From automatic vacuum cleaners to drones that deliver packages, robots are becoming more and more ubiquitous in our daily lives. As technology advances, so does our ability to handle complex tasks. They are beginning to perform tasks that were once limited only to human capabilities.
One such task is grasping an object in a dynamic and unpredictable environment, such as picking cherries from a tree. Branches are unstable, winds are unpredictable, and cherries are small objects for robots. This is a very difficult task for robots accustomed to operating in environments with fixed surfaces, such as factories where certain objects come through a stable band.
Fine manipulation of small objects It is a difficult task for robots due to perceptual error, sensor noise, and the inherently dynamic nature of the problem. On the one hand, it is a ubiquitous task in many sectors, such as manufacturing, healthcare, and agriculture, and automating it can have enormous practical and economic value.
If we think of robots that perform predetermined tasks, like robots used on factory assembly lines, we can design specific hardware for specific tasks. By analyzing the assembly process and required tools, engineers can develop robot designs that efficiently solve the problem at hand. This approach works well because the robot is not intended to be used in other factories and the objects it interacts with do not change within the factory environment. But when we try to come up with a universal solution, things change.
Suppose we need to develop a robot that can grab objects without limits in different environments. I know the environment and objects will be dynamic.Is it still possible to develop a robot that can finely grasp objects without a stable support? This is the question the authors asked and they came up with cherry bot.
cherry bot is a dynamic system for fine manipulation that learns its behavior by pre-training on approximate simulations and fine-tuning with model-free RL in the real world. It is designed to be robust against recognition errors and sensor noise, yet accurate enough to successfully handle the task. In addition, it can handle dynamic scenarios such as environment changes, object movement, etc. It also generalizes well to objects with different sizes, shapes, and textures without requiring specific hardware.
cherry bot Leverages imperfect information accessible to most robots, such as imprecise simulators and heuristic-based baseline policies, to bootstrap surprisingly sample-efficient RL training for real-world operations . Well-dynamic training tasks are designed to minimize human effort in the training process and enable more robust policies. Action Space is designed to efficiently balance learning manageability and responsiveness. The system accommodates plug-and-play recognition modules and is designed to adapt to a wide variety of objects and scenarios.
cherry bot Uses common hardware. Assembled robot arm and chopsticks. that’s it. Chopsticks are used for fine manipulation. Robotic arms aren’t perfect either. In some cases, inaccurate sensor results may be provided. Despite these shortcomings, cherry bot After just 30 minutes of real-world interaction, it demonstrates superhuman responsiveness in a dynamic, high-precision task (grabbing a slippery ball swinging in mid-air with chopsticks).
check out paperdon’t forget to join Our 19k+ ML SubReddit, cacophony channeland email newsletterWe share the latest AI research news, cool AI projects, and more. If you have any questions about the article above or missed something, feel free to email me. Asif@marktechpost.com
🚀 Check out 100 AI Tools in the AI ​​Tools Club
Ekrem Çetinkaya has a Bachelor’s Degree. He completed his master’s degree in 2018. In 2019, he graduated from Ojegin University in Istanbul, Turkey. he wrote his master’s degree. A paper on image denoising using deep convolutional networks. He is currently pursuing his Ph.D. He holds a degree from Klagenfurt University in Austria and works as a researcher for the ATHENA project. His research interests include deep learning, computer vision, and multimedia networking.
