Robots with a collective brain: The revolution of shared intelligence

In a world where automation is advancing by leaps and bounds, collaboration between robots is no longer science fiction. Imagine a warehouse where dozens of machines transport goods without colliding, a restaurant where robots serve dishes to the correct tables, or a factory where robot teams instantly adjust their tasks according to demand. This future is possible thanks to systems like the one we’ve developed: an open-source framework based on ROS2 that allows multiple robots to work together intelligently, flexibly, and safely.

From theory to practice, it's essential to research how robots learn to navigate together. The key to robot collaboration lies in their ability to communicate and make real-time decisions. Our system integrates three important features:

Autonomous navigation: Each robot calculates optimal routes using algorithms similar to those in GPS systems, but adapted to dynamic environments. With tools like GAZEBO, robots train in virtual worlds before operating in the real one. If they encounter an unexpected obstacle, such as a fallen box, they recalculate their path instantly.

Adaptable behavior: We use "behavior trees"—a kind of dynamic instruction manual. For example, if a robot gets stuck, it first tries to turn, then reverse, and if the problem persists, it requests help from the central system. This approach not only prevents collisions but also allows the system to scale—from two robots in a lab to twenty in a factory.

Computer vision and task allocation: The eyes and brain of the collaborative system ensure robots know where they are and what to do. The system combines two technologies: ArUco markers—which are like the QR codes of robotics, small printed symbols in the environment that act as reference points—and distributed cameras that detect these markers and calculate each robot’s exact position with less than 3 cm of error. It’s as if the robots carry a constantly updated internal map. The other technology is intelligent mission assignment: the system prioritizes the closest available robot, like a delivery person choosing the shortest route. If one robot fails, another automatically takes its place, ensuring tasks never stop.

To validate the system, we simulated complex scenarios. We used industrial warehouses, where robots transport packages between ArUco-marked stations while avoiding congestion. We also used restaurants, where machines serve dishes to specific tables, coordinating to avoid crossing paths in narrow hallways. Finally, we tested laboratories with heterogeneous teams—from small robots to robotic arms—collaborating on experiments.

The results were compelling, achieving precision where robots locate themselves with an average margin of error of 2.5 cm. The system showed great robustness: if a robot fails, another takes over its task within seconds. Finally, scalability—a key issue in robotics—is well addressed, as the system works equally well with 5 or 15 robots, adapting to the needs of the environment. This framework is not just for robotics experts. Being open-source and based on ROS2, a widely used platform, any company can customize it. A hospital could program robots to deliver medications, a logistics center to optimize shipments, or even a museum to guide autonomous tours. Moreover, it reduces dependence on human operators for repetitive tasks, freeing up personnel for more strategic roles.