Probably is round-trip pulse timing.
From another forum, a clever trick for measuring very short pulses ( the period between pulse initiation and echo recieved ) easily and cheaply is to quickly charge up a cap during that time. You have then translated an incredibly short pulse into a very usable voltage. Assuming 1ns for 1 foot, a cap which charges to full over 1ns and a 10-bit ADC, you can measure that foot to near 0.001 foot accuracy, equivalent to 1000GHz.
I'm sure it's not as simple in practice, but the principle is simple enough.