Uplift Mobility Device

For many people with limited mobility, the hardest part of the day is not walking — it is standing up, and knowing there is somewhere to sit. A standard walker helps with one and ignores the other, so users either carry a separate seat or plan routes around where they can rest.

Uplift set out to combine both: powered assistance through the sit-to-stand transition, with a seat integrated into the frame rather than bolted on. That made the mechanism and the human factors inseparable from the start.

The device has to be stable through a transition in which the user's center of mass moves substantially — the least stable moment is exactly the moment the user most needs support. It also has to be light enough to be manoeuvred by the person using it, which puts stability and mass in direct conflict.

Seating added geometry constraints: the seat has to be at a usable height, clear of the legs in both states, and reachable without the user turning around. Those requirements bounded the linkage design more tightly than the actuation did.

Early work was about eliminating architectures rather than perfecting one. Configurations were laid out, checked for interference through their full range, and discarded where the seat fouled the legs, the footprint grew beyond doorway width, or the linkage needed force where a user's hands would be.

What survived was an architecture where the powered legs and the seat shared a frame and a coordinated motion, so a single transition served both functions instead of asking the user to operate two systems.

I led iterative development of the conduit linkages in Onshape — the parts that route and constrain motion between the frame and the powered legs. Linkages are unforgiving: a few millimetres of change at a pivot moves the whole path and can turn a smooth transition into a bind at one end of travel.

The loop was deliberately tight. Model a revision, print it, assemble it into the prototype, run it through full travel by hand, find where it bound or flexed, and change one thing. Most revisions changed a single dimension or a single fillet.

The two processes did different jobs. FDM in PLA was for fast form and fit checks where a part only needed to hold its shape long enough to be assessed. SLA in Grey Pro was for parts that had to be loaded and actually moved, where print anisotropy and layer adhesion would otherwise dominate the result.

Choosing between them was itself a design decision. Testing a load path on an FDM part and concluding the geometry was wrong, when the process was wrong, is a mistake that costs a whole iteration.

I coordinated with the electrical sub-team to integrate motors, house electronics and protect the battery. On a device people lean on, this is structural work: motor mounts carry real load, and an enclosure that flexes changes the alignment of whatever it holds.

The interface discipline was to fix the things the other team depended on early — mounting locations, available volume, cable routing — and then let both sides iterate inside those boundaries without breaking each other's work.

A battery on a mobility device sits low, near the floor, on a machine that will be bumped into door frames and curbs. The housing was designed for that reality: protected on the exposed faces, retained so it cannot shift in a knock, and serviceable without dismantling the frame.

Serviceability was a real requirement rather than a nicety. A device that has to be partly disassembled to change a battery is a device that will be used with a failing battery.

The prototype was assessed by running the transition repeatedly and watching for the things users would notice: motion that felt abrupt, flex that felt unsafe, and geometry that put a hand somewhere it shouldn't be.

The clearest lesson was that on a human-centered device, perceived stability and actual stability are separate requirements. A mechanism can be entirely adequate and still feel wrong — and for someone deciding whether to trust their weight to it, feeling wrong is a failure.