UCLA Researchers Developed A New Deep Learning-Based Framework That Allows A Robot To Tackle Paper Folding And The Oriental Art of Origami

Everywhere we go, we come into contact with pliable, slim structures. Large deformations are a common feature of these structures when subjected to even relatively weak forces like gravity. Humans have an astonishingly deep, inherent awareness of the dynamics of such malleable objects. Getting robots to act with more human-like intuition is still a major area of study since it could lead to a wide range of useful applications for business and society.

It is not easy for robots to manipulate deformable objects because they need to predict how the object will change as it is being manipulated to succeed. However, there are presently few solid answers for the robotic manipulation of many other deformable things because most previous research has concentrated on either fabric or ropes.

Recently, a group of researchers at the University of California, Los Angeles (UCLA) developed a novel computational framework that enables a robot to take on paper folding and the Asian art of origami.

Two of the most important investigations on this subject were previously carried out by research groups at Aalto University in Finland and Bielefeld University in Germany. Their first study dealt with textiles, which are computationally easier to manage than paper. In contrast, paper is folded in the second using a complex robot system involving human-like manipulators.

The UCLA team was inspired to carry out this study due to the deficiency of simple and efficient robotic paper folding systems. Consequently, the group set out to design a straightforward but potentially useful device that could fold paper using a single robotic manipulator.

The researchers introduce a robot control technique that teaches robots behaviors from a physical perspective, allowing them to take on jobs requiring physically insightful manipulation more easily. And more particularly, they used offline environments to train artificial neural networks (ANNs) using paper-folding physics simulations. Throughout its training, the network became familiar with the “behavior” of a sheet of paper when held in various grips.

The training data was produced through mathematical and physical modeling on a computer. Subsequently, the trained neural network made quick predictions online and in real time, leading to optimal manipulation trajectories. Scaling analysis, borrowed from mathematics, is used to nondimensional the neural network’s predictions, which is another first.

Non-dimensionalization is a mathematical physics technique that eliminates the need to worry about the units of measurement between input and output. There are no units for the non-dimensionalized quantity. Therefore, changing the system’s units will not affect the analysis. It enhances the control framework’s generalization, making it possible for the robot to fold sheets of paper with varying thicknesses and geometries without separate training.

The “dimensionality” of the paper folding problem can be reduced through non-dimensionalization. In other words, it facilitates training while enhancing the real-time performance of the neural network.

One interesting result of this research is that physics-based scaling analysis and machine learning algorithms work together quite well for manipulating deformable objects with robots. When dealing with paper, for example, the computing expense of using a traditional mathematical model of physics is intractable, making real-time manipulation impossible. However, suppose machine learning is used without prior knowledge of the problem. In that case, a control scheme will be created that will only be effective for situations that match those in the training data.

According to the researchers, this framework is the first to utilize this synergistic method. They hope their study will be widely applied in various deformable manipulation tasks such as cable management, knot tying, robotic kirigami, etc. They plan to broaden their focus to include more advanced folding activities like robotic origami. Making it possible for a robot to fold paper into various shapes—paper aircraft, paper frogs, and so on—would be an intriguing endeavor.

Check out the Paper. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our 13k+ ML SubReddit, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.