The example I’m goigng to try to describe happens often around the level and I’m unsure as to why. I’ve been tweaking the RichAI settings for a few weeks to try to get more ‘life-like’ behaviour and in general it’s worked, bar this issue.
The agent will walk past a corner/ nextPathPoint and just after passing it, the RichAI system will re-path again, updating but using the previous point just passed, thus the agent turns and walks back (see images). This cycle happens anywhere between 1 - 4/5 times, sometimes even more. In the past the agents could get stuck indefinitely, but the adjustments have remedied that.
I’ve tried many variants of the PathAI/ RichAI settings. The image below shows the current config.