American developers deliberately cut the technical equipment of mini-robots and looked at what tasks such devices can perform, being “stupid”. The journal Science Advances writes about the work of scientists.
A group of researchers led by Dana Randall, professor of computer technology at ADVANCE, and Daniel Goldman, professor of physics at the Georgia Institute of Technology, wanted to show that even the simplest of robots can still perform complex tasks.
The results of performing these tasks with the help of “silly robots” (in fact, mobile free-flowing particles) exceeded the expectations of scientists. Researchers report the ability to remove all sensors, as well as communications, memory and computers, and still complete a set of tasks using the physical characteristics of robots.
BOBbots (behaving, organizing, buzzing bots – literally “behaving, organized, buzzing bots”), named after the pioneer physics of free-flowing bodies Bob Behringer, “as dumb as possible,” explains Randall. “Their cylindrical chassis has vibrating brushes at the bottom and weakened magnets at the periphery, which makes them spend more time in areas with more neighbors.” The experimental platform was complemented by precise computer simulations led by Georgia Institute of Technology physics student Shankai Li.
Despite the simplicity of BOBbots, the researchers found that when robots move and bump into each other, they “form compact aggregates capable of collectively clearing debris that is too heavy to move alone.” “While most people are building increasingly complex and expensive robots to ensure coordination, we wanted to see how complex problems can be solved with very simple robots,” says Goldman.
Their work was inspired by a theoretical model of the movement of particles on a chessboard. A theoretical abstraction known as a self-organizing particle system was developed to thoroughly study the mathematical model of BOB bots. Using ideas from probability theory, statistical physics and stochastic algorithms, the researchers were able to prove that the theoretical model undergoes a phase transition as magnetic interactions increase – an abrupt transition from a dispersed state to aggregation into large compact clusters, similar to the phase changes we see in ordinary everyday systems. such as water and ice.
“Rigorous analysis not only showed us how to build BOB bots, but also revealed the inherent robustness of our algorithm that allowed some robots to be malfunctioning or unpredictable,” notes Randall.