One major problem with a flywheel is its gyroscopic action which makes it difficult to turn the vehicle in any axis other than the direction of rotation (perpendiculat to the axis). You can use flexible mounting to allow it to "float", but then the mechanical connection to add and extract energy becomes more difficult. There is also the danger of failure and violent explosion, so you need to make it especially well balanced with good high speed bearings and a bulletproof blast shield.
Capacitors may be useful if you have lead-acid cells, and they are also needed to stabilize the DC bus at the controller to minimize ripple and high current surges through the connections to the battery pack. But such capacitors do not store much energy - just enough to keep the voltage fairly constant over one commutation cycle which is typically a couple hundred microseconds or so.
The total energy that can be extracted from braking can be calculated from mass and velocity. Here is a quick calculator:
http://easycalculation.com/physics/c...tic-energy.php
I used 30 m/s and 2000 kG and it is 900,000 Joules (W-Sec) or 0.25 kW-Hr. If you have a 144V system the capacitor needed to store this energy would be 86 F. A 3000F 2.7V capacitor is about $100:
http://www.newark.com/cornell-dubili...ohm/dp/92T9599
You would need 50 of these to get close to 144V and the capacitance would be 60F. Pretty close, but you pay $5000 to recoup 2.5 cents worth of electricity, or to get maybe an extra 1/2 mile of range.
There are some examples of KERS, including flywheel, at:
http://en.wikipedia.org/wiki/Regenerative_brake
Please check my calculations, but they seem about right.