I’m building a two-segment planar continuum robot in drake. I already write a sdf file with multiple joints working as nodes for continuum robot. Theoretically, these joints will only controlled by the kinematic model (piecewise constant curvature).

I’m working with IRIS-GCS motion planning for collision avoidance and trajectory optimization in a complex environment (specifically the vascular). The environment models are highly complex and non-convex. I found the original method for IRIS is to create collision geometry for obstacles in separate sdf file, such as the half_wall_vertical.sdf

I have a theoretical question:

I have a graph of convex sets where I am solving for the optimal path through multiple regions. I am working on improving the costs between each vertex (polygon-shaped regions). The cost to travel along the edge from vertex u to vertex v should be the distance from u to v multiplied by a scaling factor for traveling through that vectex. (The longer the edge travels through the vertex, the more effect that vertex will have on the cost)

I want to solve a non-penetration projection NLP. I have a plant with a bunch of floating bodies and I want to only selectively optimize certain parts of the object positions (e.g. translation only or only the z-component of the translation).

How can I use RgbdSensor for rendering on a server? Can I somehow use a virtual display?

What is the easiest way to use a custom fork of Drake in a Python venv?

I have a fixed SDFormat file, containing one link and representing a rigid body. How can I programmatically scale the resulting RigidBody? I’d prefer to be able to do this after plant finalization but doing this during the plant construction phase would already be helpful.

I tried to visualize the pydrake iiwa manipualtion station with the following code but failed with error ‘RuntimeError: DiagramBuilder does not contain a subsystem named plant’. How to make the visualization work in this case?

I’m a student interested in running the code from the Robotic Manipulation course locally on my Macbook, as opposed to with Deepnote. My machine is using an Apple M3 chip with macOS Sonoma 14.5, and I ran the following commands for installation: