To get braking, you need to switch all the low side FETs on at the same time. This momentarily shorts the windings in the motor. When the switches open, the voltage kickback will circulate back to the batteries and charge them.
Then you use a PWM to control the amount of braking.
In order to do this, there would need to be a PWM for the braking and a way to re-route the gate drivers. This would be all low power logic stuff, so the parts wouldn't be too expensive.
Another possible way to do it would be to build a separate braking controller that has 3 banks of FETs and a PWM. You would tie into the 3 motor phase wires. You would also need some kind of interlock to prevent braking and acceleration at the same time. This way you wouldn't need to modify the controller, but you would need big $ FETs.
To control the braking, you would need some kind of control to activate it. A throttle that goes both directions from neutral would be cool. Sort of like an RC transmitter stick.
On my Zappy, I had a separate current limiter for the braking set it at a fixed level. This worked great. But that controller was already made for braking.
Another way, that might be easier, is to use a brake switch in the lever to activate the braking. That's pretty much how it's done on my Honda hybrid. When you hit the brakes, the regen kicks in. The amount of braking could ramp up so you don't get a sudden wheel lock.
It might be a good idea to have a circuit to disable the braking if the battery voltage gets too high. This was a problem on my Zappy if I started out on the top of a big hill with a full charge.
The really crude way to do it would be to use relays to switch some really big resistors across the phase wires. No regen, but serious braking.
The problem with this is the braking force will increase with speed. Not so good when you're going full speed down a big hill. I tried this on my Zappy a long time ago. I caught the resistor on fire one time.
:
