As automation rapidly progresses, robot collaboration has moved beyond science fiction. Picture a warehouse where dozens of machines move goods without crashing, a restaurant where robots deliver meals to the right tables, or a factory where robot teams instantly adapt their tasks to meet changing demand.
Open-Source ROS2 Framework Brings Collaborative Robotics to Life
This vision is becoming reality through an open-source framework built on ROS2, which enables multiple robots to collaborate intelligently, flexibly, and safely. The research was recently published in IEEE Access.
Turning theory into practice requires studying how robots learn to navigate collectively. Successful collaboration depends on their capacity to communicate and make real-time decisions. The system incorporates three key elements:
Autonomous navigation: Each robot computes the best routes using GPS-like algorithms adapted for dynamic environments. With simulation tools such as GAZEBO, they first train in virtual settings before operating in the real world. When faced with unexpected obstacles—like a fallen box—they immediately recalculate their route.
Adaptable behavior: The system relies on “behavior trees,” which act like a dynamic set of instructions. For instance, if a robot gets stuck, it will first attempt to turn, then back up, and if the issue continues, it asks the central system for assistance. This method not only avoids collisions but also makes the system scalable—from just two robots in a lab to dozens in a factory.
Computer vision and task allocation: Acting as the eyes and brain of the collaborative setup, this component ensures robots know both their position and their assigned tasks. It combines two key technologies: ArUco markers—similar to QR codes, small printed symbols placed in the environment that serve as reference points—and distributed cameras that track these markers, calculating each robot’s location with an accuracy of under 3 cm.
It’s as though the robots maintain a continuously refreshed internal map. The second technology is smart task assignment: the system dispatches the nearest available robot, much like a courier selecting the quickest route. If one robot breaks down, another seamlessly takes over, ensuring operations continue without interruption.
Simulated Warehouses, Restaurants, and Labs Put Collaborative Robots to the Test
To test the system, researchers simulated a variety of complex scenarios. Into industrial warehouses, robots transported packages between ArUco-marked stations while avoiding traffic jams. In restaurants, machines delivered meals to specific tables, coordinating to prevent collisions in tight hallways. In laboratories, diverse teams—including small robots and robotic arms—collaborated to carry out experiments.
The results were impressive: robots located themselves with an average error margin of just 2.5 cm. The system also proved highly resilient—when one robot failed, another seamlessly took over its task within seconds.
Scalability, often a challenge in robotics, was also demonstrated, as the framework functioned equally well with five robots as with 15, adapting smoothly to different environments.
Because it is open-source and built on ROS2, a widely adopted platform, the system is accessible to any organization. Hospitals could program robots to deliver medications, logistics hubs could optimize package flow, and museums could deploy autonomous tour guides. At the same time, it reduces reliance on humans for repetitive duties, freeing staff for more strategic tasks.
Read the original article on: Tech Xplore
Read more: A Wheel-Legged Robot Moves with Improved Efficiency Across Different Types of Environments