ROS 2 control driver for KUKA LBRs
A Fast Robot Interface (FRI)-based ROS 2 control driver for KUKA LBRs (IIWA / Med) with compatibility across various FRI SDK versions.
Martin Huber
PhD, robotics software & machine learning engineer combining classical control with learned policies for robust, real-time execution on autonomous robotic systems — with a focus on computer vision and imitation learning.
Projects
Imitation learning-based velocity policy for NMPC pattern generation
An imitation-learned policy trained to search for fire extinguishers, generating planar translational and rotational velocities for a nonlinear MPC pattern generator. Deployed as a JIT-scripted model via the PyTorch C++ (LibTorch) SDK.