I like having a process. Not because projects follow it perfectly, but because a structured process gives me a way to make good decisions when the problem gets complicated.
My education was built around engineering design: requirements, brainstorming, concept development, decision matrices, analysis, prototyping, testing, and iteration. I still work that way. Each stage below links to where it happened on a real project.
The process is structured. The path through it usually isn't.
01→
Understand
the people and the problem
We learned the student loved playing with large LEGO-style blocks, so the instrument was designed around that instead of around a conventional instrument.
Adaptive instrument: Understand02→
Define
requirements and constraints
Hold a height setting through real use, let the caster swivel freely, stay easy for an operator to change, and support mowing under low obstacles like solar panels and fences.
Height-of-cut mechanism: Define03→
Explore
more than one concept
Everyone on the bike team brought their own vehicle concept from the previous semester, so we started with several complete designs, not one.
Adaptive bicycle: Explore04→
Evaluate
decide with evidence
A weighted decision matrix, weighted by what mattered most to the client. Concept A scored highest, and we still carried the strongest subsystems from every concept into the final design.
Adaptive bicycle: Evaluate05→
Build
something you can measure
A load cell calibrated on a hanging-load rig, then a rigid bench fixture designed in CAD so the sensor sits in the real force path.
Knee traction: Build06↓
Test
against criteria set up front
Shop checks, overextension with an impact tool, curb impacts recorded with an accelerometer and camera, and field use, against test plans and acceptance criteria set with the test engineers.
Height-of-cut mechanism: Test
07↺
Iterate
go back as far as the evidence says
Test↩sent the work back to03 Explore
Buttons needed a precise press to trigger reliably. We went back to concepts and switched to light sensing, so placing a block is enough.
Adaptive instrumentTest↩sent the work back to02 Define
The client sat on the alpha prototype and told us what worked and what didn't. His feedback changed features, materials, and dimensions for the beta.
Adaptive bicycleBuild↩sent the work back to03 Explore
Partway through building the first bench fixture, it no longer represented the system I wanted to test. I stopped before testing it and went back to CAD.
Knee traction