A domestic research team has developed a skin-attachable strain sensor that can accurately measure the size and direction of skin stretching simultaneously. By applying artificial intelligence (AI) algorithms to the new sensor structure, accuracy and reliability have been dramatically improved, and it is expected to be widely used in the future for electronic skin in rehabilitation, healthcare, and robots.

The Electronics and Telecommunications Research Institute (ETRI) announced on the 8th that it, together with a research team from Chung-Ang University, developed a strain sensor that can detect the amount of expansion and contraction and the direction of deformation in all 360 degrees with 98% accuracy.

Existing skin-attached strain sensors create a stretchable conductive channel by adding elastic materials such as rubber and conductive nanomaterials such as graphene and carbon nanotubes (CNTs), and then detect the size of the deformation by measuring the electrical resistance value that changes as the channel expands and contracts.

However, this structure has the limitation that it can only detect deformation applied in a specific, predetermined direction, and thus cannot accurately measure the characteristics of skin that stretches in various directions depending on the situation, even in the same area.

The ETRI-Chung-Ang University research team has developed the world's first strain sensor capable of simultaneously measuring the size and direction of stretching, and has also succeeded in predicting the size and direction with 98% accuracy by applying an AI algorithm using an artificial neural network structure.

This technology can be attached to human skin to measure human movement, and is expected to be used in industrial fields such as rehabilitation, healthcare, and robotics.

The key is a new directional sensor structure. The research team has completed a patent application for a new structure that exhibits periodic resistance increase and decrease characteristics along the 360-degree stretching direction, spanning a linear stretchable conductor channel between two rigid, non-stretching regions.

Additionally, by placing three sensors adjacent to each other in different directions, the direction of stretch and the amount of deformation of a specific area can be simultaneously extracted by combining these signals.

Previously, a large number of individual sensors were required to recognize various senses, making it difficult to avoid delays due to signal interpretation time. However, by learning and analyzing various sensor data through an artificial neural network algorithm, it was possible to extract the direction of stretch and the amount of deformation with 98% accuracy within a 30% stretch range. A method was presented to simultaneously extract various sensory characteristics by applying an AI algorithm through cognitive learning to the composite signal obtained from a single cross-reaction sensor.

The materials used are harmless to the human body and can be widely used for skin attachment and motion detection of various parts of the human body. ETRI emphasized that it has the advantage of being easy to manufacture through a printing process, low manufacturing cost, short manufacturing time, and can be used as a disposable sensor.

ETRI Flexible Electronics Research Lab Principal Researcher Kim Seong-hyeon said, “The high-precision strain sensor developed by the research team can accurately measure the complex deformation patterns of the skin even with a simple structure, so it can be widely used in fields that require electronic skin, such as rehabilitation therapy and healthcare, robots, prosthetic limbs, and wearable devices.”

Professor Park Sung-kyu of Chung-Ang University also stated, “The developed technology is a groundbreaking attempt that can simultaneously recognize various characteristics with a simple sensor module using an artificial intelligence algorithm and overcome the spatiotemporal limitations of existing methods. It can be widely applied to artificial intelligence-based systems.”

The research team said that they plan to apply this technology to measuring muscle and joint movements under various movements and to use it for diagnosis and ongoing rehabilitation treatment of musculoskeletal diseases, with the goal of commercialization within three years.

Meanwhile, this study was published online on January 5th in the world-renowned academic journal 'Advanced Materials'.