This work deals with a pneumatically actuated hand composed of four fingers and an opposable thumb. Mechanisms to convert actuator motion into phalanx rotation were studied to make each finger as similar as possible to the human specimen. Force tactile sensors are disposed along the phalanxes to allow a closed-loop force control, while the thumb position is sensorized by a potentiometer and hence position controlled. Each component is controlled by fuzzy logic. This solution allowed the existing strong nonlinearities to be easily managed. A fuzzy supervisor applies a grasping strategy whose target is an approximate identification of the shape and size of an object to grasp it most efficiently with the disposable fingers.
展开▼
