ROS

Development of an adaptive robotic assembly approach for sockets and connectors based on visual and force-based methods. First evaluated in a peg-in-hole scenario, then in more complex real-world scenarios at TTZ Nürnberger Land.

Development of an approach for separating limp cables from a bin: localising connectors, determining grasp poses, handling mutual occlusions and tracking the connector pose during handling.

Development of an approach for detecting unknown components and computing their grasp poses using RGB/depth cameras and neural networks. Validation on a robot cell with a collaborative robot at TTZ Nürnberger Land.

Development of a robotic concept for grasping, unfolding and placing paper and plastic-film interlayers into transport boxes. Validation on robot cells with collaborative robots at TTZ Nürnberger Land.

The drone used at the AerodrOHM currently localizes via external markers. The goal is a markerless, GNSS-free indoor localization using only onboard sensors (visual-inertial odometry / SLAM), building on the existing platform.

Game-based human-robot interaction with the UR12e at TTZ Nürnberger Land: Tic-Tac-Toe with camera-based board detection, game logic and safe robot control as an approachable robotics demo — optionally extendable to chess.

Analysis, technical revision, and recommissioning of a mobile C-arm robot with mecanum drive, including possible mechanical modifications and validation of system and driving behaviour.

Building a 2R robot arm with BLDC motors as an M-IAS lab experiment: model-based gravity compensation and reinforcement learning demo with ROS integration and simulation support.

Development of a system for collision-free navigation of Crazyflie drone swarms using an external 3D camera, ROS integration and reactive swarm behaviour. Testing in the AerodrOHM flight space.