@postman09 Thanks for the idea! Thankfully, our current curves have a lot of data points, so the errors don’t seem highly noticeable. (But I’m not exactly sure what the errors would manifest as.) That said, the method you suggested will be added to the to-do list.
My current progress with distance matching is shown in the below video. The start transitions aren’t terrible. But, as you’ll see, the stop transitions are non-existent. The red/blue debug sphere show predicted stop location, but the error increases as character speed increases.
(Sorry if the video is loud. I’m still using the ALSv4 test level.)
