About Victor’s research
The main objective of Victor’s PhD is to develop a new radar algorithm to classify vehicles according to a classification scheme that takes into account the length of the vehicle but also the configuration of its axles.
The main problem he is trying to solve is how to extract the physical properties of a vehicle and in particular the configuration of its axles using only tracking radar technology.
“I have always been attracted to the application of signal processing to real-world applications. This project is a perfect example! Furthermore, the project has the potential for the application of machine learning techniques which is an area I am very keen to learn in greater depth.”
“We have developed a proof of concept that demonstrates the possibility of using radar data to extract the vehicle characteristics we need to provide a vehicle classification. As there was a lot of uncertainty about the feasibility of this project at the start, this is an important step that I am proud to have taken.”
He has been working with Sensys Gatso Australia, a traffic enforcement system manufacturer based in Port Melbourne and Sydney. “Working with them has been great! Access to equipment and test sites to collect data has been greatly facilitated and access to knowledgeable people has been fabulous to keep this project on the right track.”
“This work will provide road authorities with another cost-effective and non-invasive means of collecting traffic data, and will help to improve the safety and fluidity of our road network.”
In this video we profile not ony Victor and his work, and the assistance from his industry partner Sensys Gatso, but also the iMOVE Industry PhD Program itself.
A poster for Victor’s PhD project, made for display at the 2022 ITS Australia awards.
Reflections on his PhD
This project places Victor directly in the middle of the research and industrial world which has allowed him to learn a lot about how they work and how they are linked.
“Research can sometimes take a long time and that there are many pitfalls on the way to achieving the desired results, which is quite normal. The most important thing is to enjoy doing it, to learn from these pitfalls and to try to do your best!”
He intends to continue working in the world of signal processing, and will strive to keep learning new things and apply this knowledge in real world applications.
A word from Alastair Wiggins, Technical Director, Sensys Gatso
Why is this PhD topic important to investigate?
Traffic authorities are increasingly favouring non-invasive methods for traffic data collection. However, there are no truly non-invasive methods for performing vehicle classification in accordance with the Austroads standard. The ideal solution would comprise a pole mounted sensor installed at the roadside, as this would simplify installation, maintenance and yet still avoid potential acts of vandalism.
What are the major challenges to overcome in the field?
The project has already developed an early version of the tracking radar algorithms to recognise all 12 classes of the Austroads scheme. The next phase of the project is gathering a significant amount of ground truth data, which will be used to validate and help improve the algorithms. A independent piezo-based classifier will be used as a reference system. The challenge will be to automate the corroboration between the two systems, thereby limiting the requirement for any manual verification of these results to edge cases.
Where might this work lead in the near (and far) future?
Existing tracking radar based speed enforcement cameras could be enhanced to accommodate the new algorithms and provide high quality traffic data analysis for all passing vehicles – regardless of their speed. There is great demand for a stand-alone vehicle classifier product using this technology, which could also be used as part of future tolling systems. There are also potential applications in the railway industry, which are currently being investigated.
If you’d like to contact Victor about his research, please click the button below.