At the fields of robotics, many types of sensor that is responsible to shear force as well as normal force have developed. Our improved haptic MCF rubber by combining metal particles in balloon crude rubber has high sensitivity for subtle normal and shear forces, and becomes suitable sensor for convex shape. Therefore, we proposed the technique to read convex shape by the MCF rubber sensor attached on a finger of a robot. The sensor has the response of the electric current flowing inner the MCF rubber to the shear and normal forces. The MCF rubber is also rapped by silicone oil rubber with fingerprint. The sensing depends on the shape of fingerprint. Because of the bending state of the sensor installed on the robot finger surface, the sensing enhances. We also experimentally clarified the response to the convex shape such as Japanese Braille by settling multi electrodes inner the sensor.
This paper is proposing an accuracy improvement method by using the iterative method for the previously established simple and cost effective experimental method to recover the complete 3D shapes of three different objects outer perimeter, a mannequin' s head, a cube and a cylinder that are sprayed with a speckle pattern using the air brush make up. The apparatus are basic camera equipment, a single camera, tripods, pan head and a slider and a double sided calibration board, with check marks on both sides that are aligned with each other and fitted into a photo frame. Rotational and translational movements of the subjects are performed using the pan head and the slider to produce multi-view stereo images. After iteratively calculating the 3D shapes coordinates, alignments of the sets of the 3D shapes are performed using the iterative closet point algorithm method. Finally, accuracy evaluations of the 3D models before and after using the iterative algorithm are performed.