Topic: Autopilot Could Be Better
I realize that the reality of Dev priorities means that this will never get looked at, but it is worth mentioning.
I'm sure many consider the Autopilot function a God send. And it is very useful. The problem is that is uses a rudimentary "follow the wall" algorithm that is oblivious to the high speed highways. It is so oblivious, that frequently, your bot will be jogging along parallel to the high speed roadway and NOT utilizing it. Also, by blindly using a "follow the wall" algorithm, it turns into every little pocket along the way, rather than sensibly maintaining a straight line toward the objective. In short, it can't establish efficient pathways from start to end points.
So, if there is ever any Dev time, the autopilot path finding algorithm needs, at the very least, to be aware of high speed roads and use them. Also, it needs a better algorithm to keep the bot in the middle of roadways and/or splitting the difference between obstacles as compared to skirting the perimeter of each obstacle the bot encounters.
Oh well. I guess I should be glad there is ANY autopilot function at all. I do remember when it didn't work or was very very flaky (way before Gamma destruction). It certainly has improved since then and I'm using it as is. But, it could be better.