Interesting topic for me, I spent the first 13 years of my engineering career as a stability control calibrator for Bosch and GM working on projects from pickup trucks to sports cars. SO, to condense 13 years of intimate knowledge of every line of code into a forum reply that's useful is a bit challenging! I'm trying really hard not to come off as a know-it-all but on this
specific topic I kind of am
1. Calibration of the system to a new vehicle is a full time 40hr/wk job that takes roughly 1.5yrs for a new vehicle. Maybe half or 3/4 of that is stuff you could skip on a hobby car and still get good performance but, still...
2. This calibration is extremely sensitive to the specific handling of the vehicle. The 'reactive' idea presented in an earlier post here is definitely correct but the magic is knowing the
target yaw rate to react to. i.e. how that steering wheel input you measured *should* translate to yaw when a corner is going just fine (no noticeable over/understeer). You need a yaw
target to compare too and any car in any corner is technically ALWAYS over or understeering (that's how tires work) and every car is VERY different. You have to drive thousands of situations over and over again to make that model work. (the model being: for a given speed, surface grip, bank angle of the road, steering angle, steering angle rate, brake or throttle apply - how much *should* the car yaw if everything is going well) You need that number which is constantly calculated every 5 milliseconds to know when you have too much or too little yaw in any of those situations that needs to be 'fixed' with brakes.
3. The cal is also extremely sensitive to the layout of the hydraulic system. Once it sees a yaw that is different from that model it needs to know how much fluid to pump into the brakes to get a result. This relies on a good ABS calibration which is a whole other ball of wax that has to be done first!
4. Putting a system calibrated for ANY street car (min weight for a street car equipped with ESC is maybe a 2200lb Miata??) onto a <1500lb locost with a pieced together steering system, suspension, wheel tire package, and brake system is almost certainly going to result in a system that either doesn't respond to oversteer at all or locks wheels randomly when it just *thinks* the car is oversteering because the model is wrong (the more likely scenario, IMHO).
5. To emphasize this sensitivity it might be surprising to know that almost every trim level of the same make and model car you buy has a unique ESC calibration! Even on a Miata the sport suspension car will have a separate cal from the touring suspension, the big brake car a separate cal from the small brake car, etc. The OEMs don't spend this money for fun, there are real results that are unacceptable in normal driving if you don't do it.
SO, I'm not saying not to try, I love a good experiment, but I am saying it's a HUGE crap shoot if it works or not and I can 100% guarantee that it won't work *well.* (also, insert safety warning here, randomly locked front wheels at high speed are bad and likely)
Modifying the calibration is a tempting idea but there are >1200 parameters to calibrate in hundreds of thousands of lines of code (which is why it takes an experienced pro 1.5 yrs to do it). It's just not feasible to calibrate for any reasonable cost, trust me, I've pushed on the aftermarket option for a long time and the cost is just overwhelming to even do it to a hobby car level.
Also, you have to crack the controller (don't think anyone's done this to an ESC controller yet) and learn the code first, takes maybe 3 years for a professional to learn it with all the documentation you get working at a supplier. I can't imagine making sense of a set of cracked cal tables with no documentation.
Someday maybe a supplier will do a dumbed down version with a few dozen easily understood cal tables that can work as a safety backup (I think that's possible with some engineering effort up front) but it doesn't exist right at the moment.
Hope this was helpfull.
Alex