For humans, the "sense of touch" plays a vital role in everyday activities. In tasks such as picking up objects, the sense of touch enables us to feel whether they are hard or soft, light or heavy, warm or cold, and the combination of touch and sight allows us to avoid damaging them. In the rapid development of science and technology today, has given birth to robots that can walk, run, see, speak, listen and other functions. However, research on robotic haptics is still relatively outdated and new technologies need to be developed to overcome some of these problems.
The multi-camera haptic sensor consists of four cameras that are located underneath a soft, transparent material that contains an embedded scattering of spherical particles. The cameras track the movement of these spherical particles, which are triggered by deformation of the material when an external force is applied to it.
The researchers also developed a machine learning (ML) architecture that analyzes the motion of the spherical particles in the material. By analyzing the motion of the spherical particles, the system can reconstruct the forces that cause the material to deform, also known as the contact force distribution.
The researchers explain, "We use a relatively inexpensive camera that generates a large amount of high-resolution image information totaling about 65,000 pixels, which is important for data-driven tactile sensors."
The multi-camera haptic sensor not only provides the total force value that standard force sensors used on most existing robots can, but also provides feedback on the distribution of all the forces applied to its soft surface, thereby decoupling the normal and tangential components. Due to the unique structural design, the new multi-camera haptic sensor has a larger contact surface and thinner structure than other camera-based haptic sensors because it does not require the addition of other reflective components, such as mirrors.
The use of multiple cameras allows for larger areas of arbitrary shape to be covered by this type of haptic sensor," the researchers said. This work also shows how data obtained on a subset of cameras can be transferred to other cameras, resulting in a way to obtain scalable data."
Robotic human-control sensors can be scaled to larger surfaces to create soft and sensible machine skin. The researchers discussed how their machine learning architecture could be adapted and optimized to facilitate its use in robotics in the future.
For the future, the researchers said they now plan to extend the functionality of the sensors to reconstruct information about contact with objects of complex and generalized shapes. "We believe that sensing algorithms should always be developed with the data efficiency component in mind to promote widespread use in robotics, and therefore we will be moving in this direction in our future work as well."
Room B612-59, No. 91 Hanjiang West Road, Xinqiao Street, Xinbei District, Changzhou City, Jiangsu Province, China