search

CMM-22 / a2

CMM Assignment 2 - Kinematic walking controller
1 stars 0 forks source link

Ex2 Robots bases snaps lower at start of animation. Is this intended? #31

Closed dtetruash closed 2 years ago

dtetruash commented 2 years ago

I have implemented an IK solver which seems to work. But on second viewing I found that the walking cycle is a bit off from that which is shown in the readme. Namely, the robot seems to be "crouching" more - the base is lower in the y direction. See the screenshot comparison below.

I am assuming this is not intended, and I have a bug. Is this so? Do I need to invest more time into this?

REFERENCE: image

MINE: image

dtetruash commented 2 years ago

Specifically, on initialization, the robot looks like so: image

One frame into the animation, we have: image

My assumption is that I am moving the base to it's COM?

dtetruash commented 2 years ago

And the same without mesh rendering:

image

image

Sibsoon commented 2 years ago

As from what I found in the code it is intentional, since the start position is defined somewhere in the robot configuration, where the position during animation is defined by the groundHeight and the trunk height.

eastskykang commented 2 years ago

Exactly. At the very beginning, the base position is a bit higher than the target. (I did it deliberately because in the initial configuration of the robot, the feet are under the ground which can give wrong impression to students.) So, this is very normal. You don't have to worry about it.

Remark. (As I mentioned during the tutorial) In fact, you don't need to do anything about base target, because we assume the base tracks the target perfectly. It's already handled by the skeleton code.

dtetruash commented 2 years ago

Awesome! I assumed that I was accidentally updating the base position unknowingly. Moving right along.