A self-organising swarm of miniature robots can mimic the way flocks of starlings or hordes of ants cooperate, say researchers.
Engineers in the US have created a 1,000-strong army of miniature robots – dubbed the Kilobots by the researchers – equipped with infra-red sensors and two vibrating motors that allow it to slide across a surface on its three pin-like rigid legs.
Each measuring just a few centimetres across, the robots are given no direct information about their global position but collectively construct a co-ordinated system capable of self-assembling into a variety of shapes – such as a five-point star or letters of the alphabet – using a set of fixed rules about how to interact with neighbouring robots.
The project, which is described in a paper in today’s issue of journal Science, was designed to demonstrate how complexity can arise from very simple behaviours performed en masse in the same way that starlings flock together in cloud-like ‘murmurations’ or ants create rafts and bridges with their bodies.
“The beauty of biological systems is that they are elegantly simple – and yet, in large numbers, accomplish the seemingly impossible,” said Professor Radhika Nagpal, from Harvard’s School of Engineering and Applied Sciences. “At some level you no longer even see the individuals; you just see the collective as an entity to itself.”
In February, Nagpal’s team created self-organising termite-inspired robots able to construct a variety of structures without supervision or detailed instructions, but the team has been unable to demonstrate the algorithm in a large swarm due to the complexity involved and the cost and labour of fabricating the physical devices.
With the Kilobots the team opted for a much simpler design that results in less reliable performance – the devices have trouble moving in a straight line and the accuracy of distance sensing can vary from robot to robot - but at scale, the algorithm used to control the robots is able to overcome individual limitations and guarantees, both physically and mathematically.
Self-assembly require just a single command, which programmes all robots with a two-dimensional image of the target shape and marks four ‘seed’ robots as the origin of a coordinate system.
Each robot continually transmits messages via infra-red light signals that tell neighbouring robots its X and Y co-ordinates and once the ‘seed’ robots start the process off, it generates a domino-effect of signals that propagate through the rest of the swarm.
The Kilobots can form a five-point star or letters of the alphabet
How each Kilobot positions itself depends on the distance between it and its neighbours, with robots becoming each other's reference points, building up an organised system from local interactions.
Using simple tactics such as following the edge of a group, tracking a distance from the origin, and maintaining a sense of relative location, the robots take turns moving towards an acceptable position.
The Kilobots can also correct their own mistakes – if a traffic jam forms or a robot moves off-course, errors that become much more common in a large group, nearby robots sense the problem and cooperate to fix it.
According to Nagpal the system represents a significant milestone in the development of collective artificial intelligence and could help in the design of future of distributed robotics systems.
“Increasingly, we’re going to see large numbers of robots working together, whether its hundreds of robots cooperating to achieve environmental clean-up or a quick disaster response, or millions of self-driving cars on our highways,” she said. “Understanding how to design ‘good’ systems at that scale will be critical.”
The Kilobots will also provide an essential test bed for AI algorithms designed to control multiple robots.
“We can simulate the behaviour of large swarms of robots, but a simulation can only go so far,” said Nagpal. “The real-world dynamics – the physical interactions and variability – make a difference, and having the Kilobots to test the algorithm on real robots has helped us better understand how to recognise and prevent the failures that occur at these large scales.”
The Kilobot robot design and software is available open-source for non-commercial use and has already been licensed by Harvard’s Office of Technology Development to K-Team, a manufacturer of small mobile robots.