DrDonald Dansereau

Keywords: robotic imaging, computational imaging, robotic vision, novel imaging, light fields

In order to do their jobs, robotic devices need to be able to "see". However, currently robotic vision tends to rely on off-the-shelf cameras that are not well suited to this job. Dr Donald Dansereau's research brings together concepts from computational imaging and robotics to develop new kinds of cameras and algorithms that will help robotic devices to see and do, enabling next-generation robotic devices to operate in new domains and with greater robustness and safety.

"Robotics is set to revolutionise the way we work and live, affecting transport, security, construction, agriculture and environmental monitoring. But present-day robot perception falls well short of the levels of reliability and safety we need, preventing their widespread deployment.

"For example, we will not have fully autonomous cars until they can reliably perceive their environments across all driving conditions, including rain, fog, snow, smoke, dust, and the six orders of magnitude in ambient illumination that lie between night and day. The same principal applies to pilotless aircraft, robotic agricultural devices, autonomous infrastructure and environmental monitoring equipment, and robotic medical devices used inside the human body.

"Reliable perception is a critical building block in enabling robotics to play a positive role in how humanity develops. I find this work exciting, because its aim is to ensure that robots will be safe, ubiquitous, capable assistants that represent a massive force multiplier in what humans can achieve in a sustainable future.

"I began working in this field in 2002, when I started my master's work on light field imaging. In 2014 I completed my PhD here at the University of Sydney, showing how light field cameras can help robots to see better in low light and underwater. In 2018 I returned here to take up my present role.

"My role at the University of Sydney is allowing me to broaden the scope of my work to consider new, specialised imaging modalities and new design and prototyping technologies. We expect the imaging systems I develop here to enable and help drive a very rich range of robotics applications, from autonomous driving to underwater survey and surgical robotics and beyond."