Multi-fingerd Dexterous Hand Grasping Motion Planing Research

The rapid development of robot technology was promoted by the progress ofscience and technology and automation of production needs. The design and research ofhumanoid multi-fingered dexterous hand was caused widespread concern, whosestructure and function are very similar with human. To establish motion planning canmake dexterous hand as flexible as a human hand by grasping objects of differentshapes and sizes. The grasping motion planning of dexterous hand was studied in thispaper.According to the structural characteristics of the HAUST dexterous robot hand, thekinematics and statics of dexterous hand was analyzed in the process of grasping. Theworking space of dexterous hand can be obtained by simulation analysis usingMATLAB, which provide a theoretical basis for grasping motion planning. Therelationship between the joint driving force of the dexterous hand and the force appliedto the object was obtained by analyzing statics of dexterous hand.Grasping pattern classification of HAUST dexterous hand was established andgrasping optimal position of contact points was selected. The master-slave graspingcontrol and autonomous grasping control was adopted for the sake of graspingefficiency and complexity. The grasping motion of dexterous hand can be divided intotwo steps. First, from initial point to adjacent contact point, dexterous hand complete thegrasping movement at the master-slave control mode by using data glove which canmapped movement to dexterous hand. Second, from adjacent contact point to contactpoint, the grasping movement of dexterous hand was controlled by single-chipmicrocomputer control system.The grasping motion control platform of dexterous hand which consists of master-slave motion control platform based on date gloves information collection and controlplatform based on single-chip microcomputer was built. Human-computer interactionmeasurement and control system interface was designed using C++Builder in order to achieve the function of dynamic transformation curve display of each finger joint anglesand the fingertip forces, data communication, master-slave control and motion planning.Finger control system based on F040microcontroller was developed to achieve thedynamic detection of the joint angle and the fingertip force, independent to grab controland data communication function. Related experiments and debugging based on thecontrol platform was completed and theoretical study of crawling motion planning wasverified.