Parkinson's Gait Assistant

Interdisciplinary capstone — 11 students across mechanical, electrical, and computer information systems engineering.

The brief

Parkinson's disease is a progressive neurological disorder that gradually destroys dopamine-producing neurons. As it progresses, patients lose motor control: tremors emerge, gait shortens, arm swing reduces, and falls become more likely. There is no cure. The work right now is on monitoring disease progression and supporting research into treatments that slow it down.

An external sponsor came to UMass Dartmouth with a question: could a team of senior engineering students build a wearable system that monitors a Parkinson's patient's movement throughout the day, detects tremor and gait abnormalities in real time, gives the patient a haptic alert when correction might help, and stores the data so clinicians can review trends over weeks and months?

The system the team built has four wearable devices — two for the wrists, two for the ankles — that each contain a microcontroller, an inertial measurement unit (IMU), a Bluetooth 5.4 radio, a lithium-polymer battery, and a vibration motor. The devices send motion data to a mobile app. When the system detects motion in the 4–6 Hz frequency band (the standard resting-tremor range for Parkinson's), the vibration motor fires as a feedback cue to the user. A portable charging station holds and recharges all four devices together when they're not in use.

This was the second year of the project. A previous team had built a proof-of-concept, but the haptic motor didn't reliably trigger, there was no working tremor detection algorithm, and the sponsor wanted the device significantly smaller. Our team treated it as a redesign rather than a refinement: new system architecture, new electronics, new housing.

My role

I was on the three-person mechanical engineering team. My focus was materials research and testing, with supporting CAD work on the housing and charging station revisions. The bulk of my hands-on contribution lived in the materials testing campaign — characterizing the prototype material, designing the test plans, running the analysis, and writing the production-material recommendation that follows.

The housing decision: a tradeoff, not a compromise

The customer asked for something that looked and wore like a recent-model Apple Watch — roughly a 40 mm face. We modeled and considered two housing options.

Option A was a fixed-lid design with the cover bolted in place at four corners. The lid is permanently fixed. The advantage of this design is that it makes waterproofing genuinely possible: a continuous perimeter is easy to seal with an O-ring or gasket, and there are no breakaway joints to compromise the seal.

Option B was a snap-cover design with a removable lid. A groove runs along the centerline of the housing walls, the lid snaps into it, and an ergonomic tab lets the user pop it off for battery replacement and internal maintenance. The strap interface — slots on opposite ends of the body for interchangeable wrist straps — is the same on both designs.

We chose Option B.

The reason matters: the customer explicitly wanted the user to be able to access the internals. That decision unlocks repairability and battery replacement, both meaningful for a product worn daily by a patient population. But choosing Option B also meant choosing against the higher IP ratings. A removable lid is a seam. A seam is a leak path. There is no clean way to put a continuous gasket along the parting line of a snap-fit lid while also keeping wall thickness low enough to fit inside a 40 mm wearable.

The waterproofing limitation in the final prototype wasn't a failure. It was the documented downstream effect of a customer-driven access requirement.

We tested the assembled device under IP60–IP64 conditions — light spray exposure with a standard spray bottle. The device passed: no faults, no malfunctions. We did not meet IP65–IP67 (high-pressure jet or submersion), and the report documents why: the small footprint prevented O-ring placement, the wall thinness required for compactness limited sealing options, and FDM 3D printing introduces interlayer porosity that makes printed parts inherently non-watertight regardless of geometry. Any one of those constraints would have made full waterproofing hard. Together, they made it impossible at the prototype stage.

Materials: why ASA, and what the testing showed

The prototype material was ASA (acrylate styrene acrylonitrile), a UV-resistant thermoplastic developed as an alternative to ABS. It was selected based on three considerations: UV resistance for outdoor wear, impact tolerance for a device worn through daily activity, and accessibility within the project's budget on the on-campus 3D print system.

ASA is not certified under ISO 10993 for prolonged skin contact. This is the most important caveat in the entire materials analysis. The decision to test ASA wasn't a claim that ASA was the right production material; it was a deliberate scope choice to characterize the prototype material so that the testing methodology and analysis would transfer cleanly to whatever production material the team ultimately recommended.

The tensile testing campaign followed ASTM D638-style proportions. Five dogbone specimens were 3D printed in ASA with 80% infill at a 0.01 inch layer height, with a gauge length of 27.53 mm and a cross-sectional area of 19.2 mm². Each specimen was tested at a different crosshead displacement rate (10 mm/min, 2 mm/min, 0.5 mm/min, 2 mm/min, 1 mm/min) to investigate strain rate sensitivity. Force and displacement were recorded continuously on an Instron universal testing machine until specimen fracture. Engineering stress and engineering strain were calculated from the original cross-section and gauge length.

A second test plan validated surface strain measurements using strain gauges instrumented on four 2 mm thick flat ASA plates. The strain gauge data was cross-referenced against the Instron-derived engineering strain to confirm measurement accuracy.

The data informed two conclusions. First, the prototype housing as printed meets the mechanical demands of daily wear at the stresses expected from the system's mass and the wearer's normal range of motion. Second, and more important, anisotropy from layer-by-layer FDM construction means strength varies depending on print orientation, and the material would not be reliable under repeated long-term loading at production scale.

What production would need to look like

The report includes a full analysis of why 3D printing is not a viable production process for this device, and what would need to change. The key points:

• FDM porosity — Microscopic gaps between deposited layers prevent true waterproofing regardless of geometry. Post-processing like epoxy coating or vapor smoothing can mitigate this temporarily, but a wearable that contacts skin and sweat daily will degrade those coatings and reintroduce leak paths. A sealed-but-porous device can also grow microorganisms inside the housing wall itself, which no surface cleaning will reach.
• Print-to-print consistency — Variations in printer calibration, ambient temperature, and filament batch produce dimensional and mechanical inconsistencies. For a device that needs to seat a PCB, battery, and motor in tight, repeatable cavities, that variation is unacceptable.
• Injection molding — Molten thermoplastic injected into a machined cavity produces fully dense, non-porous parts with tight tolerances and consistent mechanical properties. It is the only practical way to produce four housings per device across thousands of units on a manufacturable timeline.

If this design were to move toward production, the recommendation is: switch to an ISO 10993-certified, biocompatible thermoplastic (medical-grade ABS or polycarbonate are the obvious candidates), produce housings via injection molding with a continuous-perimeter gasket seal, and reconsider Option A (fixed lid) at that point — injection molding makes a permanent lid feasible without sacrificing manufacturability, and that change alone unlocks the full IP67 rating.

Outcome

The system worked. The wearable devices detected motion in the Parkinson's tremor frequency range and triggered haptic feedback. The Bluetooth connection to the mobile app held, motion data was transmitted and stored to the cloud, and the charging station recharged all four devices in parallel from a single USB-C input. The final prototype was demoed at the department capstone showcase with a live demonstration session.

Future improvements documented in the report include a smaller double-sided PCB to further reduce wearable size, expanded system testing across a real patient population, and the production-stage material and process changes described above.

What I took from it

The engineering problem this project poses — small, body-contact, battery-powered, sealed, electromechanical device with a vibration motor inside it — is structurally the same problem the consumer wellness and pelvic health industries are solving. The materials questions, the sealing tradeoffs, the IP rating considerations, the choice between repairability and full waterproofing, the gap between prototype processes and production processes: all of it transfers directly.

The thing I'll carry forward is the housing tradeoff. A customer asked for internal access. We honored the request and documented what it cost us in sealing performance. That isn't a failure mode of engineering. That's engineering working correctly — making the tradeoff visible instead of hiding it.