Robot watches itself like humans do in a mirror to master movements, self-repair – Interesting Engineering

Robots with self-adaptability provide valuable benefits in real-world applications. 17 hours ago18 hours ago18 hours ago18 hours ago19 hours ago2 days ago2 days ago2 days ago2 days ago2 days ago9 hours ago11 hours ago13 hours ago15 hours ago15 hours ago17 hours ago17 hours ago17 hours ago17 hours ago17 hours agoMrigakshi Dixit A robot observes its reflection in a mirror, learning its own morphology and kinematics for autonomous self-simulation. Jane Nisselson/Columbia Engineering 

 Humans are naturally aware of their bodies and can anticipate the outcomes of their movements. Robotic experts have been working to make robots self-adaptable, which could enhance the operational capabilities of these machines. Now, an autonomous robotic arm has been created that learns its physical form and movement through self-observation via camera. It’s like a robot learning to dance by watching its mirror reflection. Researchers at Columbia Engineering say this new technique allows the robot to adapt to damage, and learn new skills.“Like humans learning to dance by watching their mirror reflection, robots now use raw video to build kinematic self-awareness,” said Yuhang Hu, a doctoral student at the Creative Machines Lab at Columbia University. “Our goal is a robot that understands its own body, adapts to damage, and learns new skills without constant human programming,” Hu added. Robots with self-adaptability provide valuable benefits in real-world applications. Their ability to recover from unexpected failures is vital in areas like manufacturing, home settings, and healthcare.These self-adaptable robots can recover from collisions or damage like a vacuum bumping into furniture or a factory arm being knocked out of place.This leads to increased reliability of in-home robots and reduced downtime and costs in manufacturing — as robots can self-correct without requiring constant human intervention.“We humans cannot afford to constantly baby these robots, repair broken parts and adjust performance. Robots need to learn to take care of themselves if they are going to become truly useful. That’s why self-modeling is so important,” said Hod Lipson, James and Sally Scapa Professor of Innovation and chair of the Department of Mechanical Engineering.The key is a sophisticated AI system using three deep neural networks that mimic brain functions. These networks analyze the 2D video from a single camera and reconstruct a 3D model of the robot’s movement. It helps the robot to build a “kinematic self-awareness” – an understanding of how its body moves, like seeing itself in a mirror.This self-awareness enables robots to identify physical changes and adjust their motions to recover, effectively compensating for damage. In a typical scenario, robot training commonly uses a two-step process: first, simulated environments for initial programming and testing, then real-world application for further learning and adaptation.Engineers build detailed simulations to provide robots with a training ground that closely matches reality, allowing for a smooth shift to real-world applications.However, constructing effective simulators for robot training is a complex and demanding task, mostly requiring the expertise of highly skilled engineers.The new method addresses these difficulties and the high resource demands. The study showcases a robot can autonomously generate its own simulation by simply observing its movements through a camera.“This ability not only saves engineering effort, but also allows the simulation to continue and evolve with the robot as it undergoes wear, damage, and adaptation,” said Lipson in the press release.The new work represents the latest advancement in Columbia’s long-term research on robotic self-modeling. Over the past twenty years, the team has progressively improved the robot’s ability to create self-representations, moving from basic stick-figure models in 2006 to more detailed models using multiple cameras.The findings were reported in the journal Nature Machine Intelligence.Mrigakshi Dixit Mrigakshi is a science journalist who enjoys writing about space exploration, biology, and technological innovations. Her work has been featured in well-known publications including Nature India, Supercluster, The Weather Channel and Astronomy magazine. If you have pitches in mind, please do not hesitate to email her.Stay up-to-date on engineering, tech, space, and science news with The Blueprint.By clicking sign up, you confirm that you accept this site’s Terms of Use and Privacy Policya day ago2 days ago2 days ago2 days agoPremiumIE PROFollow

Source: https://interestingengineering.com/innovation/robot-watches-in-mirror-like-humans