Research Is the Engine

Not a side activity. Every Fitness Tech device starts as a research question — we find a gap in what existing fitness technology can measure, then build the hardware that closes it.

That ordering is deliberate. We don't build a gadget and then look for something to study. The Research team defines the question before the Hardware team builds the answer — and validation against commercial and medical-grade equipment is part of the build, not an afterthought. Members get hands-on experience with literature review, testing protocol design, data collection, statistical analysis, and model training.

01 — The cycle

How a Question Becomes a Finding

Every project moves through the same loop, and what we learn points at the next device

1

Identify a Gap

Something coaches or athletes need to know that no accessible product measures well

  • Literature review
  • Coach & athlete interviews
  • Gap identification
  • Scoping the measurement
2

Build a Prototype

Address the gap with hardware, moving through the staged build pipeline

  • Testing protocol design
  • Sensor selection criteria
  • Prototype development
  • Initial bench testing
3

Collect Real Data

With our own hardware and our own app — not someone else's black box

  • Hardware sensor deployment
  • In-app data sampling
  • User testing sessions
  • Dataset compilation & labelling
4

Analyze

Accuracy, correlation, and reliability across users and conditions

  • Data preprocessing
  • Statistical analysis
  • Benchmarking vs. reference devices
  • ML model training
5

Document & Feed Back

Present the findings, then point what we learned at the next device

  • Research documentation
  • Results visualization
  • Presentation preparation
  • Academic submissions
02 — Open questions

What We're Actually Investigating

Each question maps to a device in the portfolio — the question came first

01

How accurately can student-built sensors capture biometrics compared to commercial and medical-grade devices?

02

Can motion data from a wrist-worn IMU reliably classify exercise type and count reps across different users and different form quality?

03

What sampling rate and sensor configuration is the minimum viable one for trustworthy rep detection?

04

How does hydration — measured automatically rather than self-reported — actually correlate with training performance and recovery markers?

05

Can a load cell and flow sensor combination achieve volume accuracy sufficient for research use in a consumer form factor?

06

Does adaptive hydration targeting improve adherence compared with a fixed daily goal?

07

What does a coach need to see, and at what frequency, to make better roster decisions from live athlete data?

08

Which combination of signals gives the earliest useful warning that an athlete is trending toward overtraining or injury?

09

How much does wearability and comfort affect adherence over a multi-week training block?

Why our own hardware matters here

Commercial devices are black boxes. You can't see how the data is processed, can't customize what's tracked, and can't use them to answer new questions. Because we design the sensor platform, the firmware, and the app that receives the stream, the entire pipeline stays inspectable — which is the difference between a study we can defend and a number we have to take on faith.

See the devices these questions produced →

03 — Where the work goes

Publication & Presentation Pathways

Building a portfolio of validated, publishable work that future cohorts inherit and extend

UW Undergraduate Research Symposium

  • Our primary near-term target
  • Present device accuracy and validation work
  • Build a portfolio for grad school applications

Conference Submissions

  • IEEE Region 6 Student Conference
  • ACM Student Research Competition
  • Pathways being established now

Faculty Collaboration

In Progress

We're working to formalize faculty partnerships and UW lab collaborations for mentorship and academic guidance.

What Validation Means Here

A number isn't a finding until we know how far to trust it

Reference benchmarking
Every device is measured against commercial and medical-grade equipment before we claim anything about its accuracy.
Cross-user reliability
Does the measurement hold across different body types, different form quality, and different sports — or only for the person who built it?
Actionability threshold
Coach interviews tell us what accuracy makes a measurement worth changing a training plan over. That threshold defines the study.
Human subjects pathway
An IRB pathway is being pursued for the athlete testing work, so the data we collect can support real submissions.
04 — Get involved

Research Team Roles

No prior research experience required — the team teaches the method as it runs the studies

Literature Review

Survey existing studies on wearable accuracy, hydration science, and training-load monitoring to find the gap worth building for

Protocol Design

Design the testing procedure: what gets measured, against what reference, under what conditions, with how many subjects

Coach & Athlete Interviews

Run the discovery conversations that decide which device we build next and what the coach dashboard has to show

Data Collection

Run user testing sessions with our own hardware, and build the labelled datasets that train our classification models

Data Analysis

Process and analyze collected data with statistical methods and visualization tools; benchmark against reference devices

Documentation

Write research reports, build presentations, and prepare symposium and conference submissions

Interested in Research?

Join the team that decides what we build next — and gets to find out whether it works.

Apply Now
Find the gap. Build the device.
Collect the data. Prove the number.