You don't need to sense pedal torque, the driven wheel either has
enough traction or it slips. When it slips you have a speed difference
between the wheels that you can sense and correct. You don't need to
accurately measure the difference, just null it out.

The All Wheel Drive on my car detects a difference between front and
rear axle speed and connects power to the rear axle proportionally to
the difference. Each axle drives a pump that moves hydraulic fluid in
a circle. A speed difference lets pressure build up and operate a
clutch cylinder. There's some allowance for cornering and driving in
reverse but it's a simple concept that works very well.

For an electronic version you could sense wheel speeds as DC voltages,
subtract the main and slave voltages and power the slave wheels
proportionally to the difference, with a small dead band. You'd need
PM motors etc for tachs to sense speed, a low-gain op amp to compare
them, and an H bridge or PWM controller for the slave wheel motors.
And a kill or dead-man switch.

I love the irony of printing KILL KILL KILL KILL on the label maker
when I'm assembling *safety* equipment.

My truck has an all-mechanical 4WD with no center differential, both
axles turn at the same speed. In very slippery conditions both work
well, any difference is masked by the different weight distributions.
In marginal conditions the AWD slips a bit before the rear axle
engages but I don't feel it, only hear the tire noise. There is no
feedback to the steering wheel. At speed I think stable steering is
more important than maximum traction, having owned a Civic that
oversteered on ice.

Ever ride a dirt bike on a slippery surface? If the rear drive wheel
can slip you don't have much steering traction either. When you brake
enough that the front wheel slips you go DOWN.

A three-wheeler with an electric motor driving a rear differential
might be easy to do and work well. Separate wheel motor drives would
give you dangerous steering problems at more than lawn tractor speed.
Once you go fast enough for wheel hop things can become complicated.

ABS is really simple until you add on all the fail-safe circuits. I
built the test stations for the first analog electronic ones in the
70's.

jsw