Professor Denny Oetomo is a roboticist and mechanical and mechatronics engineer in the Department of Mechanical Engineering at the University of Melbourne. He is co-director with Professor Ying Tan of the University’s Human Robotics Laboratory, with research interests in robot design and manipulation strategies, especially where robots work with human users.
People often have a mistaken idea about robotics. They picture the clumsy giants in cages on the production line, welding cars on a conveyer in the factory, or they picture futuristic machines from sci-fi movies. For me, the point of robotics is entirely practical – it is to get a performance that would otherwise be unachievable by either the human or the robot. My research focuses on the physical interaction of humans and robots: how do you get a robot to move with humans and deal with unstructured environments?
Curiosity got me into research. I always liked robotics as a child, which may suggest that I probably never quite grew up and my toys simply became more expensive. But I want to make useful contributions for everyday applications. It’s what engineers do: they sit at the interface of science and practical problems. They find solutions. And I like to tackle new problems that I find exciting and interesting.
One of the programs I have worked on for several years is sheep shearing. In the 1980s, a sheep-shearing robot put Australian robotics on the international map. It was a fantastic achievement because it’s very difficult. A sheepskin is not a nice, even surface. There’s no perfect pattern for a sheep: this one has more wool here, that one has more wool there. Unfortunately, it did not find commercial success because of some important practical issues – for example, it took too long to strap a sheep into position.
So, working with Australian Wool Innovation, we took a different approach. We noticed that many shearers after four years have some sort of permanent injury because it really is a back-breaking labour. Shearers know they risk injury when they tire, but because they are paid per sheep, they just keep going. So, we have developed a set of sensors that shearers wear on their body to monitor their body posture and muscle activity while shearing.
The study for the past three years focused on sensor placements – which muscles to measure, what posture to pay attention to – so as to gain data about the shearer’s level of fatigue, posture and, hence, risk of injury. This resulted in a simplified system using very few sensors on the body, which can be commercialised for many shearers to wear. This part of the work is complete, has been published and is being tested across different sheep-shearing stations. The AWI is keen to improve the wellbeing of the shearers and to prolong their careers, so the measurements on the body will be used in the project’s next stage, to design and drive some physically assistive robots to help support the shearers during shearing.
Another project that has just become a commercial reality is our collaboration with Fourier Intelligence, a robotic manufacturer. It produces robots for hospitals to help with the rehabilitation of stroke patients. Hospitals have a shortage of clinicians when compared with the therapies the community needs, so we designed robots to specifically target things that benefit patients and clinicians, allowing the latter to see far more patients. Unlike a doctor, a robot can be with a patient all day, and provide far more precise measurements. Instead of checking progress every three months – when the patient visits the clinician – the robot measures whenever the patient is accessing the robot for regular therapy: how strong and how fast is the patient becoming, has the patient a better range of motion? And the patient can spend many more hours a day in monitored therapy.
In September 2021, we produced our first commercialised robot, licensed to Fourier Intelligence. We also worked very closely with the Royal Melbourne Hospital, which provided expertise in that domain. Working with the hospital since 2010, we have become one team and we understand each other’s language, which is very important. At the start, we didn’t understand the medical part and they didn’t understand robotics, which was a massive issue. Now, although disrupted by COVID-19, we have started teaching a subject together (the clinical team and the engineering team), where we jointly run the class in a panel format.
Both these projects capture nicely what human-robot interaction is about. By combining humans and robots, we find we can get a level of performance unachievable by either alone.
We need industry partners. At the University, we are not manufacturers or a commercial entity: our business is teaching and research. So, if we want our robots to be used, we need manufacturers involved, as well as clinicians and end users. I enjoy seeing what I work on being useful in the community. When I started with a therapeutic robot, it was a research prototype, with all the maintenance issues that involves, ensuring all the codes are properly maintained and that we can have access to replacement parts. Once a robot becomes a commercial venture, there needs to be sustained supply and maintenance support by a commercial entity. I am no longer faced with the traditional research robot problem of, ‘Oh, my God, it broke down. Who had the last set of software? Where’s the part for this?’
We also open our software to the public meaning research teams in other universities can use it. A benefit for us is that it’s useful for the type of research where you need international benchmarking and validation. If every researcher has their own custom-designed experimental platform (which is generally the case in our robotics field, with everyone building their own robots), it is very difficult to compare the outcomes, apple to apple. There will also be a lot of re-inventing the wheel as each research team replicates necessary functions for their own research platform. With a system that is mass produced, comparisons can be directly compared as they are performed on the same experimental platform and the same basis of software functions.
The advantage of being within the University is that I have a fertile research environment and much greater ability to access funding and support. I can access colleagues and infrastructure in a way we take for granted. Furthermore, the support for commercialisation is important, too, because I enjoy seeing my work being useful to the community. Robotics is a very applied field of research. It’s very interactive with the user. If people don’t use it, there’s no point having the robot.
As told to Barney Zwartz
The EMU Upper Limb Rehabilitation Robot that was commercialised through Fourier Intelligence (pictured top) won the 2022 Red Dot Award for Product Design.
The University of Melbourne has announced the establishment of two new major investment funds dedicated to supporting Melbourne’s world-leading researchers to turn their extraordinary discoveries and innovation into commercial reality. Find out more about the University of Melbourne Genesis Pre-Seed Fund and Tin Alley Ventures.
First published on 16 March 2022.
Share this article
Research into practice: Victorian Government Land Use Victoria ‘ePlan’ Partnership
e-Plan, a partnership between the Centre for Spatial Data Infrastructures and Land Administration, and the Vic government has modernised the Victorian land administration system.
Algorand supports cryptocurrency teaching and research
Computing and Information Systems partners with Algorand Foundation to develop learning and research in blockchain and cryptocurrency tech.