In his doctoral dissertation Kulunu Samarawickrama studied how robots can better understand and carry out multi-step assembly tasks, where several parts must be recognized, related to each other, and placed in the correct position. in his research, he developed simulation-based datasets, assembly pose estimation methods, and a learning-based framework that links perception, task-aware reasoning, and actionable outputs for robotic assembly. The key result of his research is a structured approach to “assembly cognition,” enabling robots to move beyond recognizing individual objects toward understanding how parts form a meaningful assembly. This research can benefit robotics researchers and automation developers, as well as future industrial systems that require more flexible, reliable, and intelligent robotic assembly.
The doctoral dissertation of MSc (Tech) Kulunu Samarawickrama in the field of Automation Science and Engineering titled Deep Learning Frameworks for 3D Perception and Assembly Cognition in Robotic Manipulation will be publicly examined at the Faculty of Engineering and Natural Sciences at Tampere University on 26 June 2026.
The Opponents will be Professor Mårten Björkman from the KTH Royal Institute of Technology in Sweden and Associate Professor Jens Lundell from the University of Turku. The Custos will be Professor Roel Pieters from the Tampere University.