My Expertise
Intelligent robotics; Autonomous systems; Artificial intelligence; Machine learning; Robot teams
Keywords
Fields of Research (FoR)
Artificial intelligence, Machine learning, Coding, information theory and compression, Manufacturing processes and technologies (excl. textiles), Assistive robots and technologySEO tags
Biography
Oliver Obst is an Associate Professor in Robotics in the School of Computer Science and Engineering at UNSW Sydney. His research focuses on embodied intelligence and collective robotics, with an emphasis on efficient learning systems, recurrent neural networks, and autonomous platforms operating in real-world environments.
He develops methods for learning and decision-making under uncertainty. His work covers multi-robot coordination,...view more
Oliver Obst is an Associate Professor in Robotics in the School of Computer Science and Engineering at UNSW Sydney. His research focuses on embodied intelligence and collective robotics, with an emphasis on efficient learning systems, recurrent neural networks, and autonomous platforms operating in real-world environments.
He develops methods for learning and decision-making under uncertainty. His work covers multi-robot coordination, decentralised perception, and long-horizon autonomy, with applications in health and intelligent infrastructure.
He has held senior academic leadership roles at Western Sydney University, including Director of a research centre and Head of Department, and previously worked at CSIRO/Data61 as a Research Team Leader. He served on the board of trustees of the RoboCup Federation and mentors student robotics teams. He holds a PhD in Artificial Intelligence from the University of Koblenz–Landau, Germany.
My Grants
Digital twin for the manufacture of composite structures by resin infusion (ARC DP260103722)
2026–2029. ARC Discovery Project (administered by Western Sydney University). Lead CI: Prof Brian Falzon (WSU). CIs: A/Prof Oliver Obst (UNSW) and Prof Liyong Tong (University of Sydney). This project develops a digital twin framework for resin infusion, integrating high-fidelity modelling with machine learning to predict defects in real time and support adaptive process control. The goal is improved manufacturing reliability and throughput with reduced material waste, with relevance to lightweight composite structures used in aerospace and renewable energy.
My Research Activities
We explore how robots can learn, plan, and act effectively in the physical world, individually and collectively.
Our research focuses on:
- data-efficient embodied learning and manipulation
- collective and multi-robot systems operating under uncertainty
- learning and inference methods for physical systems
We are interested in approaches that combine learning, probabilistic reasoning, and real-world experimentation, with an emphasis on efficiency, robustness, and physical grounding.
My Research Supervision
Areas of supervision
- robot learning and embodied intelligence in real-world environments
- collective robotics and robot teams (coordination and collaboration)
- learning and decision-making under uncertainty (Bayesian and belief-based methods)
- reinforcement learning and planning for long-horizon autonomy
- decentralised perception, sensor fusion, and active perception
- recurrent models for control and perception (sequence models and memory)
- robotics with applications in logistics, horticulture, health care
- digital twins and physics-informed machine learning for manufacturing processes (composites/resin infusion)
Contact
Publications
ORCID as entered in ROS
https://orcid.org/0000-0002-8284-2062