ECR Bolt-On: Training students to make robots that learn
Current autonomous robots rely on human-engineered solutions that do not practically scale into industry solutions. The approaches are too slow, inflexible and difficult to transfer between different tasks. Solving this is the focus of one of SFTI’s new ‘bolt-on’ projects which employ Early Career Researchers (ECRs) to take on key pieces of work linked to one of SfTI’s existing projects.
One of SfTI’s major Spearhead projects involved building robots that can work safely in the field. Robots working in forestry need new abilities that allow them to perceive what is happening around them and use that information to learn and adapt, even when the ground is uneven or covered in debris, or when the object the robot must collect is fragile or oddly shaped.
While working on this project, the researchers realised the growing demand for skilled robotics engineers from Aotearoa New Zealand. The problems raised by rough ground, delicate fruit, or forest slash require intelligent robotic systems that are adaptable and capable of learning complex tasks quickly and effectively.
Image: Intelligent robots need to be able to grip. Here you can see the 'robot gripper' in development.
Aotearoa needs a pipeline for early career researchers to obtain the research skills in robotics and automation that are in demand in the industry. In particular, these people are needed to help solve a wide range of automation needs from the primary industry sector as well as creating interconnectivity between autonomous manufacturing equipment and broader computer systems. SfTI is therefore funding a state-of-the-art research programme in intelligent robotics, supported by current PhD students across different universities who will gain leadership and mentoring experience. This ‘bolt-on’ project, is one of a set SfTI has designed to create an opportunity for early career researchers to develop and lead themselves while focussing on a key area that complements one of SfTI’s larger projects.
This particular bolt-on project will provide summer research projects and internships for undergraduate engineering students. The students will work on computational models that allow robots to autonomously navigate unexpected conditions and use multiple ‘fingers’ to grasp and manipulate objects. This way of applying learning techniques to robotics systems is still in its infancy and the project aims to develop a collaborative open-source platform that will place these researchers among global leaders in the industry.
“It was exactly this kind of opportunity that led me into the career that I’m in now, where I get to play with, I mean research, autonomous robotic systems across a wide variety of projects. I want to encourage as many people as possible into post-graduate studies in this area,” said Dr Henry Williams, the project leader.
Dr Henry Williams, is an early career researcher with extensive experience managing large collaborative robotics projects. He is joined by PhD students David Valencia (Ecuador) and Hoda Yamani (Iran) and the team will be mentored by Dr Jaco Fourie, Dr Armin Werner. The project will work with Women in Engineering and the South Pacific Indigenous Engineering Students organisation to recruit Māori and Pacific peoples students for the project.
This project is funded for $150,000 to 2024.