started out with a science degree in pure mathematics and geology. I guess I was the only student in the university with that particular combination at the time. I did maths out of love for the subject, but I always wanted to be involved in something more applied, and geophysics and structural geology seemed a good option. At the time the ANU did not have engineering or applied maths, so in a way I made my own applied degree — ANU has always been good at providing students with flexible study options’, says Robert Mahony.
‘After my degree I worked for a geophysics company in Canberra in the area of marine seismic exploration, but I soon realised I really wanted the challenge of a job with a research focus and I chose to go back to study a PhD. I went looking for a research group that was doing world-class applied research — I thought that if they found it interesting and exciting then so would I. I was lucky enough to end up in systems engineering working with Professor John Moore.’
Robert received a PhD in 1995 in systems engineering at ANU before completing a postdoctoral fellowship in France and a Logan Fellowship at Monash University. He returned to Canberra and the ANU in 2001, becoming a Professor in the Research School of Engineering in 2011.
‘My work is in using sensor systems to control robotic systems. I started work on motion control of vehicles during my PhD, and then in France began work on aerial vehicles’, says Robert.
Robert explains that aerial vehicles that deliver images to people on the ground are a growth industry as technology advances. ‘For example, 15 years ago the inertial measurement unit that is a key component of such a vehicle cost $3000 and weighed 500 grams. Now it costs $30 and weighs 50 grams. As the vehicles improve and become more affordable the uses are expanding. There are thousands of start-up companies world wide offering aerial vehicles for a wide range of tasks.’
‘As well as being a boon to the film and television industry, the vehicles can be used for all sorts of inspections — for example you need to regularly inspect dam walls and bridges for cracks or corrosion, normally a labour-intensive process done by abseiling. In industry there are many parts of a plant that need inspection, from large pressure vessels to cooling towers and pipes on factory ceilings. On the smaller scale, electricians or phone technicians can use aerial vehicles to inspect problems with poles or building connections.’
‘In safety and security, firefighters already use such vehicles to look through the windows or a burning building to see if there are flames or people. Police can use them to survey crowds and demonstrations, and protesters also use them to video and document police practice.’
‘A fundamental challenge in aerial vehicles,’ explains Robert, ‘is in sensing the vehicle state, a mathematical description of the position, velocity and orientation of the vehicle relative to its environment.’ He developed an attitude ‘state observer’ that has become the industry standard for quadrotor vehicles, and is now working on exteroreceptive sensing paradigms — sensor modalities that provide information relative to the surrounding environment. Robert believes that vision systems are a key technology in this area. ‘Such systems are inexpensive, light, require low power, and are rich in information. However, using vision to control aerial vehicles is complex.’ He has made key contributions in the formulation and analysis of image-based servo control for aerial robots.
A further constraining factor in current technology is that you need to be an expert pilot to fly an aerial vehicle. He is developing technology for pilot-assist systems which will make it possible for anyone to fly a quadrotor with no training.
When not working on new technology, Robert is also a musician in a several folk bands, a cyclist and sailor.