pith. sign in

arxiv: 2507.08973 · v4 · submitted 2025-07-11 · 💻 cs.HC

Analytical Study on the Exposedness of Potential Positions for External Human-Machine Interfaces

Jose Gonzalez-Belmonte , Jaerock Kwon This is my paper

Pith reviewed 2026-05-19 04:44 UTC · model grok-4.3

classification 💻 cs.HC
keywords autonomous vehiclesexternal human-machine interfacesvisibility simulationpedestrian crossingvehicle surface exposuresignaling designUnity simulation
0
0 comments X

The pith

Simulation of pedestrian viewpoints shows wheels, front fenders and headlights are the most visible vehicle surfaces.

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

This paper sets out to determine the best locations for external displays on autonomous vehicles by measuring how often different surfaces can be seen by pedestrians on the sidewalk. It uses a Unity simulation to test fifteen vehicles across fifty-seven scenarios that vary vehicle position, camera angle, distance, and number of other cars on the road. The results indicate that wheels, front fenders, and headlights appear in view most frequently. A reader would care because placing signals where people actually look could make interactions with self-driving cars safer and more intuitive. The work also provides a simulation platform that others can reuse for similar visibility studies.

Core claim

The analysis of simulated viewpoints concludes that the areas of the vehicle most often seen by pedestrians on the sidewalk attempting to cross the road were the wheels, front fenders, and headlights. Based on these results, the paper suggests implementing displays on at least two of the following regions: windshield, front fenders, side mirrors. These findings aim to help ensure that pedestrians are able to see signaling on approaching autonomous vehicles.

What carries the argument

Unity-based simulation that records visibility frequency of vehicle surfaces by positioning virtual cameras at multiple sidewalk locations with varying distances and directions.

If this is right

  • Designers should consider placing external human-machine interfaces on high-visibility areas like the front fenders and headlights.
  • Using at least two display regions increases the likelihood that pedestrians notice the signals.
  • The simulation method can be applied to test visibility in additional traffic or environmental conditions.
  • Vehicle manufacturers may use these exposure rankings to standardize interface placements.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Real-world validation of the simulation results could lead to regulatory guidelines for display placement on autonomous vehicles.
  • Extending the model to include cyclist or driver viewpoints might reveal different optimal positions.
  • The emphasis on visible surfaces could influence how other road users like cyclists or drivers are signaled to in mixed traffic.

Load-bearing premise

The simulation's camera placements, distances, and traffic setups accurately reflect the typical viewpoints of real pedestrians observing vehicles from the sidewalk.

What would settle it

Conducting an experiment where real pedestrians wear eye-tracking devices while watching approaching vehicles in actual traffic and comparing the observed gaze points to the simulation's visibility counts.

Figures

Figures reproduced from arXiv: 2507.08973 by Jaerock Kwon, Jose Gonzalez-Belmonte.

Figure 1
Figure 1. Figure 1: Top-Down View of the Virtual Street Layout. [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Screenshot of the simulation running. Red lines indicate the trajectory of the [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Top-Down View of the fifteen vehicles used in the simulations. Color denotes [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Diagram illustrating the ten possible sidewalk camera positions, and the eight [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Diagram illustrating all sixteen setups based on the [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualization of the three ranges used for [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Illustration of the different exterior parts of a sedan, and the terminology used [PITH_FULL_IMAGE:figures/full_fig_p019_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of the right side of the heatmap of all vehicle types that is produced [PITH_FULL_IMAGE:figures/full_fig_p020_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Results showing how many times a Grid Point was recorded, across all scenarios where the vehicle was a SUV driving: (a) North, on the right side of the virtual environ￾ment; (b) East, out of the alleyway on the left side of the virtual environment; (c) South, on the left side of the virtual environment; (d) West, out of the alleyway on the right side of the virtual environment. 19 [PITH_FULL_IMAGE:figures… view at source ↗
Figure 10
Figure 10. Figure 10: Results showing the heatmap of the SUV at each facing direction, when the [PITH_FULL_IMAGE:figures/full_fig_p022_10.png] view at source ↗
read the original abstract

As we move towards a future of autonomous vehicles, questions regarding their method of communication have arisen. One of the common questions concerns the placement of the signaling used to communicate with pedestrians and road users, but few works have been fully dedicated to the matter. This paper uses a simulation made in the Unity game engine to record the fifteen different vehicles under fifty-seven different scenarios each for the first time, in order to find how often its forward-facing exterior surfaces can be seen by a pedestrian on the sidewalk. Variables include the vehicle type, position, number of vehicles on the road, camera position and direction, as well as its minimum and maximum distance from the recorded points. It was concluded that the areas of the vehicle most often seen by pedestrians on the sidewalk attempting to cross the road were the wheels, front fenders, and headlights. Based on these results, a suggestion is made to implement displays on at least two of the following regions: windshield, front fenders, side mirrors. These findings are valuable in the future design of signaling for autonomous vehicles in order to ensure pedestrians are able to see them on approaching vehicles. The software used provides a platform for similar works in the future to be conducted.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper describes a Unity-based simulation study of 15 vehicles across 57 scenarios each to quantify the frequency with which forward-facing exterior surfaces are visible to a virtual pedestrian camera positioned on the sidewalk. It concludes that wheels, front fenders, and headlights are the most visible surfaces and recommends placing eHMI displays on at least two of the windshield, front fenders, or side mirrors.

Significance. If the simulation parameters accurately capture real pedestrian viewpoints, the work could supply concrete guidance for eHMI placement on autonomous vehicles. The scale (15 vehicles, 57 scenarios) and the provision of a reusable simulation platform are positive features that could support follow-on studies.

major comments (2)
  1. [Simulation description] Simulation description (abstract and methods): no details are supplied on the visibility quantification procedure (ray-casting thresholds, occlusion test, surface sampling, etc.), nor are any statistical tests, error bars, or sensitivity results reported on the surface-exposure counts. This directly affects the robustness of the ranking that identifies wheels, front fenders, and headlights as most visible.
  2. [Simulation scenarios] Camera and scenario parameters: the chosen camera heights, lateral offsets, distances (min/max), and directions are presented without empirical grounding in measured pedestrian eye heights, typical sidewalk-to-lane distances, or head-orientation data, and no cross-validation or sensitivity analysis against field observations is described. Because the central claim and design recommendation rest on these frequency counts, the lack of validation is load-bearing.
minor comments (2)
  1. [Abstract] The abstract states that variables include 'camera position and direction' but does not enumerate the exact ranges or distributions used; a table or supplementary figure listing these parameters would improve reproducibility.
  2. [Conclusion] The recommendation to place displays on 'at least two' of the listed regions would benefit from a brief justification of the number two and from clarifying whether the three candidate regions are ranked by visibility or by some other criterion.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the detailed and constructive review of our manuscript. The comments highlight important areas for improving the transparency and robustness of our simulation study, and we have revised the paper to address these points directly.

read point-by-point responses

  1. Referee: Simulation description (abstract and methods): no details are supplied on the visibility quantification procedure (ray-casting thresholds, occlusion test, surface sampling, etc.), nor are any statistical tests, error bars, or sensitivity results reported on the surface-exposure counts. This directly affects the robustness of the ranking that identifies wheels, front fenders, and headlights as most visible.

    Authors: We agree that the original methods section provided insufficient detail on the visibility quantification procedure. In the revised manuscript we have added a complete description of the Unity ray-casting implementation, including the exact visibility thresholds, occlusion testing logic, and surface sampling method. We have also included statistical tests on the exposure counts, error bars on the reported frequencies, and sensitivity results that support the robustness of the ranking for wheels, front fenders, and headlights. revision: yes

  2. Referee: Camera and scenario parameters: the chosen camera heights, lateral offsets, distances (min/max), and directions are presented without empirical grounding in measured pedestrian eye heights, typical sidewalk-to-lane distances, or head-orientation data, and no cross-validation or sensitivity analysis against field observations is described. Because the central claim and design recommendation rest on these frequency counts, the lack of validation is load-bearing.

    Authors: We acknowledge the value of empirical grounding for the camera parameters. The revised methods section now cites peer-reviewed sources for average pedestrian eye heights, typical sidewalk-to-lane distances, and head-orientation ranges to justify the chosen values. We have additionally performed and reported a sensitivity analysis that varies these parameters within documented ranges and shows limited impact on the surface-exposure rankings. A new field-observation cross-validation study lies outside the scope of the present simulation paper; the sensitivity results nevertheless provide quantitative support for the stability of our conclusions. revision: partial

Circularity Check

0 steps flagged

No circularity; simulation outputs are independent of inputs

full rationale

The paper describes a Unity-based simulation that places cameras at chosen positions, heights, distances, and directions across 15 vehicles and 57 scenarios, then tallies visibility frequencies of surfaces such as wheels, front fenders, and headlights via ray-casting or occlusion checks. The resulting ranking and eHMI placement suggestion are direct empirical counts from these runs. No equations, fitted parameters, or self-citations are invoked to derive the ordering; the methodology is self-contained as a forward simulation whose outputs are not forced by construction to match any prior result or definition within the paper.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The reported conclusions rest on author-chosen simulation parameters and the unverified assumption that the game-engine model captures real pedestrian visibility sufficiently well for design recommendations.

free parameters (2)
  • Number of vehicles modeled
    15 vehicles selected to represent variety; exact selection criteria not stated in abstract.
  • Number of scenarios
    57 scenarios per vehicle chosen to vary position, surrounding traffic, camera angle, and distance.
axioms (1)
  • domain assumption Unity game-engine rendering and camera placement produce visibility counts that generalize to real-world pedestrian observation of moving vehicles.
    No validation against physical tests or real-world data is mentioned.

pith-pipeline@v0.9.0 · 5743 in / 1460 out tokens · 65339 ms · 2026-05-19T04:44:27.901635+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Defining an Evaluation Method for External Human-Machine Interfaces

    cs.RO 2026-04 unverdicted novelty 7.0

    The authors define and demonstrate a 223-question evaluation method for eHMI proposals, tested on four existing designs plus a kinematic baseline, suggesting a hybrid kinematics-plus-text approach as potentially stron...

  2. Using Unwrapped Full Color Space Recording to Measure the Exposedness of Vehicle Exterior Parts for External Human Machine Interfaces

    cs.RO 2026-04 unverdicted novelty 3.0

    Simulations show front vehicle parts are often obstructed, so eHMIs should use a distributive placement on windshield and fenders.

Reference graph

Works this paper leans on

13 extracted references · 13 canonical work pages · cited by 2 Pith papers

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

    work page

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

    work page

  3. [3]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

    work page

  4. [4]

    , author Kwigizile, V

    author Alhawiti, A. , author Kwigizile, V. , author Oh, J.S. , author Asher, Z.D. , author Hakimi, O. , author Aljohani, S. , author Ayantayo, S. , year 2024 . title The Effectiveness of eHMI Displays on Pedestrian – Autonomous Vehicle Interaction in Mixed - Traffic Environments . journal Sensors (Basel, Switzerland) volume 24 , pages 5018 . https://www.n...

    work page doi:10.3390/s24155018 2024

  5. [5]

    , year 2022

    author Aversa, D. , year 2022 . title Unity artificial intelligence programming : Add powerful, believable, and fun AI entities in your game with the power of Unity . edition Fifth edition ed., publisher Packt Publishing , address Birminham, UK

    work page 2022

  6. [6]

    , author Dodou, D

    author Bazilinskyy, P. , author Dodou, D. , author de Winter, J. , year 2019 . title Survey on eHMI concepts: The effect of text, color, and perspective . journal Transportation Research Part F: Traffic Psychology and Behaviour volume 67 , pages 175--194 . https://www.sciencedirect.com/science/article/pii/S1369847819302293, :10.1016/j.trf.2019.10.013. [ [...

    work page doi:10.1016/j.trf.2019.10.013 2019

  7. [7]

    , author Dietrich, A

    author de Clercq, K. , author Dietrich, A. , author Núñez Velasco, J.P. , author de Winter, J. , author Happee, R. , year 2019 . title External Human - Machine Interfaces on Automated Vehicles : Effects on Pedestrian Crossing Decisions . journal Human Factors volume 61 , pages 1353--1370 . https://journals.sagepub.com/doi/10.1177/0018720819836343, :10.117...

    work page doi:10.1177/0018720819836343 2019

  8. [8]

    title Office of Highway Policy Information - Policy Federal Highway Administration

    author Federal Highway Administration , . title Office of Highway Policy Information - Policy Federal Highway Administration . https://www.fhwa.dot.gov/policyinformation/tmguide/tmg_2013/vehicle-types.cfm. [ [@fishermans] (2024) ] fishermans_ford_2024 author [@fishermans] , year 2024 . title Ford Edge 2006 lowpoly for 3D -printing . https://sketchfab.com/...

    work page 2024

  9. [9]

    , author Lyu, W

    author Guo, F. , author Lyu, W. , author Ren, Z. , author Li, M. , author Liu, Z. , year 2022 . title A Video - Based , Eye - Tracking Study to Investigate the Effect of eHMI Modalities and Locations on Pedestrian – Automated Vehicle Interaction . journal Sustainability volume 14 , pages 5633 . https://www.mdpi.com/2071-1050/14/9/5633, :10.3390/su14095633...

    work page doi:10.3390/su14095633 2022

  10. [10]

    , author van der Kint, S

    author Rodríguez Palmeiro, A. , author van der Kint, S. , author Vissers, L. , author Farah, H. , author de Winter, J.C.F. , author Hagenzieker, M. , year 2018 . title Interaction between pedestrians and automated vehicles: A Wizard of Oz experiment . journal Transportation Research Part F: Traffic Psychology and Behaviour volume 58 , pages 1005--1020 . h...

    work page doi:10.1016/j.trf.2018.07.020 2018

  11. [11]

    , author Boissieux, L

    author Troel-Madec, M. , author Boissieux, L. , author Borkoswki, S. , author Vaufreydaz, D. , author Alaimo, J. , author Chatagnon, S. , author Spalanzani, A. , year 2019 . title eHMI positioning for autonomous vehicle/pedestrians interaction , in: booktitle Adjunct Proceedings of the 31st Conference on l' Interaction Homme - Machine , publisher Associat...

    work page doi:10.1145/3366551.3370340 2019

  12. [12]

    , author Bazilinskyy, P

    author de Winter, J. , author Bazilinskyy, P. , author Wesdorp, D. , author de Vlam, V. , author Hopmans, B. , author Visscher, J. , author Dodou, D. , year 2020 . title How do pedestrians distribute their visual attention when walking through a parking garage? An eye-tracking study . journal Ergonomics volume 64 , pages 793--805 . https://www.tandfonline...

    work page doi:10.1080/00140139.2020.1862310 2020

  13. [13]

    , author Dodou, D

    author de Winter, J. , author Dodou, D. , year 2022 . title External human–machine interfaces: Gimmick or necessity? journal Transportation Research Interdisciplinary Perspectives volume 15 , pages 100643 . https://www.sciencedirect.com/science/article/pii/S2590198222001051, :10.1016/j.trip.2022.100643

    work page doi:10.1016/j.trip.2022.100643 2022