Alex Fraley Aerospace · Robotics · Test
Contact

Projects · Selected work

Engineering stories.

Selected work, told with its objective, method, and result. The deepest stories come from the NIST chapter; the rest span a student-built race car, a lunar rover concept, and the independent engineering I do now under Curiosell Systems.

From the NIST chapter

Five stories from the field.

Test methods

Aerial tests responders can rebuild

Operational tasks restated as buildable, repeatable field tests, complete enough to travel between sites.

Apparatus CADRepeatable procedures
The full story

ObjectiveDetermine whether an operational task, such as search, inspection, or delivery, could be restated as a test another team could reproduce without the original authors present.

What I didDeveloped and validated aerial test methods with fabricated field apparatus, machine-readable image targets, documented procedures, and CAD models carrying the geometry.

ResultMethods that traveled. The CAD and procedures were complete enough for another site to build the same setup and obtain comparable numbers.

Operational meaningAn agency evaluating a new aircraft can run the same test at its own site and compare its numbers directly with published results.

Apparatus CAD
Scoring software

Scoring partial views honestly

Flight cameras rarely capture the whole target. The scoring workflow accumulates evidence instead of demanding a perfect frame.

Automated scoringFrame-level traceability
The full story

ObjectiveEstablish whether automated scoring could remain fair and defensible when no single frame contains the full target.

What I didDesigned and implemented an experimental computer-vision scoring workflow in Python that samples robot-camera video, applies partial-target logic, and resolves detections through a versioned score schema, with human review retained where judgment mattered.

ResultEvaluation events scored across the program, with every score traceable to its exact frames.

Operational meaningA disputed score is settled by opening the frames behind it, not by argument.

Score trace view
Field program

Outdoor small-UAS drop testing

Outdoor releases measured by non-contact video trajectory analysis, run as a field program end to end.

Field program end to endNon-contact measurement
The full story

ObjectiveMeasure how released objects actually behave from release to ground rather than assuming it.

What I didThe full program: release procedure, redesigned release and stop hardware and guide-rail concepts to reduce run-to-run variability, field administration, and the video-based trajectory measurement behind the results.

ResultMeasured terminal-velocity results across test objects, documented in internal NIST technical reporting.

International events

Quadruped trials at RoboCupRescue, ICRA, and IROS

Mobility trials run live while the event moved around them.

3 event series2024 arena CAD
The full story

ObjectiveCharacterize quadruped mobility on standardized terrain, and establish whether the terrain itself could travel between venues.

What I didTrial execution at three international event series, and CAD of the 2024 championship arena including a metric 3D floorplan.

ResultComparable trial data across venues, and arena geometry documented thoroughly enough to reconstruct elsewhere.

Quadruped mobility trial
Wildfire · 2023–2025

Adapting the tests for wildfire response

Thermal in place of visual, uncontrolled in place of controlled, and three separately testable tasks in place of one. The full case study covers the method.

3 task classesIEEE CCWC 2025
The full story

ObjectiveIdentify which parts of a validated aerial test survive a move into wildfire response, and which have to be redesigned.

What I didLed the NIST and XPRIZE collaboration; co-led the student apparatus team behind the Treefire Post, the Thermal Campfire Omni, and the concentric-C thermal targets.

ResultA published IEEE CCWC 2025 paper, co-authored with Ethan Nowery, McCarthy Devine, Adam Jacoff, and Paul Oh.

Operational meaningWildfire programs evaluating drone capability gain test methods designed for thermal sensing and uncontrolled sites, not benchmarks borrowed from the lab.

Thermal apparatus build
Read the full case study
Earlier engineering · University of Maryland

Terps Racing: Formula SAE Electric.

Formula SAE Electric · Low Voltage team

Whole-vehicle engineering, learned on a student-built electric race car.

Before the robotics work, I was a member of the Low Voltage team on Terps Racing's Formula SAE Electric car. Electrical, mechanical, aerodynamic, safety, packaging, and manufacturing decisions all had to coexist in one vehicle, and the vehicle then had to survive being driven hard.

It was my first sustained look at what happens when a subsystem that works on the bench meets the rest of a vehicle. The build-and-test culture of that shop carried directly into the test engineering that followed.

Low-voltage system sketch. Illustrative, not a team drawing.

Team photograph
Competition vehicle
Low-voltage hardware
Independent engineering · Curiosell Systems

Current independent work.

Current technical work spanning release engineering, robotics software, procedural simulation environments, computer vision, and edge and embedded systems.

Release engineering · Curiosell Systems

CURI: a release system that verifies itself

Software distribution with test-engineering habits. If the installer claims it, the system can check it.

Deterministic buildsRelease evidence
The full story

ObjectiveDetermine whether a release pipeline could verify its own work deterministically, end to end.

What I didCURI, an independent Curiosell Systems project covering deterministic build workflows, release verification, installation and installed-state verification, and reversible deployment with rollback, uninstall, and uninstall preview.

ResultA working release-engineering pilot in which each stage leaves release evidence that can be independently verified.

Release verification output
Simulation infrastructure · Curiosell Systems

DRONIEUS: procedural simulation environments

Procedural drone-simulation environments and infrastructure tooling built around SketchUp Ruby generators, reusable asset pipelines, and Unreal-ready scene workflows.

SketchUp Ruby generatorsUnreal-ready workflows
The full story

ObjectiveGenerate high-fidelity simulation infrastructure procedurally instead of hand-modeling every scene.

What I didSketchUp Ruby generators for infrastructure and terrain elements, a reusable environment-asset pipeline, and scene workflows prepared for Unreal Engine consumption.

ResultA growing library of reusable environment assets and generated scenes feeding the simulation work, developed incrementally.

Procedural infrastructure assets
Generated environment view
Project repository
Flight simulation · Development paused

Unreal Engine flight simulation

A drone flight-dynamics simulator in Unreal Engine 5.7, building on the DRONIEUS environment work. Development is paused on a documented pawn-class/FDM defect.

Unreal Engine 5.7Development paused
The full story

ObjectiveExplore how far a custom Unreal Engine 5.7 environment can carry flight-dynamics and drone-scenario simulation work.

What I didSimulator development in Unreal Engine 5.7: flight-dynamics work and scene workflows that consume DRONIEUS-generated environments.

ResultA baseline simulator whose development is paused on a documented pawn-class/FDM defect. No validated flight model and no completed SITL/HITL work is claimed.

Simulator scene capture
Robotics software · In development

JetBot Control: mobile robot control

An iOS app for driving and monitoring JetBot robots: an MVP validated on device, with Android work in progress.

iOS MVP device-validatedIn development
The full story

ObjectivePut dependable teleoperation and telemetry for a JetBot platform in an operator's hand.

What I didThe iOS MVP, validated on device against real hardware. Android builds and tests exist alongside the iOS work.

ResultA device-validated iOS MVP moving toward release. Not yet distributed through the Apple App Store.

App icon and device control
App Store listing
Edge and embedded

Edge AI and embedded prototyping

A working bench of edge and embedded platforms: Jetson-class edge AI, JetBot robotics, and ESP32-class microcontroller work.

Jetson · ESP32Edge computer vision
The full story

ObjectiveKeep the computer-vision and automation work honest by running it on real edge hardware.

What I didExperimentation across Jetson Nano and Orin Nano, JetBot platforms, and ESP32 and ESP8266 microcontrollers: edge computer vision, IoT telemetry, and local automation.

ResultPrototypes that feed the robotics and simulation work, kept deliberately small and public-safe.

Bench and edge hardware
Aerospace capstone · University of Maryland

Lunar rover concept

The aerospace capstone at Maryland: autonomy and systems design for a vehicle operating beyond physical reach.

AutonomySystems design
The full story

ObjectiveBalance autonomy, power, mobility, and mission constraints inside one coherent vehicle design.

What I didThe autonomy and systems design work for the rover concept, as the capstone of an aerospace engineering degree.

ResultA first complete exercise in systems engineering: every subsystem answering to the whole vehicle.

The full sequence is on the timeline.

Open the timelineGet in touch