My interests lie in the development of vision systems for long term
robotic applications ranging from underwater and ground to airborne
autonomous platforms. The overarching theme of my work is in robust
visual perception for robotics and the major areas I have contributed to
1. Multiple robotic platforms that communicate
visual maps to other robots and humans building a consolidated
representation of a large scale outdoor environment. Vision was the
primary sensor for perceiving the world with a focus on representations
that could be communicated efficiently, fused with other information
consistently, in a decentralised manner.
2. A full-sized
commercial vehicle able to navigate and localise in urban environments.
The vehicle was able to obey traffic rules and drive within lanes,
strongly aided by visual information. This was the only Australian
entrant to the USA Urban DARPA Grand Challenge which was a major
demonstrator for the capabilities of robotics in civilian spaces.
Dragline payload density estimation for optimising mining production.
Robotics and automation play a major role in mining and perception in
unstructured environments is not only crucial for robust operation but
also for streamlining the entire productivity pipeline. This work
demonstrated real-time estimation of payload density, enabling accurate
calculations of total mined product and illustrated how to improve
4. Robust multiview vision algorithms
for robots in harsh and challenging environments. I have developed
short baseline stereo algorithms that are robust to vibrations and
pressure changes between stereo cameras with views of environments at
long ranges. This work is particularly suited to vision only pose
estimation and 3D reconstructions of visual observations from Unmanned
Aerial Vehicles and autonomous submersibles.
This work falls under the research area of the CyPhy lab
which focuses on robotics in general.