I am currently in the forumlating stages. The current build plan is as follows:
1) Find a suitable motor for prototyping: currently considering the MBG36F 36V motor: http://www.goldenmotor.com/
2) Find a suitable open source motor driver with regenerative braking -- this is in the running, have contacted developer: http://vedder.se/2014/01/a-custom-bldc-motor-controller/#comment-8737
3) Find suitable capacitor array + make sure driver can handle it
-- This seems like a good candidate: http://www.digikey.com/product-detail/en/BCAP0350%20E270%20T11/1182-1016-ND/3079280
-- would need 36 / 2.7 = 13.3 = 14 caps. That means about $100 of super caps (when buying them in bulk), $150 when prototyping.
-- This would give us about 5 Watt Hours of power, which could run the 250W wheel for 71.4 seconds at full power. It would run a 150W wheel for 2 full minutes, which might be the wheel we design for in the end. (again, we are aiming for light weight that helps smooth out the ride, we are not turning the bike into a scooter!)
-- The caps would weigh 60g * 14 = 0.84 Kg, or 1.85 pounds.
4) Hook all these things together and get them to work. Start at step 1 until things look good.
5) While 1 - 4 are happening, work on Open Source Motor design and fabricate it. Use and test it.
6) When the above are finished, work on designing the hub to integrate all of them
-- Hub designed to act as heat sink
-- should also include high power resistors to do heat dump braking (if the caps are full or if the regenerative braking fails).
The goal is for the total added weight (compared to a traditional hub) to be < 5 pounds.
-- Look for organization / non-profit / grant to fund project. It is a good project and may get some interest.
-- connector to charge caps + add a battery pack (for those who want a scooter)
-- connector to plug USB device into like phone or light so that can charge.