hanzeracingdivision / driverless Goto Github PK

When track mode with cublic splines is attempted, some of the passed targets will become unpassed spuriously.

This is because it causes a mismatch between the length of the non_passed_targets list and the non_passed_dists list. If the two lists differ in length, the code attempts to find the argmin of an empty list and the code breaks.

Will try added a clause to the for loop (where the argmin occurs) such that a prerequisite condition to continue is the two lists have the same length.

Or It might be better to simply add an if statement where if the length of the first list = 0, it manually empties out the other list.

When beta (raw angle to cone) is in (0,90) and the car angle is in (-90,-180), the alpha computed is incorrect and the car drives off

this also happens when beta is in (0,-90) and the car angle is in (90,180) .

Tempted to just fix with an if-then statement but it will be less efficient than doing it properly

once all targets are passed, in track mode the first 'lap' is complete and all targets are set to 'unpassed'. Therefore if the target behind the car is closest the car will aim for that one, which is clearly the wrong choice.

possible fixes:
-include the direction of the targets when choosing next target.
-have the previous cones not change to unpassed after lap (doubtful about this method)