In a recent study at the University of Groningen, physicist Elisabetta Chicca has explored the navigational capabilities of insects, focusing on their potential influence on energy-efficient computing. Insects, despite having very small brains, demonstrate impressive navigational abilities such as avoiding obstacles and navigating through narrow spaces. Chicca’s research aims to understand how insects manage these tasks with limited cognitive resources, a concept that could inform advancements in computing efficiency.
A key aspect of an insect’s navigation involves the perception of motion, which helps in estimating distances. Chicca explains that insects use the relative speed of moving objects to assess their environment. This is similar to observing how trees appear to move faster than distant houses from a moving train. To handle complex navigational challenges, insects modify their behavior, for instance, by flying in straight lines and then turning, effectively managing their limited brainpower for navigation.
Chicca, along with PhD student Thorben Schoepe and neurobiologist Martin Egelhaaf of Bielefeld University, developed a model to understand these neural mechanisms. This model, which is based on steering towards areas with the least apparent motion, was then applied to a small robot. The robot was tested in various environments, including a corridor lined with walls featuring a random print, and other virtual environments with obstacles or small openings. The robot’s behavior in these settings mirrored that of insects.
Chicca’s research also touches on the distinction between learned and innate behaviors in robotics. While much of modern robotics focuses on machine learning and the development of skills over time, Chicca points out that insects are capable of efficient navigation from birth, a trait that is hardwired into their brains. This insight opens up possibilities for more efficient computing solutions. Chicca’s team has previously developed a small, efficient chip, and there are aspirations to integrate insect-like navigational behaviors into similar hardware. This approach suggests a future where specific, purpose-built hardware could perform tasks more efficiently than general-purpose computers.
Photo credit: Leoni von Ristok, University of Groningen
