How Robots with Stiff Necks Can Still Play Tag Without Crashing
Imagine you are trying to play a game of Tag with three friends, but you are all wearing stiff neck braces. You can’t turn your head to see who is behind you.
If you can’t see your friends, how do you stay in a perfect formation without crashing?
The "Field of View" Problem
This is the "Field of View" (FOV) problem. Usually, scientists pretend robots have eyes in the back of their heads—like a "circular sensing range," or a bubble of sight that goes 360 degrees around them.
But real robots, like the 4 DJI Matrice 100 drones used in a recent study, usually have cameras that only point one way. This creates a "directed interaction graph," which is like a one-way street for information. If Robot A can see Robot B, but Robot B can't see Robot A, the whole group can get wobbly and fall apart.
The "Glow-Up" for Robot Brains
To solve this, a team of researchers came up with a "glow-up" for robot brains. They didn't just tell the robot where it was; they added "virtual points" to its map.
Think of these like invisible "stunt doubles" or ghost points that represent the very edges of what the robot can see. By tracking these ghosts, the robots can stay perfectly balanced.
Pratik
Mukherjee
This allows us to derive a modeling of the multi-robot system with limited field of view, which we will refer to as the extended system that is amenable to the theoretical framework in [3].
Testing in the Real World
The team took their 4 drones to a 30m x 20m indoor space and then out into the wild at the Virginia Tech Drone Park.
Surviving the Wind
Even when 10 mph winds tried to push them over—creating "jitter," which is like a robot shivering because it's nervous—the drones stayed steady.
How the "Extended System" Worked
The "Extended System" worked. The robots used a "potential-based control," which is like an invisible magnetic force field. If a drone got too close to the edge of its friend's sight, the "force field" pushed it back into the right spot.
Result
Scientists saw that the drones could fly super close to each other without ever hitting.
Resilience
They survived the wind and the "noise" (which is like static on a radio that makes it hard to hear) from their GPS-like tracking system.
Unsolved Mysteries
The researchers used 4 drones, but they don't know if this trick will work for a "swarm" of 100 or 1,000 robots.
Also, they used a special radio system called UWB to help the drones "see" each other's position. In the future, they want to see if the drones can do this using only their own camera "eyes," even when those eyes can't turn.
Source: "Experimental Validation of Stable Coordination for Multi-Robot Systems with Limited Fields of View using a Portable Multi-Robot Testbed," Pratik Mukherjee, Matteo Santilli, Andrea Gasparri, and Ryan K. Williams (2019).