How robots will collaborate with humans: an interview with Roman Meshcheryakov.

— Roman Valeryevich, you are a specialist in "cyber-physical systems." What does that entail?

— Cyber-physical systems (CPS) are complex organizational and technical systems that operate within a physical environment. They include computational systems that receive, process, and output information. They also have the capability to sense information and interact with the physical environment.

Some well-known examples include IoT (Internet of Things). This technology has been around for a while and encompasses grid systems, industrial internet, and smart home technologies. In this century, the paradigm of IoE (Internet of Everything) is actively developing.

Wearable electronics on "smart clothing" are also considered cyber-physical systems. "Smart hospitals," "smart greenhouses" — everything is "smart." There are also extensions of cyber-physical systems, such as sociocyber-physical systems, which incorporate humans as both subjects and objects of control.

In our laboratory, we explore various aspects of the theory and practice of cyber-physical systems, focusing on three major areas.

The first area involves robotic complexes based on various environments (land, air, water, space) that operate in groups and interact with one another. On board an aircraft (which could also be a ground, water, or underwater vehicle), there are several processors that create a unified computational environment. They function both autonomously and as a group. Our research focuses on control methods, interaction, communication, task distribution, and more.

The second major area is ergatic systems, where humans are involved. In addressing such tasks, we define the human's role: what they need to do versus what should be performed by the robot (automated and/or information system). Humans not only make decisions but also carry out tasks assigned to them by technical means (specifically, robots). It is crucial for humans to have a visual representation of data and knowledge, as well as coherent data processing.

The third area is cybersecurity, which ensures the safe operation of cyber-physical systems under various conditions. In addition to maintaining confidentiality and data integrity during transfer and interaction, special attention must be given to maintaining availability — ensuring that information is transmitted between subjects in accordance with time requirements that allow the system to function effectively.

— Could you please explain where our developments in this field stand on the world map? How good are we?

— In the field of human-machine interaction, we are undoubtedly at the forefront of science. In terms of biointerfaces, which facilitate communication between humans and computers using not only speech and voice but also biological signals derived from humans, our research and developments are highly valued abroad.

We are conducting advanced research in task distribution between humans and robots. Currently, how do robotic systems operate? Either the robot operates, or the human does — sequentially over time or simultaneously in different spaces. Tasks are rigidly distributed, and mixing is not allowed. When humans and robots intersect in a small space at the same time, questions arise regarding predicting and forecasting robot behavior. What should its behavior scenario be?

In the field of human-machine interaction, we are undoubtedly at the forefront of science. Our research and developments in these areas are highly valued abroad.

For example, in Russia, it is currently prohibited for unmanned and manned aircraft to occupy the same airspace simultaneously. However, we understand that soon robots will be part of our environment just as smartphones and smart devices are. We have world-class developments in distribution and interaction. We also have solid advancements in the accessibility of cyber-physical systems, ensuring telecommunication requirements in various noise conditions, ensuring the reliability of communication channels, and presenting information to humans.

The lack of computational resources and the peculiarities of the electronic component base lead us to delve deeper into models and algorithms of behavior. We excel in mathematics and are capable of creating methods and models. The challenges posed by the Russian component base are both a hindrance and an asset.

Just like in the 1990s, when we had limited computing technology, we solved problems by developing effective software and algorithms. This led to an explosive growth of mathematical and programming schools in computational systems.

Abroad, it was easier to write a program that produced a solution in an hour within five minutes, while we did the opposite: we spent an hour writing a program that solved a problem in five minutes. In fact, we are returning to that model. The difficulties in material and technical support prevent us from fully disengaging and implementing everything we want, so we must think through and develop what is promising. Some classes of problems cannot be solved even with promising tools, including neuromorphic or quantum computers. For example, NP-complete problems are addressed using metaheuristics.

— Could you provide examples of such solutions that we can be proud of?

— The first example that comes to mind is the management of a group of unmanned aerial vehicles in solving complex tasks. Depending on the environment and the internal state of the UAVs, they can redistribute functions among themselves. The first task is to create such a group. The second task is to maintain its current status during operation, as unforeseen situations may arise: someone might run out of battery, or there could be a malfunction — a propeller might break, or there may be changes in weather conditions such as precipitation. We know how to manage this, and it is certainly one of the most advanced things that exist.

The second example involves various types of control interfaces using biological signals, including electroencephalography (EEG). We conduct these studies in collaboration with our partners — Voronezh State University, the Institute of Medical and Biological Problems of the Russian Academy of Sciences, the Yuri A. Gagarin Research and Testing Center for Cosmonaut Training, MIPT, and others.

— Where do the tasks originate from, and how do they evolve?

— Tasks arise from real-life situations, from observing how technologies develop and where the world is heading. Initially, these are exploratory or applied tasks, but they lead to fundamental problems that require in-depth development of mathematical, modeling, methodological, and algorithmic support, as well as the need to create new energy, propulsion, and other physical solutions (as part of the physical world of cyber-physical systems).

Our main goal is to improve people's lives. Not just to make it easier, but to make it better. The social sphere demands that as life expectancy increases, robots (or other technologies) become assistants to humans in their routine operations.

Here, we cannot overlook the role of artificial intelligence technologies, which are necessary for implementing functions so that the system operates in a manner that is most understandable to humans. Users want to know what will happen next, enabling them to trust the system.

Our main goal is to improve people's lives. Not just to make it easier, but to make it better.

In the laboratory, we are conducting a series of Russian Science Foundation grants that establish the fundamental foundations for constructing elements of cyber-physical systems, such as mixed teams, biological interfaces, robotic medical systems, ensuring cybersecurity in cyber-physical systems, visual analytics, and intelligent data analysis.

We have topics related to the analysis and behavior of cyber-physical systems as part of a larger class of complex systems, where the number of connections between system elements significantly exceeds the number of possible options for analyzing all existing operational scenarios. Here, we are developing one of the most promising directions today — using robots in areas where humans should not be present. For example, underwater, in space, on volcanoes, during emergencies, or in any aggressive environments.

— Can we say that in your field the logic of development goes from practical tasks to fundamental ones?

— How does development progress? When a new class of robots emerges, such as those operating in underwater environments, tasks arise concerning computational resources, underwater energy, decision-making, sensor cleaning, and various other events, for which we lack sufficient technical, computational, and other means to track.

For instance, there are currently developments in a very large class of biomimetic robots. If we are talking about aquabots, they must be designed in such a way that fish do not get scared. We cannot use classic rotor engines in water, as aquatic inhabitants are very sensitive to acoustic noise, which can be harmful to them. We need a device that swims like a fish, emits a biomimetic signal that is more familiar to underwater inhabitants; we need information transfer to occur on waves that do not affect the surrounding environment and its inhabitants; we need the device to integrate smoothly and naturally into the environment. Thus, there is a task — to create a behavior model. All of this forms new subclasses of cyber-physical systems in which we will live.

If we are talking about aquabots, they must be designed in such a way that fish do not get scared.

There is also the task