Three-axis tactile sensing has advantages for grasping an object of unknown mass and hardness. We developed a new three-axis tactile sensor that possesses a simple structure to endure a large applied force from a powerful grasp. Normal force distribution is measured based on grayscale values obtained by image data processing, as with previous three-axis tactile sensors. Tangential force distribution is determined by the linear movement of image data calculated by optical flow. The sensing characteristics of this sensor are dominated by the configuration and material of fine conical feelers formed on a silicon rubber sheet. By UV-LIGA, we obtain a fine mold of a silicon rubber sheet. In evaluation experiments, we applied both normal and tangential force to the sensor and confirmed this tactile sensor's ability to acquire normal and tangential forces. In its design, we utilize a USB microscope that has a CMOS camera and a light source. In a series of experiments, we performed normal and tangential force tests to obtain its basic characteristics. The linear relationship between the grayscale value and the normal force is obtained from the normal force test. If the average optical flow is under 0.2 mm, the tangential force is proportional to the average optical flow. The slope of the relationship between the tangential force and the average optical flow increases with additional normal force. Finally, we derive a series of equations for three-axis force calculation.

This paper proposes a population control architecture for a multiple robotic system inspired from biological immune networks. The authors' architecture organizes suitable population balance against dynamic environment independently of any centralized control, while each robot decides its action without a global world model. This group behavior, suitable population balance, is organized as side-effect of the interaction of the individual robots in the world.

A rapid, stable decoupled algorithm for solving semiconductor hydrodynamic equations is proposed. First, the Newton iterations to solve each set of equations are omitted. This causes no change in obtained solutions nor in the block convergence. Second, in place of the exact Jacobian matrix, a reduced Jacobian matrix is constructed and used to solve the energy balance equation. This reduced-Jacobian-matrix method is applied to two different hydrodynamic formulations and shown to be very effective, in both formulations, in accelerating the block convergence speed of the decoupled algorithm. The method provides an efficient version of the decoupled algorithm, which is very simple and applicable to a large variety of hydrodynamic formulations.>