Researchers at the Université Libre de Bruxelles have developed a robotic system capable of morphing to suit the task at hand. The ‘mergeable nervous system’ is made up of separate robots that can detect and replace faulty units.
The notion of separate robots coming together to form a sum greater than their parts has always appealed to fans of science fiction. But advances in robotics and autonomous systems have edged us closer toward the reality of swarm technology.
The robotic nervous system developed in Brussels is made up of components capable of assuming and ceding control, forming different shapes to suit the circumstances and breaking off with broken companions.
It relies on a combination of self-determination and a central control. The overall robot can be seen as a series of units coming together to form a single ‘brain’ with a single purpose. The process of ceding control only occurs when one unit connects to another, so each individual bot can act independently until it becomes part of the whole.
Lead author of the study, Marco Dorigo, has labeled this unique control method a “mergeable nervous system”. As shown in the video below, robots illuminated in red are assuming control.
Read more: Robot tax could ease drawbacks of automation
The applications of swarm technology
For an idea of what can be achieved with smart robotic swarms, we need only look to the natural world for inspiration. Everyday insects, such as ants or bees, are able to achieve incredible feats by working as a single unit.
Speaking to Popular Science, Dorigo said that this kind of robotic teamwork could help robots think on their feet and problem-solve more easily. “Take moving on a very rocky terrain, for example. One alone would get stuck, but attached to each other, they become more stable and they can move on the rough terrain.”
The research paper suggests that the mergeable nervous system could transform the way we think about robotics and develop solutions for specific tasks. “Building on the Mergeable Nervous System method,” it says, “robots of the future will display a new type of adaptivity by autonomously choosing appropriate morphologies for the tasks and environments they encounter.”
Although this technology proves that robots can be adaptable, the research team recognizes that understanding exactly how to adapt to fit the situation is the real challenge. “To solve the same problem on the fly, we might be able to rely on ever increasing computing power and advances in evolutionary computation techniques.”
The study concludes that similar robotic nervous systems could eventually replace the need for task-specific bots. A crate of these adaptable units could turn up and shift into whichever shape best solves the problem at hand. “Our vision is that, in the future, robots will no longer be designed and built for a particular task. Instead, we will design composable robotic units that give robots the flexibility to autonomously adapt their capabilities, shape and size to changing task requirements.”