Earlier this year Australia experienced one of its deadliest and most widespread bushfire seasons. Around the globe fires engulfed areas that had never previously been burnt, such as the dense tropical Amazon and suburban Los Angeles.
One aerospace engineering graduate turned robotics visionary, Vera Somers, is hoping her research will one day help detect bushfire prone areas and survey at-risk areas by creating drones for bushfire surveillance and intervention.
“The large-scale impacts and real-world risks make bushfire research extremely important – this last bushfire season stresses how vital this research is,” said Ms Somers, who joined the University of Sydney from the Netherlands to complete a PhD at the Australian Centre for Field Robotics under the supervision of Associate Professor Ian Manchester.
Ms Somers hopes UAVs and automation can offer faster and more efficient situational awareness, potentially making firefighting efforts safer by containing fires before they become uncontrollable.“Although many fire propagation models and simulators already exist, the link connecting them to automated decision making and UAV path planning is still missing,” she said.
“We use optimisation methods to study how bushfires spread and how to control them. By creating risk maps and UAVs we can find the areas most in need of monitoring, and then use this information to deploy targeted resource allocation for both prevention and firefighting.
“We actually model and control the spread of fire in a similar way to the modelling of epidemics. It is fascinating how both bushfires and epidemics – as well as other spreading processes – have very similar underlying models.”
Ms Somers also recently assisted the post-bushfire effort by helping rebuild fences that had been lost in the January 2020 fires.
Amrit Sethi, who joined the University of Sydney from India and who is a graduate of a Bachelor of Mechatronic Engineering (Space) and a recipient of the University Medal, is working with clean energy expert and aerospace engineer from the School of Aerospace, Mechanical and Mechatronic Engineering, Associate Professor Dries Verstraete in the development of a technology that aims to extend the lifetime of hydrogen fuel cells.
“Rising pollution levels and their impact on climate change highlight the need for clean energy sources” said Mr Sethi, who is the recipient of an International Strategic Scholarship and is completing a PhD in Aerospace Engineering.
“Hydrogen fuel cells produce clean energy by combining hydrogen and oxygen with water being the only by-product, he said.
“However their short lifetimes do not justify their high costs for many real-world applications. My research focuses on developing a prognostics and health monitoring framework that can help extend the lifetime of hydrogen fuel cells to make them more efficient, hopefully leading to a wide-scale industrial adoption.
“Our research is based on the use of machine learning techniques that identify short-term and long-term fuel cell health issues, which can then be used to improve the fuel cell's lifetime.
“Tapping clean energy sources is the need of the hour, and with my research, I aspire to help the world be a little greener.”
As part of a research group focused on improving the landing operations of vertical take-off and landing aircraft led by Associate Professor Ben Thornber, Master of Aeronautical Engineering student Jack (Hee Sung) Park has developed a lightning-fast computational method to assist both manned and unmanned aircraft to operate and land on platforms such as ships, oil rigs and buildings.
“The landing of helicopters and other aircraft on platforms such as ships, oil rigs and urban buildings is inherently risky. My research aims to better understand how aircraft descend and ascend in various conditions to improve their safety and performance,” said Mr Park, who joined the University of Sydney from Handong University, South Korea in 2018.
“Previous simulations have required complex recalibration, whereas the method we have developed computes simulations in a fraction of the time.
“I hope my research results in an accurate and computationally efficient way of predicting helicopters and unmanned air vehicle aerodynamics.”
As part of the Intelligent System Transport Group led by Australian Centre for Field Robotics academic Professor Eduardo Nebot, Colombian PhD student Julie Stephany Berrio Perez is developing an algorithm that helps autonomous vehicles’ in-built mapping systems detect changes in the environment and reflect those changes into the map so they can navigate roads safely.
"One of the requirements for autonomous vehicles to be able to operate persistently in typical urban environments is to maintain high accuracy position information over time. In other words, the capability of an autonomous vehicle’s mapping and localisation system must adapt to changes in the environment,” said Ms Berrio Perez.
"A map for autonomous driving applications which is created today might not be entirely valid next month, or even next week due to changes in the environment caused by construction sites, speed zone changes, new pedestrian paths, changing signals and new road infrastructure,” she said.
"Autonomous vehicle maps require continuous updates to operate safely. An alteration of the surroundings could cause the vehicle to stop operating autonomously or even cause an accident. It is then indispensable to have an algorithm capable of reflecting environmental changes on the vehicle's map."
Ms Berrio Perez’s work has been implemented on the University of Sydney Campus Dataset, which was collected for more than a year. The autonomous, electrical vehicle was retrofitted with multiple sensors and was driven on a weekly basis around the University of Sydney campus.