The least squares average is for determining the speed of a vehicle. I have a 12 bit binary counter which has its coutn latched at 400 Hz. I will have about 400 to 800 samples to use for the algorithm. I'm running the counter at 150 Mhz and frequency doubling it by latching the output of the xtal itself. The 400+ samples that I will have will only differ by only 4 or 5 counts due to the high speed nature its operating at. So I'm looking for a fit of 2 dimensional data that yields a gradient. I would appreciate if anyone already has an algorithm that does this or is similar to what I want.


cheers

Ryan