Bio-inspired swarm robots

“Swarm robotics is one of the main applications of collective animal behaviour studies,” says Liang Li. It is designed to solve the tasks that are difficult for individual robots to achieve. However, most current studies on swarm robots mainly solve artificial questions, such as asking a group of robots to swim in a circle or over a thousand robots to line up in a starfish pattern. Moreover, the algorithms and parameters for swarm robots are largely optimized on the basis of simplified mathematical modelling.

As a result, swarm robots that are extensively inspired by collective animal behaviour are still largely unexplored. In the project Bio-inspired swarm robots, Liang Li and his colleagues constructed a platform of swarm robots to apply algorithms extracted from animal systems. “We applied the sensory motor control of leader-follower behaviour to robots and demonstrate the powerlessness of the biological controller,” Li explains. For him, the main innovation is the ability to extract the biological controller and apply it to robotics directly. “With this, we can quantify the control performance. It helps us to understand why the life system evolves out of this control strategy,” he says.

The platform has also been utilized in Liang Li’s semester course Biorobotics, and a general testbed which will help generate and test hypotheses in collective animal behaviour will soon be up and running.