A swarm of autonomous robots that avoids collisions and bumps?
It seems that the new algorithm developed by researchers from Northwestern University can bring a real breakthrough, significantly improving the operation of autonomous vehicles or automated warehouses.
If autonomous cars are to become a reality someday, then manufacturers must work hard on their safe and error-free navigation, without bumps or causing unnecessary traffic jams. Fortunately, they are not alone in this, because research centers from around the world are involved in similar projects - for example, Northwestern University, whose scientists have just developed the first decentralized algorithm with a guarantee of freedom from collisions and stagnation.
Scientists tested their algorithm on a simulation with 1024 robots and a real swarm of 100 robots in the laboratory. We are talking here about the so-called swarm robotics, a concept in which instead of one robot we have a whole swarm that solves problems together. As an example, the search and moving of a large object is often given here - robots go in different directions and when one finds a purpose, it transfers information to others and the whole group gets to work together. In the case of autonomous vehicles, the matter looks a bit different, but cars certainly need to master efficient and safe movement in such a "swarm".
During the tests, the scientists achieved very good results, because the robots reliably, safely and efficiently began to form a predetermined shape, which took them less than a minute. According to one researcher, Michael Rubenstein: If you have many autonomous vehicles on the road, you do not want them to collide or create congestion. Understanding how to control our robots to create specific shapes will help us understand how to control a fleet of autonomous cars when they interact with each other.
Researchers also emphasize the key importance of decentralization, because when control rests with the hands of one robot or swarm, but managed by one individual, problems occur more often: - If the system is centralized and the guide robot stops working, the whole system falls. With decentralization, there is no one leader who tells others what to do. Each robot makes its own decisions. If one of them falls, the swarm can still do its job.
And since robots still need coordination to avoid collisions and stagnation, the algorithm sees the surface under the robots as a grid, and thanks to GPS-like technology, each robot is aware of its place on it. So before making a move decision, it uses sensors to communicate with its neighbors, thus checking whether nearby grid areas are free or occupied. - The robots refuse to move until the space has slowed down and they are not sure if another robot will take it. They are careful and reserve their place well ahead of time. This algorithm can be used in fleets of autonomous vehicles or automated warehouses, where hundreds of robots perform a similar task - add scientists.