Interview Dr Gregory Yovanof, Scientific leader at IDE
HERMES is a revolutionary technology that offers low-cost, low power consumption high bitrate chipset working in subterahertz.
Can you please introduce yourself, Dr. Yovanof?
Certainly. I am an electrical engineer and communications expert by training, and I received my PhD from Southern California University in Los Angeles after completing my undergraduate studies in Greece. After that, I spent nearly 10 years working for HP Labs and Eastman Kodak in Palo Alto, as well as for a few startups in Silicon Valley during the early days of the internet. My work focused on developing hardware and chips for the DVD market in the 1990s, and I also worked on deep space communication programs, specifically developing communication links for the Hubble telescope and the defense industry. This is how I became involved in hardware development.
I have produced and designed chips for companies like Akai, and was a pioneer in the DVD market industry. In 2001, I returned to Greece to join a private university established by the founder of Intracom, collaborating with Carnegie Mellon. I worked there for 17 years as head wireless communications and innovation management, developing a program for early-stage entrepreneurs on how to take an idea from the lab to the market. This program ran for 12 years.
Since 2019, I have been working part-time for IDE, with a focus on wireless communication, and I am also the head of a network called Strategis - a maritime ICT cluster in Piraeus, Greece. Our goal is to network companies from the ICT domain with applications in shipping and the blue economy.
Can you tell us about HERMES?
Sure. IDE's contribution to HERMES is in the development of applications in spectrum sensing. We want to take advantage of the fact that Walsh transform offers the opportunity to do some of the signal processing in the analog domain to speed up the spectrum sensing functionality, which is critical in dealing with issues of spectrum scarcity and the design of systems in cognitive radio. With cognitive radio, you have to rapidly sense and detect existing wireless systems operating in a certain area and take measures to facilitate the coexistence of several wireless systems in that area. Walsh transform offers advantages in doing fast processing in the analog domain, which can lead to advancing the stage of spectrum sensing. This technology will be useful for detecting a large number of signals simultaneously in a given area and then taking measures to operate. We need to co-integrate the work of the partners so that the entire system of wireless communication and spectrum sensing perform seamlessly and effectively over a large spectrum area from low frequencies up to the sub-terahertz bands. That is what attracts us and motivates our work in HERMES. We hope that Walsh will help us in detecting a large number of signals in a very short time period, which is critical for applications such as maritime surveillance and the control of a swarm of aerial and sea drones.
What is IDE's role in HERMES?
We contribute to WP1 with the specification of scenarios, taking into account the requirements of spectrum sensing. In WP4, we help with the development of the detection of several signals in the presence of interference, the overall system architecture, defining use cases, and testing chipsets with maritime surveillance in mind.
What benefits will HERMES bring to European citizens?
HERMES is a revolutionary technology that offers a low-cost, low power consumption, high bitrate chipset working in sub-terahertz band. If successful, it will revolutionize machine-to-machine communication and IoT communication. It will also enable human-to-device communication, offer the opportunity for the development of infrastructure that facilitates metaverse or web 3.0 applications, and provide the European industry with a huge advantage if funded by Europe.
For you IDE is it important that this new technology is being funded by Europe for Europe?
If Walsh proves to be as efficient as we expect it to be, it will give a huge advantage to the European industry. The EU funding will allow us to develop this technology internally and raise it in Europe.
Can you speak to the importance of radio in spectrum monitoring for IDE applications and maritime surveillance?
Absolutely. In the context of maritime surveillance, there are often numerous wireless systems operating in close proximity, including IoT communication, cellular networks, satellite links, and drones. It is essential to identify the various signals in the area through spectrum sensing in order to increase awareness and take appropriate action to avoid interference and improve spectrum efficiency. In such dynamic environments, a clear understanding of the existing signals is crucial for situation awareness and effective communication.
What benefits will HERMES' high precision RF provide?
HERMES will provide advantages in detecting signals very accurately, identifying the presence or absence of a signal within a range of frequency, and identifying any additional interfering signals that happen to coexist in a certain area. The Walsh spiky signal of HERMES will give an advantage over a traditional Fourier based spectrum sensing approach, allowing for the detection of desired and interfering signals simultaneously.
What amount of time and effort will be needed to transition HERMES into a tool for European border surveillance?
Usually for new technologies it is not an easy task to bring the final product to the market. Transitioning HERMES into a tool for European border surveillance may take two to three years after completing the project in the best-case scenario.
There is currently a lot of discussion about the increasing prominence of ChatGPT and AI. Is AI being utilized for spectrum monitoring, and will Hermes play a significant role in this usage? What are the potential advantages and drawbacks of implementing such a system?
AI is certainly an exciting prospect, and it has the potential to be a game-changer. We are amazed at how rapidly AI is advancing, especially with ChatGPT. In comparison to where we were just six months ago, there has been a significant leap forward. We are hoping to see similar progress in signal processing, which is why we are experimenting with AI in this field. We are very excited and eager to see the results of implementing AI in end-to-end signal processing through Hermes.
We are working closely with KUL to create a library of signal signatures that can help us classify signals and use AI to detect signals in noisy environments. We receive signals in a Walsh receiver and use AI to distinguish between various types of signals that may coexist. We are optimistic that this approach will yield results, but it will require a great deal of effort. Training AI models requires a massive amount of data, and we need to find a way to create or simulate data to train our algorithms effectively. We have several years to work on this, and we are collaborating with KUL to train their AI under various scenarios.
The potential benefits of applying AI in signal processing would be substantial. It would significantly enhance the capabilities of current spectrum sensing devices that rely on traditional mathematical algorithms.
As for the risks, I don't see any in our case. AI bias is a concern in other applications, but in our scenario, we are simply using AI to detect signals. The signals will exist regardless of AI, so it is merely a matter of observation and detection. AI will not interfere with what is happening; it will be a passive observer attempting to identify what is occurring. I cannot see how that would pose any risk.
If Walsh proves to be as efficient as we expect it to be, it will give a huge advantage to the European industry.