In a previous paper we developed an optical three-axis tactile sensor that can acquire normal and shearing forces to be mounted on a robotic finger. Normal and shearing forces applied to the sensing element were detected separately; when we examined the repeatability of the present tactile sensor with 1,000 loading-unloading cycles, the respective error of the normal forces was 2%. In the present paper, the three-axis tactile sensor is mounted on a robotic finger of three degrees of freedom to evaluate it for dexterous hands. A series of three kinds of experiments were performed. First, the robotic hand touches and scans flat specimens to evaluate the sensing ability of the friction coefficient. Second, it detects the contour of parallelepiped and cylindrical objects. Finally, it manipulates a parallelepiped case put on a table by sliding it on the table. Since the present robotic hand was able to perform the above three tasks with appropriate precision, we expected that it would be applicable to dexterous hands in subsequent studies.

This paper presents an application of a low force interaction method in a control scheme of robot manipulation based on tactile sensing. Our aim is to develop an intelligent control system that can distinguish the hardness of unknown objects so that robotic fingers can effectively explore the object’s surface without altering its physical properties or causing damage. Initially we developed a novel optical three-axis tactile sensor system based on an optical waveguide transduction method capable of acquiring normal and shearing forces. The sensors are mounted on the fingertips of the multi-fingered humanoid robot arm. We proposed a new control scheme applying low force interaction to distinguish the hardness of unknown objects in robot manipulation tasks based on tactile sensing. The scheme utilized new control parameters obtained by calibration experiments using hard and soft objects that enable robot fingers to precisely control grasp pressure and define the slippage sensation of the given object. Finally, verification experiments of the proposed control scheme using a humanoid robot arm were conducted whose results revealed that the finger’s system managed to recognize the hardness of unknown objects and complied with sudden changes of the object’s weight during object manipulation tasks.

This paper presents the application of a six-axis force sensor and a novel optical three-axis tactile sensor to humanoid robot navigation system which is based on contact interaction towards supporting visual-based navigation. The force sensors are mounted on humanoid robot arms to perform grasping in self-localization task to define the robot's position and orientation. The grasping results guided the robot locomotion and avoid it from collision. Meanwhile the optical three-axis tactile sensors are mounted on cooperative two-finger system for object handling tasks. The tactile sensor is capable of acquiring normal force and shearing force. Experiment with hard and soft objects are performed which results revealed good performance of the integrated robotic fingers and tactile sensor system to recognize and grip the objects. The presented control algorithms for both sensors are capable of preventing the probability of damage to the sensors and objects during robust grasping and object handling tasks.

We are developing an optical three-axis tactile sensor capable of acquiring normal and shearing force, with the aim of mounting it on a robotic finger. The tactile sensor is based on the principle of an optical waveguide-type tactile sensor, which is composed of an acrylic hemispherical dome, a light source, an array of rubber sensing elements, and a CCD camera. The sensing element of silicone rubber comprises one columnar feeler and eight conical feelers. The contact areas of the conical feelers, which maintain contact with the acrylic dome, detect the three-axis force applied to the tip of the sensing element. Normal and shearing forces are then calculated from integration and centroid displacement of the gray-scale value derived from the conical feeler's contacts. To evaluate the present tactile sensor, we have conducted a series of experiments using a y-z stage, a rotational stage, and a force gauge, and have found that although the relationship between the integrated gray-scale value and normal force depends on the sensor's latitude on the hemispherical surface, it is easy to modify the sensitivity according to the latitude, and that the centroid displacement of the gray-scale value is proportional to the shearing force. When we examined repeatability of the present tactile sensor with 1,000 load-unload cycles, the respective error of the normal and shearing forces was 2 and 5%