Realization of Intelligent Grasping on a Traditional Hydraulic Manipulator

Abstract

In the literature, studies on dexterous manipulators are mainly focused on multi-finger multi-joint structures. These manipulators often have difficulties with grasping heavy objects due to the delicateness and complexity of the structures. This paper describes the authors' attempt on implementing dexterous grasping on a traditional two-single-joint- finger hydraulic manipulator. Hydraulic driven manipulators have higher power output. However, dexterous grasping re- alization on this type of manipulators represents a very high technical challenge as it is difficult to measure the changes in the grasping force and the object attitude. In the authors' study, an angle sensor and a force sensor are used to capture the changes in the finger opening distance and the grasping force. An elastomer and photoelectric device based slip sensor is designed to measure the relative movement between the fingers and the object. A fuzzy control algorithm is developed and implemented on a microcontroller unit to serve as the higher-level controller of a two-level hierarchical control sys- tem for the finger movement. The algorithm decides the feeds of the fingers based on the outputs of the angle, force and slip sensors. The lower-level controller is an electrical grasping motion control device that applies the output of the fuzzy controller to control the movement of the fingers. With a 10Mpa hydraulic power rate, the maximum grasping force of the manipulator is over 2500N. Experiments show that this hydraulic dexterous manipulator can grasp both heavy objects and eggs, glasses and other fragile objects.