Innovation Project Overview

Safer Mapping for Fragile, Hard-to-Reach Archaeological Sites

Our team designed a rover concept that helps archaeologists document dark, unstable environments—like caves and ancient structures—without putting people or artifacts at risk. We combined field research, rapid prototyping, CAD iteration, and expert feedback to build a solution that feels ready for real-world testing.

Prototype Platforms: LEGO + CAD
CAD Tools: Tinkercad → Onshape
Fabrication: Bambu Lab P1S (planned)
Expert Input: Archaeologist Interview
  • Remote operation to reduce risk We designed for safer documentation when visibility is limited and terrain is unpredictable.
  • Modular components for field repairs Attachments and camera mounts can swap quickly to adapt to different spaces and mission goals.
  • Mapping-ready sensor placement We prioritized camera height and blind-spot reduction based on expert feedback.
Identify — Anthony

We started by learning what can go wrong in real archaeology.

Before building anything, we investigated the biggest challenges archaeologists face when documenting fragile history. We reviewed documentaries, read articles, studied artifacts at the Metropolitan Museum of Art, and discussed how excavation and documentation can unintentionally damage objects.

We found that artifacts are extremely sensitive to environmental changes. Conservation guidance commonly emphasizes stable conditions and warns that higher humidity can increase the risk of mould growth in collections [2]. Professional conservation groups also recommend “middle range” humidity targets for mixed materials to reduce long-term deterioration [1].

One challenge stood out as both urgent and solvable: mapping dark, unstable, and hard-to-reach spaces such as caves and ancient structures. In robotics research, cave surveying is often described as a “total darkness” problem that requires specialized sensors and lighting [5]. That evidence helped us define our Innovation Project goal:

Safer exploration Better visibility Reduced artifact impact

Problem Statement

Archaeologists need a safer way to document and map dark, unstable, tight spaces—without increasing the risk to people or causing accidental damage to fragile cultural heritage.

Project Plan & Team Collaboration — Kyle

We planned like an engineering team—clear roles, fast brainstorming, tight organization.

We built a detailed project plan for a rover-based solution and brainstormed using IDEO-style rules: defer judgment, build on others’ ideas, stay visual, and go for quantity. We sketched multiple rover layouts, discussed tradeoffs, and tracked tasks in a shared spreadsheet so deadlines and responsibilities stayed visible.

Project Plan — Snapshot
To Do In Progress Done Define scope & interview expert Build cardboard mockup Finalize camera placement Prepare test plan LEGO stability tests CAD sync + iterate Timeline: Weeks 1–4 (progress shown)
Our assignments, status and timeline at a glance.
Ideation Wall — Sketches
Camera Height Raise 10–15 cm to capture walls Modularity Quick-swap mounts for field repairs Lighting & Sensors Integrate LED ring + distance mapping
Early ideas and decisions from our brainstorm.

Team Roles

Mobility: stability for uneven terrain • Sensing: camera/sensor ideas for mapping • Durability & Safety: protective bumpers + repairability • Documentation: engineering notebook + media

Create — Julissa

We built multiple rover versions—physical + digital—to prove the idea is real.

We began with a cardboard model to lock in size, layout, and component placement quickly. Then we built LEGO prototypes to test stability, turning radius, and how the rover behaves on different “field-like” surfaces.

In parallel, we created CAD models—first in Tinkercad for fast revisions, then in Onshape for a more detailed, assembly-style layout. This let us experiment with camera placement and proportions before finalizing our concept.

Design Priorities

Compact & agile Camera + sensor integration Remote operation Modular attachments
Cardboard / top view
Cardboard / Top View
Early layout of the camera, hub, motors and protective bumper.
LEGO tracked prototype
LEGO Tracked Prototype
Angled (3/4) photo showing terrain and stability testing.
LEGO wheeled prototype / assembled view
LEGO Wheeled / Assembled View
Tight turns and turning radius on a smooth surface.
CAD / left side
CAD / Left Side View
CAD model of the rover, left side.
Iterate — Brian

We used feedback like professionals: collect → prioritize → update → re-test.

After our early prototypes and CAD models were ready, we shared them with other teams, mentors, family members, and two key reviewers: Grace Howe (archaeologist) and Rawlins (NYC event organizer). Grace recommended raising the camera to capture wall and ceiling detail, and Rawlins pushed for modularity so repairs would be easier in the field.

Based on that feedback, we improved terrain handling, reduced blind spots by repositioning sensors, and updated our CAD models so the design matched our real revisions.

Key Improvements

  • Adjusted wheel/track placement for stability on uneven terrain
  • Raised and repositioned camera to reduce blind spots
  • Improved modular attachments for faster swaps and repairs
  • Synced Tinkercad + Onshape models to match the newest version
Code & Graphs — Data Visualization

Scan-processing Python code and a browser graph of the same data

Below is the Python script used to turn range scans into a point cloud, a browser-rendered scatter plot that reproduces the output, and a Tinkercad screenshot showing the CAD layout.

Python: download

import math
import matplotlib.pyplot as plt

scan = [
    {'back_mm': 2000, 'angle_index': 0, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': -1, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': -2, 'front_mm': 402},
    {'back_mm': 2000, 'angle_index': -3, 'front_mm': 428},
    {'back_mm': 2000, 'angle_index': -4, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': -5, 'front_mm': 1149},
    {'back_mm': 2000, 'angle_index': -6, 'front_mm': 150},
    {'back_mm': 296,  'angle_index': 0, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': 1, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': 2, 'front_mm': 333},
    {'back_mm': 2000, 'angle_index': 3, 'front_mm': 2000},
    {'back_mm': 2000, 'angle_index': 4, 'front_mm': 1141},
    {'back_mm': 2000, 'angle_index': 5, 'front_mm': 1124},
    {'back_mm': 2000, 'angle_index': 6, 'front_mm': 190},
]

# --- TUNE THIS ---
# angle_index isn't degrees; it's a step counter.
ANGLE_STEP_DEG = 15
NO_HIT_MM = 2000

front_x, front_y = [], []
back_x, back_y = [], []

for s in scan:
    a_deg = s["angle_index"] * ANGLE_STEP_DEG
    rf = s["front_mm"]
    if rf is not None and rf < NO_HIT_MM:
        th = math.radians(a_deg)
        front_x.append(rf * math.cos(th))
        front_y.append(rf * math.sin(th))
    rb = s["back_mm"]
    if rb is not None and rb < NO_HIT_MM:
        th = math.radians(a_deg + 180)
        back_x.append(rb * math.cos(th))
        back_y.append(rb * math.sin(th))

plt.figure()
plt.scatter(front_x, front_y, marker="o", label="Front sensor hits")
plt.scatter(back_x, back_y, marker="x", label="Back sensor hits")
plt.scatter([0], [0], marker="s", label="Robot (0,0)")
plt.axis("equal")
plt.grid(True)
plt.xlabel("mm forward (x)")
plt.ylabel("mm sideways (y)")
plt.title("Distance Scan Point Cloud")
plt.legend()
plt.show()
          

Browser Graph (replicates the Python point cloud)

Interactive scatter plot rendered with Chart.js — front/back sensor hits.

Tinkercad Screenshot

Tinkercad screenshot

Screenshot showing the CAD layout used for measuring camera height and mount positions.

Communicate — Jonathan

We explained the “why” and the “how,” not just the build.

We focused on telling a clear story: what problem we chose, what evidence supports it, how our rover works, and why our design choices reduce risk to people and artifacts. Our presentation included physical LEGO prototypes, CAD models from Tinkercad and Onshape, sketches, and drawings so reviewers could see our thinking at every step.

3D-printed rover, assembled
3D-Printed Rover
Assembled version 1, designed in CAD.
Final rover prototype, corner view
Final Prototype
The wooden rover that maps its surroundings with a camera.

Impact Summary

Our rover concept supports safer documentation by keeping people at a distance, improving visibility with deliberate camera placement, and enabling modular repairs so the work can continue without risky improvisation inside sensitive sites.

Evidence & Research

We anchored our design in conservation and robotics research.

Our project sits at the intersection of heritage preservation and robotic mapping. Conservation guidance commonly recommends mid-range humidity targets for many cultural materials [1], and warns that high humidity can increase the risk of mould growth [2]. That supports our emphasis on minimizing unnecessary exposure and handling.

On the robotics side, cave surveying research explicitly addresses mapping in “total darkness” using lights and sensors [5], and surveys of robotics in archaeology describe how tools like LiDAR and 3D modeling help reach inaccessible sites and improve documentation precision [3].

45–55% RH
A commonly recommended museum target range for many mixed cultural materials [1].
65%+ RH
Canadian conservation guidance notes mould growth is promoted in higher humidity ranges (e.g., 65–100%) [2].
Total darkness
Cave surveying research describes darkness as a core challenge requiring sensors + lighting [5].
Humidity Risk Bands (Illustrative from published guidance) Illustrative humidity guideline ranges

Note: This chart visualizes guideline ranges, not a prediction model. Always follow the policies of the museum/site you work with.

References

Sources cited on this page

  1. International Institute for Conservation (IIC) & ICOM-CC. Environmental Guidelines (Declaration). Link
  2. Canadian Conservation Institute (CCI). Mould Growth on Textiles (notes that RH 65–100% promotes mould growth). Link
  3. Kyriakoulia, P. (2025). Survey on the Application of Robotics in Archaeology (Open Access, PMC). Link
  4. Klehm, C. et al. (2024). Remote, Rugged Field Scenarios for Archaeology and the Field Sciences (NSF-hosted PDF). Link
  5. Tabib, W. et al. (2020/2021). Autonomous Cave Surveying with an Aerial Robot (mapping in total darkness; arXiv/IEEE T-RO). Link

Additional sources are listed in the project repository.