Three-Dimensional Shape Construction in a Soft Large-Scale Vision-based Tactile Sensor with a Single Rotational Camera

Soft haptic sensing interfaces based on optical methods using cameras are characterized by their simplicity in design with less complication in terms of wiring, maintenance, and so on. Also, thanks to the fact that the mechanical properties of the flexible outer skin used in such vision-based tactile sensor are not compromised, promising in high reliability with respect to noise. On the other hand, when the sensing area increases, two or more cameras need to be setup behind the soft skin for monitoring markers' displacements. Previously, we developed a barrel-shaped tactile sensor with two cameras setup at two ends of the device for extraction of three-dimensional (3-D) movement of markers. However, simultaneously processing multiple images from cameras requires much computational effort. In this paper, we describe a 3-D estimation method using the rotation of a fisheye camera setup inside the soft skin's boundary, active by a simple servo motor. Based on rotation of the camera, the sensing time, sensitivity can be adjusted actively. The proposed method not only improves the usable volume inside the sensor, but also shows more robust measurement performance with respect to deformation of the sensor's outer skin.