Adapting NIST Aerial Drone Tests for Thermal Identification, Inspection and Suppression of Wildfires
2025 IEEE 14th Annual Computing and Communication Workshop and Conference (CCWC 2025)
Case study · UAS · Test method design
Adapting Validated Robot Tests to Wildfire Response
Thermal identification, inspection, and suppression
Standard aerial robot tests assume a controlled environment; wildfire response does not. Validated test methods were adapted for a domain where sensing, terrain, and consequence all change the question.
01
Mission / context
Adapting standard aerial robot tests for thermal wildfire identification, inspection, and suppression scenarios.
The work begins from methods that were already validated, and examines what happens to them when the operational domain changes underneath them.
02
The engineering problem
A validated test method is validated for a context. Moving it to a new operational domain is not a matter of running it outdoors; the assumptions underneath it have to be re-examined one at a time.
01
The sensing modality changes.
Wildfire response leans on thermal sensing. A test built around visual targets does not carry over unmodified, and the parts that do carry over are not obvious in advance.
02
The environment is uncontrolled and hazardous.
Terrain, smoke, heat, and scale are not laboratory variables and cannot be held constant between trials.
03
The task set is different.
Identification, inspection, and suppression are three distinct operational tasks, and each one needs its own decomposition. Treating them as a single "wildfire capability" produces a test with no defined measurand.
04
The consequence of a wrong evaluation is high.
An overstated capability in this domain costs more than a wrong benchmark result does.
03
My role
Test engineer, collaboration lead, and co-first author of the paper this work produced.
I led the NIST and XPRIZE wildfire drone-response collaboration and co-led the student team that designed and built the Treefire Post and Thermal Campfire Omni test apparatuses introduced in the paper. I designed the 3D-printed concentric-C thermal acuity targets the method uses to make thermal detection measurable rather than impressionistic.
I prepared the manuscript for NIST editorial review board submission and drove the technical revision of the text through the review cycle. Alex Fraley, Ethan Nowery, and McCarthy Devine share first-author credit; the published order remains Fraley, Nowery, Devine, Jacoff, and Oh.
04
Technical approach
Start from validated methods, not from scratch.
The premise is adaptation of existing aerial drone test methods, carrying forward the repeatability properties they already had rather than inventing a new evaluation and hoping it holds.
Decompose the domain into task classes.
Thermal identification, inspection, and suppression as three separable evaluation targets rather than one undifferentiated capability.
Re-examine the sensing assumptions.
Thermal as the primary modality, which changes what a target is, what counts as a detection, and what the apparatus has to present.
Build the apparatus the adapted method needs.
Where the original test elements did not transfer, new ones were designed and built: the Treefire Post and Thermal Campfire Omni apparatuses, and 3D-printed concentric-C targets that give thermal acuity a graded, repeatable object to resolve.
Validate in the field, with domain collaborators.
Wildfire-oriented aerial evaluation and collaboration exercises with people who work the domain.
Write it up for a technical audience.
Technical writing as an explicit part of the method, not an afterthought. The point of adapting a method is that someone else can adopt the adaptation, which is why the work ends in a published conference paper rather than an internal note.
Primary figureMethod adaptation
Method adaptationValidated → re-examined → adapted
1Validated test method
Existing aerial robot test methods, validated for a controlled context, carrying the repeatability properties they already had.
- Targetsvisual
- Contextcontrolled
- Outputa score that survives repetition
2Domain assumptions re-examined
What survives the move, what has to be redesigned, and how to tell the difference.
- Sensingvisual→thermal
- Environmentcontrolled→uncontrolled, hazardous
- Task setone capability→three task classes
- Consequencebenchmark error→overstated capability
3Adapted method
Built where the original elements did not transfer; validated in the field with domain collaborators; written up so someone else can adopt it.
- Targets3D-printed concentric-C thermal acuity targets
- ApparatusTreefire Post · Thermal Campfire Omni
- Outputa published conference paper
Split across three evaluation task classes
Task class 1Thermal identification
Task class 2Inspection
Task class 3Suppression
05
Test / analysis method
- Wildfire-oriented aerial evaluation and collaboration exercises.
- Scenario adaptation of validated aerial robot test methods.
- Thermal and terrain test apparatus development: Treefire Post, Thermal Campfire Omni.
- 3D-printed concentric-C thermal acuity targets.
- Field observations and research analysis.
Toolssmall UASthermal sensingvalidated aerial robot test methodsadditive manufacturing
06
What the evidence shows
Reusable wildfire-response evaluation concepts, purpose-built thermal test apparatuses, and a published conference paper covering thermal identification, inspection, and suppression.
Publication
Co-first author, with Ethan Nowery and McCarthy Devine.
07
Results
| Result | Value |
|---|---|
| Output | Reusable wildfire-response evaluation concepts, purpose-built thermal test apparatuses, and a published conference paper |
| Task classes addressed | Thermal identification · inspection · suppression |
| Primary sensing modality | Thermal |
| Apparatus introduced | Treefire Post · Thermal Campfire Omni · 3D-printed concentric-C thermal acuity targets |
| Program period | 2023–2025 |
| Publication | IEEE CCWC 2025, Las Vegas, NV; shared first authorship with Ethan Nowery and McCarthy Devine |
| Outcome or performance metric | None established |
08
Why it matters
This is the clearest demonstration on the site of what standards-grade test methods are actually for.
The aerial test-method work establishes that I build repeatable evaluation methods. This project carried those methods into wildfire response, where the cost of an overstated capability is high. That transfer is the evidence that the methods were built to be portable across operational contexts rather than specific to one site. Standards work exists precisely so that an evaluation survives a change of context; this project demonstrates that claim rather than asserting it.
The most interesting part is not what transferred. It is what did not, and how you find out: a method that quietly fails to transfer produces results that appear precise but do not measure the intended capability.
The publication is also independently verifiable: the paper is public, and the venue is named.