>>1501Theory is cheap.
Testing is expensive.
Testing anything at 1000mph is also probably expensive.
I was going to say use accelerometer and integrate but
>For instance, we can integrate the output from a longitudinal accelerometer, but experience with Thrust SSC showed that small errors can accumulate so the results from this method are never better than approximate. Fuck.
Broadcast a signal from the nose of the car, measure it at the rear. Ask Doppler what he thinks. Boom, you have a velocity.