Decided on spending more time this winter in trying to develop a system to aid drivers (mainly myself at first) get faster around (firstly) an auto-slalom course (and maybe later a race track). Hence the entrieds I made relating to motion tracking and GPS positioning.

Not sure yet how this will finally work, but I guess it’s all about the journey…

Some general thoughts so far:
GPS is quite inaccurate. 100m error is too much.
Differential GPS is A LOT better at around 3-5m, but still not enough. I need accuracy of at most a car width (that might still be too much).
GPS is also only updated once a second… need many more updates then that.
So something else needs to be used to help out.
Accelerometers are also not accurate when doing a double integral (for distance), but apparently they update very quickly (talking in the 50Hz to 200+Hz).

Due to the GPS inaccuracies, I’m willing to not even use it.
Maybe a combination of accelerometers, electronic compass and vehicle wheel speed?

The video tracking option is also not very good. From my tests so far, results are not promissing. Tracking is not too reliable, and requires LOADS of computing power!
If I can find a way to simplify, I might consider adding it to the list of aids.

The other problem is finding the location of all the cones. This one seems like a tougher nut to crack, but without it, the above part might be useless.