Watch 2 robots work together to make a hot dog

February 19th, 2020 Posted by Boston University

In this video, graduate researchers Guang Yang and Zachary Serlin teach robots Jaco and Baxter to work together to safely cook, assemble, and serve a hot dog to a human. (Credit: Getty Images)

An advance in machine learning allows two robots, named Jaco and Baxter, to make a hot dog.

Teaching robots to perform complex tasks involves a framework that could apply to a host of tasks, like identifying cancerous spots on mammograms or better understanding spoken commands to play music. But first, as a proof of concept, they’re making franks.

Researchers still don’t fully understand exactly how machine-learning algorithms work. That blind spot makes it difficult to apply the technique to complex, high-risk tasks such as autonomous driving, where safety is a concern. In a step forward published in Science Robotics, Calin Belta, professor in the Boston University College of Engineering, and researchers in his lab taught two robots to cook, assemble, and serve hot dogs together.

Their method combines techniques from machine learning and formal methods, an area of computer science that is typically used to guarantee safety, most notably in avionics (electronics for aircraft) or cybersecurity software. These disparate techniques are difficult to combine mathematically and to put together into a language that a robot will understand.

Belta, a professor of mechanical, systems, and electrical and computing engineering, and his team employed a branch of machine learning known as reinforcement learning. When a computer completes a task correctly, it receives a reward that guides its learning process. Although the steps of the task are outlined in a “prior knowledge” algorithm, how exactly to perform those steps isn’t. When the robot gets better at performing a step, its reward increases, creating a feedback mechanism that pushes the robot to learning the best way to, for example, put a hot dog in a bun.

Integrating prior knowledge with reinforcement learning and formal methods is what makes this technique novel. By combining these three techniques, the team can cut down the amount of possibilities the robots have to run through to learn how to cook, assemble, and serve a hot dog safely.

Source: Liz Sheeley for Boston University

Feeling fond of robots makes for better teams

Trash talk from a robot dings human performance

Robotic gripper is gentle enough to handle eggs

Trash talk from a robot dings human performance

Robotic gripper is gentle enough to handle eggs

September 12th, 2019 Posted by Charlotte Hsu-Buffalo

The robotic gripper attaches to a commercially available robot arm. (Credit: Douglas Levere/U. Buffalo)

A new robotic gripper can alter its grip depending on what it’s holding.

Human hands have remarkable skills that allow manipulation of a range of objects. We can pick up an egg or a strawberry without smashing it. We can hammer a nail. One reason our hands can perform such a variety of tasks has to do with our ability to alter the firmness of our grip.

Researchers designed the new two-fingered robotic hand to do the same thing.

robotic gripper responds to push to avoid breaking the dry spaghetti noodle it's holding — The robotic gripper uses repulsion between magnets to adjust the stiffness of its grip and absorb energy from collisions. This helps improve safety in industrial settings, and prevents objects like this dry piece of spaghetti from breaking. (Credit: Douglas Levere/U. Buffalo)

The design of the robotic hand allows it to absorb energy from impacts during collisions, researcher say. This prevents whatever the robot is holding from breaking, and also makes it safer for people to work with and near the machines.

Such grippers would be a valuable asset for human-robot partnership in assembly lines in the automotive, electronic packaging, and other industries, says Ehsan Esfahani, associate professor of mechanical and aerospace engineering in the University at Buffalo School of Engineering and Applied Sciences.

“Our robotic gripper mimics the human hand’s ability to adjust the stiffness of the grip. These grippers are designed for collaborative robots that work together with people. They’re going to be helpers, so they need to be safe, and variable stiffness grippers help to achieve that goal.”

Magnets are the secret behind the robotic gripper’s versatility, Esfahani says. Instead of having two fingers fixed in place, each of the gripper’s fingers has a magnetic base that sits between two neodymium magnets that repulse, or push against, the finger.

The air gap between the magnets acts like a spring, creating a little give when the hand picks up an object or collides with an external force. User cans also increase or decrease the space between the magnets to adjust the stiffness of the grip.

In one test, the engineers placed a short stick of spaghetti lengthwise between the fingers of the robotic hand. When the gripper crashed into a fixed object, the device detected the external force, which caused the magnets to adjust their position, temporarily reducing the stiffness of the grip and allowing the gripper to absorb some of the energy from the collision.

The end result? The spaghetti stick stayed in one piece.

It’s possible to attach the gripper to commercially available robot arms already in use in many facilities, Esfahani says. That could lower the cost of adapting the technology for companies interested in improving the safety and capabilities of existing machines.

The study appears in IEEE Transactions on Industrial Electronics.

Source: University at Buffalo