Hm. Certainly low speed anti cogging must be inherently easier. I would recommend looking around on google scholar first, to try to see what others have published. There are some guys on Hackaday that have published approaches. I think fundamentally, you want to understand the physics of the motor to one or two levels of approximation, and then measure and model things. If you just measure the torque ripple and compensate you may run into trouble at different current levels etc. and create more problems than you solve.
I think if you can build a primitive mathematical model in python or even a spreadsheet of a motor which takes various inputs and tells you the torque vs. rotational position, then you may find that the inputs to the model can be measured, and a reasonably good compensation system could then be built, which may even work at higher speeds. You will get more help if you are working on something that more people have a use for, so this approach may be less work than it first appears.
It’s probably advisable to set up your system and experiment with things so you can see ways forward and learn hands-on. I often find this opens doors I did not expect.
What is your “standard driver”? This is just the power stage or what?
BTW you seem to be possibly mixing up the influence of cogging on angular position with the influence on torque. At low speed, the angle control mode of SimpleFOC, if used with an actual optical encoder (might not work with a magnetic angle sensor due to calibration issues that remain unsolved and are in turn influenced by cogging with the current approach), would largely eliminate the apparent influence of cogging on angular position. However it will still influence the torque, so if you want force feedback then yes anti-cogging may be valuable.
I think sooner or later you will need to measure torque. You could do this indirectly using the inertia and observed acceleration of a free spinning motor, however that could get complicated. Some kind of sensor seems warranted, idk what the options are there…