Innovative rat robot inspires researchers

For Graz researcher Thomas Schmickl, the work is remarkable for several reasons, he writes in a commentary.

For example, the idea of ​​using robots that fit into a group as naturally as possible to conduct behavioral research, to control animal groups to a certain extent or to gain new insights into the interaction between animals, people and artificial intelligence (AI) systems is tempting for many researchers and investigators. The fact that such systems actually behave more or less authentically, are accepted to some extent by their living models, and elicit responses is technically very difficult to implement, as the team led by Qing Shi from the Beijing Institute of Technology ( China) writes in their work. If this succeeds, you would be dealing with a ‘biocompatible’ or ‘bioeffective’ robot, as the head of the University of Graz’s ‘Artificial Life Lab’, Thomas Schmickl, explains in a perspective article on the publication.

The “SMuRo” system learns by observation
The Chinese scientists have now built a flexible robot that resembles a rat and moves on a base with wheels. This allows SMuRo to mimic typical movements. However, these are not programmed into the system, but are learned by observing real rats. With the help of machine learning, SMuRo processes what it sees and experiences and gives it its own meaning, as it were. The scientists, including Zhenshan Bing and Alois Knoll from the Technical University (TU) of Munich, used an approach modeled on social or imitation learning in humans and animals.

The group from Munich has been working for years on the development of robots that can perform the movements of mice as authentically as possible. The team and colleagues from China presented their neurorobotic mouse (NeRmo) in the journal ‘Science Robotics’ last year. The current SMuRo system partially takes over this work.

Robo-rat in “behavioral dialogue” with real animals
By observing the rats’ interactions with each other, the robot gradually acquired knowledge about its ‘fellow rats’. He collected data about the spatial position of body parts during movements or about the way the animals behaved in a room. Based on this, the new system essentially developed its inner image of rats. This allows him to actively and authentically imitate movement sequences. What’s more, it turned out that the robot’s actions influenced the behavior of the rats that came into contact with it.

This is the first time the social learning cycle has been demonstrated in a robot that interacts freely with animals, Schmickl writes. SMuRo shows unprecedented social learning ability in a long-term “behavioral dialogue” with real rats, according to the Graz scientist, who recently wrote an article in “Science Robotics” about possibilities for teamwork between robots and wildlife.

Expert: “Impressive step forward”
In the case of SMuRo, it was shown that the robotic animal learned the typical rat behavior of holding, jumping and social nose contact and used this in such a way that it motivated its counterpart to show more exploratory behavior and stimulated their curiosity. When the robot held the animals more tightly, they made more negative sounds, while more jumping and nose contact resulted in positive feedback.

For Schmickl, “an autonomous robot that can handle such long and complex interaction patterns” is an “impressive step forward” – especially considering how mercilessly animals reject a bad imitation of themselves: “It will be interesting to see where the door that SMuRo opened, still guides us.” Such “biohybrid systems,” as the study authors put it, could in the future help “understand the complex interactions between sensory perception, behavioral decisions, and behavioral decisions.” “internal states” and draw conclusions about the underlying brain functions. Ultimately, this could also be used to better design the interactions between humans and AI systems, according to Qing Shi’s team.