterogeneous platforms with a shared map and a documented, repeatable relocalization procedure is a meaningful feasibility result. But the note is right that the current wording overreaches; the feasibility claim is plausible, not yet quantified.

This is for people working on field robotics, inspection, or disaster response. It deserves a serious peer review rather than a desk reject; a referee could reasonably ask for language changes and one metric. I would bring it to our reading group as an anchor for a discussion on what counts as multi-robot coordination.

Referee Report

3 major / 5 minor

Summary. The paper is a field report describing a simulated nuclear-inspection mission in which two heterogeneous UGVs (a large Warthog and a smaller HD2 with a Kinova arm) are teleoperated from a single base station. The Warthog first maps the building exterior; the HD2 is then given an operator-supplied approximate position in that map and teleoperated indoors to perform a simulated radiation-detection task, with lights standing in for radioactive sources. The authors report 5.6 hours and 5.86 km of experiments, describe networking upgrades from 5 GHz to 915 MHz mesh radios, and candidly document failures such as a camera/lidar field-of-view mismatch. The central claim is that the robots produced a unified map output, demonstrating the feasibility of coordinated multi-robot missions for mission-critical operations.

Significance. If the central claim is adequately supported, this would be a useful integration-and-deployment contribution: it shows a path toward a single-operator, multi-robot indoor/outdoor inspection workflow and reports concrete operational lessons (network choices, teleoperation modes, sensor-misalignment effects). The paper's strengths are its real deployment data, the explicit reporting of failures rather than only successes, and the reproduction of established components (ICP, follow-the-gap, Teach-and-Repeat) as tools. However, the feasibility claim currently rests on a manual operator step that is not quantitatively evaluated, and no map-accuracy or detection-accuracy metrics are reported. The significance as a demonstrated feasibility result is therefore not yet fully established.

major comments (3)

[Section II, Step 3 (and Figures 3 and 5)] The only mechanism that places the HD2 map into the Warthog map's coordinate frame is the operator manually inputting an approximate position in the GUI. The paper reports no measurement of the accuracy, repeatability, or sensitivity of this manual relocalization, no alignment residual in the overlap region of the two maps, and no test of whether the procedure works when the operator does not already know the environment. Because the abstract's central claim is that the robots produced a unified map and that this demonstrates feasible coordination, the burden is on the authors to quantify this handoff or to soften the claim to 'sequential teleoperation with operator-supervised map merging.' I recommend adding repeated-trial data on initialization error versus final map alignment, or an explicit statement that the manual step is a known operational constraint rather than a demonstrated coordination capability.
[Section II, Step 6 and Figure 7] The simulated radiation detection uses a hand-adjusted 44% grayscale threshold and projects readings only to lidar points whose field of view overlaps the camera. The paper itself reports that the camera/lidar FOV mismatch causes missing detections when the robot is close to a source (Figure 7). Nevertheless, the conclusion states that the mission to 'investigate the safety of a radioactive area' was successfully demonstrated. There is no ground-truth detection rate, precision/recall, or even a count of sources placed versus sources detected. Without such statistics, the radiation-detection component of the demonstration remains anecdotal. Please add detection metrics or explicitly characterize this part as a qualitative proof-of-concept with known failure modes.
[Section II, Table I and overall methodology] The only quantitative metrics reported are distance traveled, duration, and average physical area covered. There is no metric for map quality, such as loop-closure error, ICP convergence residual, or trajectory error against ground truth. Since the paper's core claim is a 'unified map output,' the absence of any quantitative map-accuracy evaluation leaves the central claim unverified. At minimum, please report the alignment residual between the Warthog and HD2 maps in the overlap region and any map-consistency measures, or state explicitly that no quantitative map evaluation was performed and adjust the claims accordingly.

minor comments (5)

[Section III.A] The sentence 'the raw camera stream and 3D map did not have to pass throught the network' contains a typo: 'throught' should be 'through.'
[Section II, Step 6] The distance-bin description is internally inconsistent: red points are '2 m or closer,' orange points are 'between 2 m and 3 m,' and yellow points are 'at least 4 m,' leaving the 3-4 m interval undefined. Please correct to 'between 3 m and 4 m' or 'at least 4 m' with a matching bin definition.
[Table I] The table columns are difficult to parse, especially the 'Local' and 'Teleoperated' entries (e.g., dashes and 'Y' are not explained). A caption or footnote defining the entries and the meaning of dashes would significantly improve readability.
[Figure 4c] The text states that the average grayscale value is compared with the grayscale threshold, but the figure caption does not explain the axis labels, the threshold line, or how the 44% (112 in 8-bit) value is represented. Please annotate the figure or the caption.
[Section III.B] The sentence about the user needing to 'remain aware and take appropriate actions' is vague; please specify which control modes exist (manual, follow-the-gap, Teach-and-Repeat) and how switching between modes is handled operationally.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is a field report whose parameters are openly hand-adjusted and whose self-citations are tool references rather than load-bearing premises.

full rationale

The paper is a field report with no predictive or derived quantities that reduce to their own inputs. The unified map is produced by an explicit operator step (Section II, Step 3) in which the operator inputs an approximate position for the HD2 in the Warthog map; this is an operational input, not a quantity claimed to be predicted independently, so the map alignment is not presented as a derived result. The 44% radiation threshold and the 2 m/3 m/4 m distance bins are openly described as particular to the light-and-camera setup and are not used to forecast an external result, so they cannot constitute a fitted-input-called-prediction. Self-citations (Pomerleau [5][8], Baril et al. [9], Fortin et al. [10]) are used as background method references for ICP, but the mapping result is not justified by those citations; no uniqueness theorem or load-bearing premise is imported from the authors' prior work. The manual relocalization step weakens the strength of the 'coordinated multi-robot missions' feasibility claim, since the coordination relies on sequential teleoperation with operator assistance, but that is a validity and generalization concern, not a circularity. The derivation chain is therefore self-contained and the central claim, while arguably overstated, is not circular.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central claim rests on standard SLAM, a light-as-radiation proxy, and a manual relocalization step; none of these are new entities. The only hand-tuned constants are the detection threshold and distance bins, which the authors disclose.

free parameters (2)

Grayscale threshold for radiation detection = 44% (112/255)
In Section II, Step 5, the authors state that an area is considered radioactive when the average grayscale intensity exceeds 44 percent, a value 'particular to the combination of lights and camera at hand, but was adjusted to simulate a directional Geiger counter.' This is hand-tuned for the specific camera and lights.
Distance bin boundaries for radiation projection = 2 m, 3 m, 4 m
In Section II, Step 6, color-coded distance bins (red within 2 m, orange 2-3 m, yellow at least 4 m) are chosen to represent radiation strength; these are arbitrary visualization choices.

assumptions (3)

domain assumption ICP is suitable for map merging and localization across the two platforms
The paper uses ICP [8] for all mapping without evaluating its accuracy in this environment; the unified map assumes ICP converges correctly with the manual initial guess.
domain assumption The brightness of a light source is a valid proxy for radiation intensity
The simulation uses lights to represent radiation sources, so any conclusions about radiation detection are limited by this proxy, acknowledged in Section II.
domain assumption The operator can provide an approximate map position for relocalization
Step 3 requires the operator to input the approximate position in the GUI; the success of map merging depends on this manual step being feasible in a mission-critical context.

how reviews work

0 comments

Cite this review

Pith. "Pith review of ASAP-MO:Advanced Situational Awareness and Perception for Mission-critical Operations." pith.science (2026). https://pith.science/paper/DYDMEVWC

@misc{pith2026250501547,
  author       = {Pith},
  title        = {Pith review of: ASAP-MO:Advanced Situational Awareness and Perception for Mission-critical Operations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DYDMEVWC}},
  note         = {Machine review of arXiv:2505.01547}
}

read the original abstract

Deploying robotic missions can be challenging due to the complexity of controlling robots with multiple degrees of freedom, fusing diverse sensory inputs, and managing communication delays and interferences. In nuclear inspection, robots can be crucial in assessing environments where human presence is limited, requiring precise teleoperation and coordination. Teleoperation requires extensive training, as operators must process multiple outputs while ensuring safe interaction with critical assets. These challenges are amplified when operating a fleet of heterogeneous robots across multiple environments, as each robot may have distinct control interfaces, sensory systems, and operational constraints. Efficient coordination in such settings remains an open problem. This paper presents a field report on how we integrated robot fleet capabilities - including mapping, localization, and telecommunication - toward a joint mission. We simulated a nuclear inspection scenario for exposed areas, using lights to represent a radiation source. We deployed two Unmanned Ground Vehicles (UGVs) tasked with mapping indoor and outdoor environments while remotely controlled from a single base station. Despite having distinct operational goals, the robots produced a unified map output, demonstrating the feasibility of coordinated multi-robot missions. Our results highlight key operational challenges and provide insights into improving adaptability and situational awareness in remote robotic deployments.

Figures

Figures reproduced from arXiv: 2505.01547 by the authors.

**Figure 1.** Illustration of a mission. Where (S1) and (E1) represent the start and end route of the first robot, that scouts the outside environment, and (S2) and (E2), show the plan start and end of the second robot, that accomplish the detection task. The full operation is controlled from the base station highlighted in red. This aerial view captures the realworld setting where the final demonstration was executed. could be … view at source ↗

**Figure 2.** Figure a shows the medium platform, HD2, equipped with: (1) Kinova Gen3 robotic arm, (2) the OAK camera and (3) Hesai XT-32 lidar. Figure b shows the Warthog from Clearpath, (4) Lidar RS128 and (5) ZED X stereo camera. Figure c shows the mobile base station, to transport and operate the devices during experiments. (i.e., bright construction lights) as shown in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** The green map and orange trajectory represent [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: The process to convert a camera image to a directional [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 6.** Figure 6: Teleoperation tools installed in the base station. The [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]

**Figure 7.** Figure 7: Example of a misleading environmental reading. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

10 extracted references · 8 canonical work pages

[1]

Decision processes in military command and con- trol,

J. V . Baranski, W. M. Petrusic, and A. M. Bonnel, “Decision processes in military command and con- trol,” in Proceedings of the NATO RTO Human Factors and Medicine Panel (HFM) Symposium , vol. MP- HFM-136, NATO Science and Technology Organiza- tion (STO), 2006, pp. 11–1–11–16

work page 2006
[2]

Lessons learned: Symbiotic autonomous robot ecosystem for nuclear environments,

D. Mitchell, P. Baniqued, B. Lennox, et al. , “Lessons learned: Symbiotic autonomous robot ecosystem for nuclear environments,” English, IET Cyber-Systems Robotics, vol. 5, no. 4, Dec. 2023

work page 2023
[3]

Present and Future of SLAM in Extreme Environments: The DARPA SubT Challenge,

K. Ebadi, L. Bernreiter, H. Biggie, et al. , “Present and Future of SLAM in Extreme Environments: The DARPA SubT Challenge,” IEEE Transactions on Robotics, vol. 40, pp. 936–959, 2024

work page 2024
[4]

Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile rescue robots,

K. Nagatani, S. Kiribayashi, Y . Okada, et al. , “Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile rescue robots,” Journal of Field Robotics , vol. 30, no. 1, pp. 44–63, 2013

work page 2013
[5]

Robotics in snow and ice,

F. Pomerleau, “Robotics in snow and ice,” in En- cyclopedia of Robotics , M. H. Ang, O. Khatib, and B. Siciliano, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 2023, pp. 1–6

work page 2023
[6]

Intuitive wireless control of a robotic arm for people living with an upper body disability,

C. L. Fall, P. Turgeon, A. Campeau-Lecours, et al. , “Intuitive wireless control of a robotic arm for people living with an upper body disability,” in 2015 37th Annual International Conference of the IEEE Engi- neering in Medicine and Biology Society (EMBC) , 2015, pp. 4399–4402

work page 2015
[7]

A novel obstacle avoidance algorithm: “follow the gap method

V . Sezer and M. Gokasan, “A novel obstacle avoidance algorithm: “follow the gap method”,” Robotics and Autonomous Systems , vol. 60, no. 9, pp. 1123–1134, 2012

2012
[8]

Comparing ICP Variants on Real-World Data Sets,

F. Pomerleau, F. Colas, R. Siegwart, and S. Magnenat, “Comparing ICP Variants on Real-World Data Sets,” Autonomous Robots, vol. 34, no. 3, pp. 133–148, Feb. 2013. Presented at the 2025 IEEE ICRA Workshop on Field Robotics

work page 2013

Show all 10 references

[9]

Kilometer-scale autonomous navigation in subarc- tic forests: challenges and lessons learned,

D. Baril, S.-P. Desch ˆenes, O. Gamache, et al. , “Kilometer-scale autonomous navigation in subarc- tic forests: challenges and lessons learned,” Field Robotics, vol. 2, no. 1, pp. 1628–1660, Mar. 2022

2022
[10]

Uav- assisted self-supervised terrain awareness for off-road navigation,

J.-M. Fortin, O. Gamache, W. Fecteau, et al. , “Uav- assisted self-supervised terrain awareness for off-road navigation,” in Proceedings of the IEEE International Conference on Robotics and Automation (ICRA) , 2025. Presented at the 2025 IEEE ICRA Workshop on Field Robotics

2025

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Decision processes in military command and con- trol,

J. V . Baranski, W. M. Petrusic, and A. M. Bonnel, “Decision processes in military command and con- trol,” in Proceedings of the NATO RTO Human Factors and Medicine Panel (HFM) Symposium , vol. MP- HFM-136, NATO Science and Technology Organiza- tion (STO), 2006, pp. 11–1–11–16

work page 2006

[2] [2]

Lessons learned: Symbiotic autonomous robot ecosystem for nuclear environments,

D. Mitchell, P. Baniqued, B. Lennox, et al. , “Lessons learned: Symbiotic autonomous robot ecosystem for nuclear environments,” English, IET Cyber-Systems Robotics, vol. 5, no. 4, Dec. 2023

work page 2023

[3] [3]

Present and Future of SLAM in Extreme Environments: The DARPA SubT Challenge,

K. Ebadi, L. Bernreiter, H. Biggie, et al. , “Present and Future of SLAM in Extreme Environments: The DARPA SubT Challenge,” IEEE Transactions on Robotics, vol. 40, pp. 936–959, 2024

work page 2024

[4] [4]

Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile rescue robots,

K. Nagatani, S. Kiribayashi, Y . Okada, et al. , “Emergency response to the nuclear accident at the Fukushima Daiichi Nuclear Power Plants using mobile rescue robots,” Journal of Field Robotics , vol. 30, no. 1, pp. 44–63, 2013

work page 2013

[5] [5]

Robotics in snow and ice,

F. Pomerleau, “Robotics in snow and ice,” in En- cyclopedia of Robotics , M. H. Ang, O. Khatib, and B. Siciliano, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 2023, pp. 1–6

work page 2023

[6] [6]

Intuitive wireless control of a robotic arm for people living with an upper body disability,

C. L. Fall, P. Turgeon, A. Campeau-Lecours, et al. , “Intuitive wireless control of a robotic arm for people living with an upper body disability,” in 2015 37th Annual International Conference of the IEEE Engi- neering in Medicine and Biology Society (EMBC) , 2015, pp. 4399–4402

work page 2015

[7] [7]

A novel obstacle avoidance algorithm: “follow the gap method

V . Sezer and M. Gokasan, “A novel obstacle avoidance algorithm: “follow the gap method”,” Robotics and Autonomous Systems , vol. 60, no. 9, pp. 1123–1134, 2012

2012

[8] [8]

Comparing ICP Variants on Real-World Data Sets,

F. Pomerleau, F. Colas, R. Siegwart, and S. Magnenat, “Comparing ICP Variants on Real-World Data Sets,” Autonomous Robots, vol. 34, no. 3, pp. 133–148, Feb. 2013. Presented at the 2025 IEEE ICRA Workshop on Field Robotics

work page 2013