Rasmus Haugaard

raha@raha.dk raha.dk scholar.raha.dk

Profile

PhD and former Assistant Professor with work at the intersection of Computer Vision, Deep Learning and Robotics with a focus on modeling robot-vision problems as well-posed learning problems towards more reliable robot systems. First-author publications in top-tier venues (CVPR, ICCVW, CORL, ICRA, IROS).

Experience

2023 - 2025

Assistant Professor / Postdoctoral Researcher, SDU Robotics

  • Research on pose estimation from tactile feedback for easy robot cell fixture calibration with uncertainty quantification
  • Worked on an autonomous learning pipeline for trash sorting
  • Applied methods developed during the PhD in internal projects
  • Taught software development and deep learning courses
  • Supervised bachelor/master projects and other student projects
2019 - 2023

Ph.D. in Object Pose Estimation, SDU Robotics

  • Research on object pose estimation, including pose estimation from correspondence distributions, extending that approach to multi-view, and estimating pose distributions - an expressive pose knowledge representation which integrates well with probabilistic frameworks and sensor-fusion
  • Research on visual servoing for peg-in-hole, screw- and pcb component insertion, including a sim2real- and an autonomous data collection pipeline
  • CV and autonomous cable management for the 2021 IROS challenge
  • Assisted teaching numerical methods & deep learning
  • Change of environment at the Technical University of Denmark (DTU)
2015 - 2021

BSc & MSc in Engineering (Robot Systems), SDU

  • Master project (2021) on establishing multiple correspondences in face of ambiguities for object pose estimation
  • Bachelor project (2018) on training object detectors from synthetic data
  • Exchange at Norwegian University of Science and Technology (NTNU)
  • Extracurricular course in entrepreneurship

Skills