A joint research team jointly led CityU of Hong Kong It has developed a new soft-touch sensor with skin-like properties. The robotic handle with a sensor attached at the fingertip can accomplish difficult tasks such as firm grasping fragile objects and threading a needle. Their research provided new insights into haptic sensor design and could contribute to various applications in robotics, such as smart prostheses and human-robot interaction.
Dr. Chen YajingOne of the leaders involved in the study was an associate professor in the Department of Biomedical Engineering (BME) at CityU. The results were recently published in the scientific journal RoboticsTitled “Soft magnetic leather for ultra-fine touch sensor with force self-disconnect“.
Mimic the properties of human skin
The main feature of human skin is its ability to feel shear force, which means the force that causes two objects to slip or slide over each other upon contact. By sensing the size, direction, and subtle change of shear force, our skin can act as reactions and allow us to modify the way we should hold the object steadily with our hands and fingers or how far we should grip it.
To mimic this important feature of human skin, Dr. Dr. Ban Jia, A collaborator from Hong Kong University (HKU), developed a new device, a soft-touch sensor. The sensor is in a multi-layered structure like human skin and includes a specially flexible and magnetic film with a thickness of about 0.5 mm as the top layer. When an external force is applied to it, it can detect the change of magnetic field due to the deformation of the film. Most importantly, it can “separate”, or decompose, the external force automatically into two components – the normal force (the force applied perpendicular to the body) and the shear force, providing an accurate measurement of these two forces respectively.
“It is important to separate the external force because each force element has its own effect on the object. It is necessary to know the exact value of each component of the force to analyze or control the static or moving state of the body,” he explained. Yan you canBME PhD student and first author of the paper.
Improving the accuracy of deep learning
Moreover, the senator has another human-like characteristic – tactile “super-precision” that allows it to locate the stimulus as precisely as possible. “We developed a tactile ultrafine algorithm using deep learning and achieved a 60-fold improvement in contact positioning accuracy, which is the best among the ultra-high-resolution methods reported to date,” said Dr. Shin. This effective ultra-fine touch algorithm can help improve the physical accuracy of the tactile sensor array with the fewest number of sensors, thus reducing the number of wires and the time required to transmit the signal.
He added, “As far as we know, this is the first touch sensor that has achieved the capabilities of self-separation and high accuracy simultaneously.”
The robotic hand with the new sensor completes the challenging tasks
By attaching the sensor to the finger tip of the robotic gripper, the team showed that robots can accomplish challenging tasks. For example, a robotic clutch steadily holds a fragile object such as an egg while an external force tries to pull it away, or a needle threader via remote operation. “The superior precision of our sensor device helps the bionic hand to adjust the contact position when holding an object.” Dr. Shin explained that the robotic arm can adjust the force magnitude based on the power separation ability of the touch sensor.
He added that the sensor can be easily extended to the shape of sensor arrays or even the continuous electronic skin that covers the entire body of a robot in the future. The sensitivity and measurement range of the sensor can be adjusted by changing the magnetization direction of the upper layer (magnetic film) of the sensor without changing the thickness of the sensor. This electronic skin enabled it to have different sensitivity and different measurement range in different parts, just like human skin.
Also, the sensor has much shorter manufacturing and calibration processes compared to other haptic sensors, making it easier for actual applications.
“This proposed sensor could be useful for many applications in robotics, such as adaptive gripping, fine manipulation, texture recognition, smart prostheses, and human-robot interaction. They can make home robots a part of our daily life.
Dr. Shin and Dr. Ban are the paper’s opposite authors. CityU team members include PhD students Yan Youcan and He is my zhi From BME and Dr. Yang Zhengbao, Assistant Professor from the Department of Mechanical Engineering. Other collaborating researchers are from Carnegie Mellon University and Southern University of Science and Technology.
The research was funded by the National Natural Science Foundation of China, the Hong Kong Research Grant Council, and the Shenzhen (China) Principal Fundamental Research Project.