Mobility scooters are usually wheeled chairs. However, many potential users don't have a need to be seated - they have perfectly good legs and being constantly seated makes for poor posture as well as placing the user at a height disadvantage against people able to stand.
'Perfectly good legs' sounds absurd in the context of a disability scooter, but for people with Type 2 Diabetes the problem is not leg strength but balance : they are unable to maintain good balance - especially when standing still - due to diabetic neuropathy, which attacks nerve endings in the sole of the feet, forcing them to rely on inner ears and visual cues.
A good compromise would be a Segway, but these are forbidden in some countries due to poor regulatory oversight and may well be insufficiently stable or well-controlled for a person with limited balance to use. An experimental platform aimed at optimising for poor balance and recovery from out-of-balance condition would be a good match for an intelligent servo controller such as TAPAS.
Two specific aims would be :
Bridge small gaps - such as an entrance to public transport - using a mixture of rapid acceleration, controlled braking and a mechanical bridging lever deployed in sync with the acceleration to hop over a variable-height gap.
Roll up and down stairs, braking and reversing as required to drive the scooter into the stair corner whilst keeping it upright. This requires powering forward until the scooter drops over the edge of a step with almost zero forward speed, powering backward to retain a stable position in the step corner, and powering to a balanced upright position until ready for the next step, or flowing forward to roll down the steps at a controlled speed.
There is considerable overlap with another project suggestion 'self balancing wheelchair' though the aims are different. A collaborative project might be appropriate. Like that idea, a pair of boards to operate two BLDC motors would be required.