Nicolas Nabrink

Co-authored a paper extending the thesis work to uncertainty-aware pose estimation, submitted to IROS 2026 (currently under review)

Title: Synthetic Data-Driven Machine Learning for Object Pose Estimation in Robotic Manufacturing (Grade: 1.0) - Designed and implemented a synthetic data pipeline in NVIDIA Isaac Sim to generate photorealistic industrial scenes with ground-truth 6D poses in BOP format for controlled robustness studies - Fine-tuned a state-of-the-art pose estimation model (SurfEmb) and executed sim-to-real validation across synthetic benchmarks and a real industrial dataset - Deployed the pose estimation pipeline for infere

- Developed software tests in Python for Advanced Driver Assistance Systems (ADAS), automating validation workflows - Built internal Python tools and GUIs for test execution and reporting - Worked with Hardware-in-the-Loop setups to validate Electronic Control Unit functions

Research Title: Shared Control for Teleoperation of Automated Vehicles in Dynamic Environments (Grade: 1.7) - Designed an MPC safety controller for EDGAR teleoperation in dynamic environments - Implemented time-to-collision logic and adaptive speed control for critical scenarios - Developed the controller in Python with ACADOS and integrated it in C++ with ROS2 - Validated the end-to-end control stack in CARLA simulation, demonstrating improved safety and performance in complex scenarios

- Contributed to the development of a unilateral knee exoskeleton for hemiparetic patients, with a focus on gait assistance and rehabilitation-oriented feedback - Designed and implemented a vibrotactile biofeedback concept to improve user awareness and interaction with the assistive device - Developed embedded sensing and data acquisition on ESP32 using Arduino/C, integrating IMU signals for motion analysis - Built a MATLAB-based GUI for data visualization and post-processing to support experimental analysi

- Worked in the Mechatronics Core of the USP Junior Consulting Firm, supporting end-to-end delivery of customer-driven engineering projects - Led project conception and technical scoping, translating customer ideas into feasible mechanical and electrical solution concepts - Developed detailed CAD models in SolidWorks and built functional prototypes for validation and demonstration - Coordinated client communication, proposal alignment, and execution planning to connect commercial requirements with technical

Student Council at Highschool. In charge of creating cultural and sports events for more than 1000 students.

Master Thesis: Synthetic Data-Driven Machine Learning for Object Pose Estimation in Robotic Manufacturing (Grade: 1.0) - Paper submitted to IROS 2026 - Focus on AI-driven robotics applications, including perception, control and real-world deployment.