Robots could soon do daily home tasks on their own. A new method helps them move both arms together and adjust when things change.
Researchers at Universidad Carlos III de Madrid have developed a method that allows robots to coordinate their arms more flexibly by combining observational learning with communication between the robot’s limbs. The approach improves how service robots adapt their movements when conditions change.
The method is designed to help robots handle daily household tasks such as setting and clearing the table, ironing, and cleaning the kitchen. By enabling both arms to share information during a task, the system supports smoother and more stable motion.
Teaching robots to perform everyday activities remains a central challenge in robotics. Traditional programming requires defining each movement in detail, often through thousands of lines of code. Imitation learning offers a different path. Instead of coding every step, a human demonstrates the task. This can be done by physically guiding the robot’s arm or by recording the action for the robot to observe.
Through this process, the robot learns to perform tasks such as pouring water or organizing objects on a shelf. However, direct copying is not enough. If a robot learns to grasp a bottle at a fixed position and the bottle is moved slightly, a system based only on repetition may fail because it repeats the original trajectory.
The new technique addresses this limitation by enabling learned movements to adjust when the target changes. The motion adapts smoothly, similar to a rubber band shifting shape while keeping its structure. This allows the robot to reach a new position while preserving key aspects of the task, such as keeping a bottle upright to prevent spilling.
By focusing on adaptation rather than rigid repetition, the method moves service robots closer to operating reliably in home environments where object positions and conditions often vary.
