In June 2024, physicist Dr. Dovilė Čibiraitė-Lukenskienė joined the FTMC. She studies terahertz, which is located in the electromagnetic wave spectrum between microwaves and infrared radiation. She and her colleagues are developing high-frequency technologies with high expectations in key areas such as quality control, security and faster wireless communications.
After completing her Bachelor's and Master's studies at Vilnius University, in 2016 D. Čibiraitė-Lukenskienė went to Germany: she defended her PhD thesis at the Johann Wolfgang Giotte University in Frankfurt, and later continued her research career at the Fraunhofer Institute for Industrial Mathematics ITWM.
Over the years, she has been interested in the properties of terahertz waves and related research in the application of sensors in spectroscopy, imaging, communication or other scientific systems. In one of her projects at Fraunhofer ITWM, terahertz waves came in handy in a seemingly unexpected area: the idea of sorting bulky wood waste in a more environmentally friendly way. The aim is for artificial intelligence at the conveyor belt to recognise which waste is recyclable and which is not, using sophisticated sensors.
Dovilė presented a presentation on this topic at the high-level optoelectronics science conference APROPOS 19, which took place at FTMC in autumn. It is also mentioned in the special book "Female Researchers in Focus - We Create Change", published on the initiative of the community Fraunhofer Gesellschaft. The publication is designed to showcase the work of female researchers in the institute's community and to encourage young people to take up technology-related careers, which, despite the challenges, are very interesting.
And we talk to her about a research project that is relevant to everyone.
(Dr. Dovilė Čibiraitė-Lukenskienė. Photo: FTMC)
What is the current problem with bulky wood waste?
The current automated sorting procedure is as follows: old furniture and other bulky waste brought to the larger recycling plants is placed on a conveyor belt by an excavator operator, and the staff removes items from the mass of wood waste that belong to another category (often with clearly visible metal, plastic or other additional waste). The wood is then moved on to the waste management site and shredded down to a few centimetre pieces.
The next step: X-ray machine is used to scan the pieces for any remaining metal fragments that are invisible to the naked eye, and compressed air is used to dislodge them. Finally, the wood is further crushed and transported to the incineration sites.
That's all well and good - but from an economic and environmental point of view, incineration should be the last resort once you know that there's nowhere else to put it. And if we want to get more value for money, we need to make sure that the right waste can be reused and can go back into the economy. For example, we could recycle good quality wood from old furniture into pallets or paper. In this way, it is as if we are giving a second life to the material before it is burned. However, current technology does not yet allow us to efficiently and accurately determine the suitability of waste.
(Photo: Pexels.com)
How can terahertz waves help here, and what was your task?
We wanted to automate the sorting process itself, so that it would not require human intervention - and so that more wood waste suitable for recycling would be selected. And four sensor systems (one of which I developed) can help to improve this quality: thermal, near-infrared, visible-light and terahertz. In this way, we monitor bulky waste moving along a conveyor belt in different ways, and record the resulting images, which we then provide as a learning tool for the artificial intelligence. One of the problems preventing the automation of such work so far has been the lack of visual databases. It was therefore necessary to create them.
Currently, 50-60% of all wood is sorted. With our former project partners, we have estimated that a robot or machine could separate recyclable wood waste with around 90% accuracy using the new technology. I expect that if this is put into practice, together with the extraction of additional waste, at least 70-80% of the wood could be sorted.
Another advantage: the algorithm would not only sort the wood waste, but also identify which are less or more polluted. This is also important, because we could reuse wood that is "healthy" several times over, compared to that which is contaminated with carcinogenic varnishes and has to be burned immediately at extremely high temperatures.
(Photo: Pexels.com)
How has this phase ended for you and the German researchers - and what's next for them?
A demonstration device was developed which generated many real images of bulky waste and different wood textures. Researchers involved in the project are now trying to select which computer neural network algorithms work best and which electromagnetic wave sensor systems are most useful for automated sorting.
The next step is to go to a bulky waste sorting plant with a conveyor. The developed device will be tested to see how it works in real working conditions.
I look forward to the results of this project and I hope that the relationship with the Fraunhofer community will continue with new international projects.
Witten by Simonas Bendžius