This line of research investigates the detection of sensor information that may be compromised. Fundamentally, any such detection relies on accurate knowledge of both the system dynamic behaviour, and judicial modelling of system observations.


As system automation is being applied to increasingly complex tasks, autonomous systems are now operating in dynamic and non-predictable environments. This shift in the domain of operation requires autonomous systems to actively perceive the world in which they operate and make decision based on the information they gather.

In this project, the Autonomous Systems Research Centre engaged with a local industry partner (MRA) to enable perception of coal vessels using a variety of on-board sensor information. Using a mixture of Bayesian and machine learning techniques, the shiploader systems were enabled to autonomously detect points of interest, such as hatches, on coal ships.

Stockpile profile mapping with MRA and PWCS

Ports and mines use large stockpiles of bulk material as a storage buffer. The process of adding new material to a stockpile is called stacking, and the removal of material is called reclaiming. Stacking and reclaiming are typically performed by large machines that interact with the stockpile and these machines form part of the critical path and directly effect port efficiency. To automate and optimise stacking and reclaiming it is vital that the machines are aware of the stockpile profile, especially since the profile can change over time due to changing environmental conditions.

To this end, the Autonomous Systems Research Centre engaged with local partner MRA and to deliver real-time modelling of stockpiles based on a fusion of mass-flow principles and Radar and Laser scanning sensors. This approach is proving to be very successful and is currently undergoing commercialisation. This work is in collaboration with Port Waratah Coal Services and Abbot Point Coal Terminal.


Automated safe loading magnetite onto a Barge is challenging. In this project, the Autonomous Systems Research Centre engaged with a local industry partner (MRA) to enable perception of magnetite using a variety of on-board sensor information. Using primarily Bayesian techniques, the research group delivered accurate estimates of the barge list (or roll) and also an accurate profile (or map) of the magnetite.

In this project, we developed new Simultaneous Localisation and Mapping (SLAM) techniques that employ optimisation methods to offer superior scan matching.


One of the major challenges for automated loading and unloading of bulk material trains is estimating the position and velocity of individual wagons.  

Motivated by this, the Autonomous Systems Research Centre are investigating the next generation of wagon speed and position measurement based on 2D laser scanning technology.

This work has been commercialised by our industry partner MRA. They now have several successful installations throughout Australia.


Our background and vision


The research team


A list of research publications


Contact us

Close Menu