From de140127b8501eebc5dc54a1f24f98f9b1ae0dba Mon Sep 17 00:00:00 2001 From: m-guberina <gubi.guberina@gmail.com> Date: Tue, 21 Nov 2023 00:34:09 +0100 Subject: [PATCH] wrote commetns on what to try --- python/examples/fixes_for_demo.md | 13 +++++++++++++ 1 file changed, 13 insertions(+) create mode 100644 python/examples/fixes_for_demo.md diff --git a/python/examples/fixes_for_demo.md b/python/examples/fixes_for_demo.md new file mode 100644 index 0000000..aa3489d --- /dev/null +++ b/python/examples/fixes_for_demo.md @@ -0,0 +1,13 @@ +1. i'm pretty sure why it goes so far with movej is because you gave it a tight error. + well, give it a lax error then. atm there is a single instance where you're running this command, + so it's fine, you cares. bottom line is, be above the starting point, and then go down with + dmp which has the force feedback and will ease into it. +2. throw the low-pass filter away, rock the moving average, that's what worked. +3. if it will give you trouble, just remove it. write on the bord, that's precise enough. +4. put in new the new translation and rot mat into old code. + replace all (BUT PATH) 2's with 1's to go from the z-axis into the y-axis. + +one of these fixes the problem instantly. +fuck the tuning for the most part, it's mostly ok. +feel free to play around with it AFTER IT WORKS. +but whatever is in there now is not what makes the difference. -- GitLab