As Austria's largest research and technology organisation for applied research, we are dedicated to make substantial contributions to solving the major challenges of our time, climate change and digitalisation. To achieve our goals, we rely on our specific research, development and technology competencies, which are the basis of our commitment to excellence in all areas. With our open culture of innovation and our motivated, international teams, we are working to position AIT as Austria's leading research institution at the highest international level and to make a positive contribution to the economy and society.

Our Center for Vision, Automation & Control located in Vienna invites applications for a master’s thesis. The Center for Vision, Automation & Control exploits the opportunities provided by automation and digitalisation to initiate and advance innovation for industry, primarily in Austria and Europe.

For this master’s thesis you will join our Competence Unit Complex Dynamical Systems, which focuses on the development and deployment of algorithms to control various types of systems - starting from low-energy applications like electronics and drive systems, to (large-scale) robotics and heavy industrial applications. The automation of work machines, such as cranes and forklift trucks, is a strategic research goal within our unit. In the future, these machines will take over repetitive and dangerous tasks. The aim is twofold: firstly, to relieve the burden on employees and enhance their role, as they will be responsible for planning and monitoring in the future, and secondly, to counteract the shortage of skilled workers. To enable automation, many complex issues need to be investigated, ranging from environment recognition and interpretation to machine control and human-machine interaction. For validation and testing purposes, AIT has set up its own open-air test site for autonomous machines in Seibersdorf.

• Forestry cranes operate in highly unstructured environments where logs must be grasped individually or in groups under varying conditions. Current semi-automated systems rely on simple heuristics and operator input, limiting throughput and increasing operator fatigue. To overcome these limitations, this thesis focuses on developing and evaluating reinforcement learning (RL) based control policies for multi-log grasping on a forestry crane, trained in a physics-based simulation environment.
• Under the guidance of our researchers, you will design and implement RL policies with appropriate reward shaping, curriculum learning, and domain randomization strategies to generalize across varied log configurations.

• You will benchmark your approach against heuristic baseline methods, evaluating grasping success rate, cycle time, and robustness to disturbances.
• You will analyze sim-to-real transfer challenges and investigate mitigation strategies for deployment on real forestry crane hardware.
• You will broaden your knowledge and skills in the fields of reinforcement learning, robot simulation, and learning-based control for robotic manipulation.
• Optionally, you may present your results at a top robotics conference (e.g., IROS or ICRA).