Raphael Thys
Image: © titima157/Stock.adobe.com
Confirm researcher Dr Iman Abaspur Kazerouni has set his sights on smart robots for industry 4.0.
Dr Iman Abaspur Kazerouni completed both a bachelor’s and master’s degree in electronic engineering in Iran’s Hakim Sabzevari University. It was during the latter that he discovered an interest in image processing, leading him to do a PhD researching and designing a smart system for early breast cancer detection from thermal images.
After a few more years as an assistant professor in Iran, Kazerouni arrived at the Centre for Robotics and Intelligent Systems at the University of Limerick as a postdoc researcher in 2019. Last summer, he joined the SMART 4.0 programme at Confirm, the Science Foundation Ireland research centre for smart manufacturing.
Co-funded by the EU’s Marie Skłodowska-Curie Actions, SMART 4.0 is a postdoctoral training programme seeking to drive the fourth industrial revolution forward in Europe through creating the next generation of smart manufacturing leaders.
Kazerouni’s research continues to focus on image processing as he is now working to bring vision capabilities to robotics and smart industrial machines.
‘The use of smart mobile robots creates the flexible and autonomous industrial automation needed to create smart factories’ – DR IMAN ABASPUR KAZEROUNI
Image: Dr Iman Abaspur Kazerouni
What inspired you to become a researcher?
During my master’s, my supervisor guided me to image processing which was an interesting area for my new research. Researching robot navigation then became important for my own curiosity and understanding, and later more seriously.
Research in artificial intelligence is the best opportunity for me to build my fantasy world. As a researcher, I can develop my idea and find the answers to the questions about human intelligence. I learn new things in my research path every day and every day is a new day to move forward.
What research are you currently working on?
Daily, millions of tons of industrial feedstocks, components and ingredients flow into plants and warehouses while equal measures of packages, products and waste streams flow out. Each material movement is handled by either a person or manned vehicle, requiring vast amounts of workers and vehicles across every sector.
I am working on a vision system of smart mobile robots in the industrial environment to help these robots detect objects and find their paths automatically.
I am part of SMART 4.0 Marie Skłodowska-Curie fellowship research at University of Limerick and my supervisors are Prof Daniel Toal and Dr Gerard Dooly.
In your opinion, why is your research important?
The use of smart mobile robots creates the flexible and autonomous industrial automation needed to create smart factories, where the greatest asset is the exchange of information made possible by the integration of the latest intelligent technologies into robotics.
One major advantage of smart mobile robots is their computer vision capabilities. The complex array of cameras and sensors used by mobile robots to detect their unknown surroundings, allows them to accurately observe their environment in real-time systems. This is especially valuable in industrial settings that are constantly changing and shifting.
We can use smart robots in industrial environments and stores for any world crisis. As the smart robots do not need to keep a safe distance to slow the spread of diseases, they can keep working in any lockdown situation.
What commercial applications do you foresee for your research?
The commercial robotics market is expected to reach a value of €53bn by 2026. Commercialisation of smart mobile robots due to technological advancement in robotics is expected to drive the market over the next years.
Mobile robots have made remarkable advances in recent years. By improvements in computer vision and overall economics, robotics applications have become more and more useful and more convenient for a variety of repetitive and heavy tasks. Although most of the growth in robotics has happened in industrial applications, public applications are important for researchers and companies.
What are some of the biggest challenges you face as a smart manufacturing researcher?
For my current research, I work on deep learning techniques which are the-state-of-the-art algorithms in computer vision. These models need a high-performance computer for training and testing data. Thousands of researchers work on these methods and I must keep myself up to date by reading papers and implementing new codes.
Another challenge can be testing in real world, which is so tough in this situation with restrictions on travel and work in the industry area.
Are there any common misconceptions about smart manufacturing research?
Artificial intelligence and deep learning are still in their early development stages. It has a long way to go.
The experiments are likely performed in a laboratory setting, with clean and proper data, and the examples featured may have been selected to try to increase the hype surrounding the technology. The reality is that it is much harder to build a fully functional model that can work in the real world. Some of the most basic human tasks in computer vision and natural language processing have yet to be cracked by deep learning algorithms.
For mobile robots and self-driving cars domain, it’s a challenging race to see who can develop a car that can drive completely autonomously. It is a long way from reaching the market. Real-life situations have millions of variables and you need millions of examples, of all the possible solutions to train a mobile robot or self-driving car on how best to react in those situations.
What are some of the areas of research you’d like to see tackled in the years ahead?
With regards to smart mobile robots, I am eager to see more research in innovative machine vision and self-driving car solutions. I would like to see more development and uses in people’s daily lives.