Transforming Human-Robot Interaction with Modular Robot Skin

Feb 14, 2024 | Available Tech

Lawson WongThe future of robotics is highly dynamic, with rapidly increasing scenarios for humans and robots to interact daily. For example, collaborative robots, or “cobots,” assist on factory floors, handling repetitive tasks and allowing a human worker to focus on more intricate and mentally demanding assignments. In the medical field, robotic assistance can help surgeons perform delicate procedures, and exoskeletons and prosthetics can transform mobility.

Each of these scenarios is enabled by the promise of modular robot skin – a technology that enables advanced, real-time tactile sensing solutions, revolutionizing how robots sense and engage with their environments and enabling safe and fluid human-robot interaction.

Traditionally, robotic skins have been plagued by limitations such as wired connections, complex manufacturing processes, and a lack of adaptability. However, researchers at Northeastern University’s Generalizable Robotics and Artificial Intelligence Laboratory (GRAIL) have been hard at work pioneering a new approach. Led by Assistant Professor Lawson L.S. Wong from the Khoury College of Computer Sciences, the team has developed a modular robot skin with a wireless data processing unit that promises to transform how robots interact with their environment and with humans.

Modular Robot Skin

GRAIL’s research in robot tactile perception started approximately four years ago, with the last two years focused on developing modular robot skin hardware. Currently, three Ph.D. students and several other graduate and undergraduate students are involved in robot skin-related research.

Traditional robots rely on vision for safety, but visual perception can be affected by lighting conditions and occlusions. Therefore, physical sensing is much more reliable, dynamic, and safe.

One of the key advantages of modular robot skin is its scalability. Modular skin allows for easy customization, unlike traditional integrated skins, which are manufactured as one piece and are limited in size and shape. By adding or removing modules, researchers can tailor the skin to fit the unique contours of any robot, from industrial arms to prosthetic limbs.

The modular skin is also much easier to maintain and repair. In the event of a sensor malfunction, only the affected module must be replaced, minimizing downtime and reducing costs. This starkly contrasts traditional skins, which often require a complete replacement.

Additionally, while typical robotic skins require wired connections for data transmission and power supply, the modular robot skin technology is wireless and can transmit data in real-time, allowing for safer and more dynamic applications.

modular robotic skinTactile sensing provides an extra dimension of perception to robots and improves the performance. For instance, tactile sensing enables collision detection in human-robot collaboration scenarios. The ability to sense collisions in real-time helps prevent injuries and increases workplace safety. For robots with highly mobile tasks, such as climbing stairs, fine tactile sensing is crucial in achieving balance and whole-body control.

Lastly, enhanced tactile sensing could be applied to applications such as robot pets. The ability to react when you touch them can make the interaction feel much more natural and enjoyable. With the support of AI technology, pet robots could respond to various tactile gestures, such as patting or stroking, unveiling an entirely new way of communicating with robots.

“As researchers, we aspire to challenge cutting-edge technology and expand the boundaries of human knowledge, and robot tactile perception is rich with possibilities,” says Lawson.

Commercialization Plans for Modular Robot Skin

GRAIL has established several technical collaborations with other laboratories, facilitated connections with robotics companies, and submitted patent applications with the help of the CRI. The technology is currently in the application process for a U.S. patent and has provisional protection.

“In the next several years, we will witness widespread market applications for modular robot skin, driven by the development of AI technology and the proliferation of robotic products,” says Shuo Jiang, a Ph.D. student leading the commercialization efforts for the project. “The robot skin system includes not only the skin hardware but also robot control algorithms, computational acceleration, simulator design, complementary mechanical designs, and software supporting the entire ecosystem.”

With the provisional patent, the team is seeking licensing, development, and commercial partnership opportunities within the research and industry settings.

Written by Elizabeth Creason