Replies: 1 comment
-
|
I'm wondering if the issue is that I'm using the default object material, and once the fingers start rubbing against the surface, the forces can grow very high, preventing the robot from moving and causing it to stall or freeze. For now, I’ll try simulating the contact forces by keeping a small offset above the surface and using an internal function that takes the distance from the object’s curvature and returns a simulated force, instead of relying on Genesis’s force readings. This way, I can track the object’s curvature without needing to accurately model the exact material properties. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
I’m trying to simulate a robot arm tracking a surface by using force feedback from its fingers. If the force increases, the arm lifts slightly to reduce the force; if the force is too low, it moves down a bit. This way, I can track a surface precisely without relying on complex image processing to estimate the exact surface topology.
Humans do this all the time using force sensing, for example when cleaning and scrubbing a surface.
So far, I haven’t been able to achieve smooth tracking. The robot arm jumps up too much, and the motion is not smooth.
Here are some images from a recording I made:


Here the robot hand is jumping up trying to reposition:
Here is the graph of Z and the detected forces on finger 1 and 2:

The travel surface is a wedge with the ramp going up the farther away from the robot base.
The full recording can be found here:
https://github.com/lovecode1/Genesis-Tracking-Surface-Issue/raw/refs/heads/main/force_contact_travel.mp4
Here is the code I have so far if you want to take a look:
https://github.com/lovecode1/Genesis-Tracking-Surface-Issue
I guess the Inverse kinematics algorithm returns new joint angles that cause the robot arm to move away too much instead of tracking the surface.
How accurate is the simulation of forces on the robot fingers?
How noisy are the readings?
How this can be improved?
Beta Was this translation helpful? Give feedback.
All reactions