USC R&D Develops MOTIF Robotic Hand, Overcoming Traditional Robot Hand Technology Limitations

As humans grow, they learn to apply just the right amount of force to push a box and to avoid touching a hot pot with bare hands. Now, a robotic hand equipped with these instincts has emerged.

Image Source: USC Viterbi School of Engineering

According to foreign media reports, the MOTIF Hand, developed by a team of students in collaboration with Daniel Seita, an assistant professor of computer science at the USC Viterbi School of Engineering, is built based on a multimodal concept and possesses multiple sensing capabilities. Among its most prominent abilities are temperature and force sensing, with built-in depth, force, and temperature sensors that enable the robotic hand to perceive and respond to these factors. These features not only open up new prospects for research on robotic hands but also extend the device's lifespan by preventing temperature damage. Its mechanical sensing capabilities have also demonstrated unexpectedly practical value in real-world applications.

"In factory and similar settings, robots need to push objects to target locations, which requires precise force measurement," Seita explained. "This type of force sensor ensures that robots apply the appropriate amount of force." He added, "No one has developed this type of robotic hand before."

The MOTIF robotic hand is built on the LEAP robotic hand developed by a research team at Carnegie Mellon University in 2023. The core breakthrough of MOTIF lies in its newly added human-like perception capabilities. Seita stated that this robotic hand, which possesses more precise human-like characteristics and functions, can be applied in various fields such as factory work, cooking, and welding.

The robot's temperature sensing capability comes from a thermal imaging camera embedded in its palm. Seita and the graduate team from the USC Viterbi School of Engineering are dedicated to creating a hand device that can simulate human temperature perception.

"When encountering hot cookware while cooking, we first use our hands to check if it's safe to touch, to avoid burns," Seita explained. "We hope to translate this intuition into a robotic system."

The co-author Hanyang Zhou of the research paper "MOTIF Hand: A Multimodal Observing Robotic Hand with Thermal, Inertial and Force Sensing" pointed out that this intuitive system requires the hand to be close to the surface of the object to be tested. The paper has been published on the arXiv preprint server.

"We were thinking at the time: is there a way to receive signals without touching anything? So we directly embedded an infrared camera in the palm of our hand," added Hanyang Zhou.

In other words, the MOTIF robotic arm can detect temperature through this thermal imaging camera without needing to touch the object; it can complete the task simply by bringing its palm close to the camera for scanning.

The research of Seita, Zhou, and their team aims to make the testing process for temperature and force perception more natural, in other words, closer to human real experiences of these phenomena. For example, force is something that humans cannot see but can only perceive. The design concept of the MOTIF robotic hand is derived from the core mechanism of human perception, which is the tactile feedback conveyed through the mechanical properties of objects, such as weight, enabling a more realistic robotic mechanical response.

"Humans cannot rely on visual discernment; they must depend on tactile perception. But how can robotic hands achieve this?" Zhou raised the question, "When I am unsure if a bottle is full, I just give it a light tap. I would definitely shake it, right?"

The MOTIF robotic hand integrates an Inertial Measurement Unit (IMU) sensor to introduce this simple test into the field of robotics. This bionic hand, much like a human palm, can determine the weight of an object with just a flick or a shake.

To promote the further development of this perception technology, Seita and his team have committed to open-sourcing the MOTIF robotic hand. "Open-source research results are crucial for advancing academia," Seita stated. "The more people use our robotic hand, the more beneficial it is for research."

Zhou referred to the breakthrough in perception technology of the MOTIF robotic arm as a "platform," hoping that the entire robotics community can continue to innovate on this foundation in the future. "As long as research teams are interested in this platform, we should make it more accessible and easier to use," Zhou emphasized.