The Design And Fundamental Property Research Of A Reconfigurable Robot

For the reason of fast changes of modern manufacturing condition and tasks in order to adapt market demands, manufacturing industry needs new reconfigurable manufacturing system which can quickly adapt the changing tasks. As components of it, robot must be also reconfigurable, so reconfigurable robot become a hot topic of robot research. A reconfigurable modular robot system consists of a collection of basic modules that can be assembled into many different kinematic configurations according to the requirement of special tasks. In this thesis, the division and design of the module, the optimization of configurations, kinematics and dynamics of series reconfigurable robot are studied.First of all, this thesis divides the modules of robot based on the functional features of reconfigurable robot and the basic principles of module division, and all modules are designed in concept. The standard connection interface of the basic modules is designed in order to ensure quick and accurate connection to each other of the basic modules.Moreover, the representation of configurations using graph theory is proposed, and the configuration of reconfigurable robot is expressed based on a improved configuration incidence matrix of graph theory. On this basis, genetic algorithm is selected as the method of the optimal design of configurations, and the effectiveness of this approach is demonstrated by a computation example.The forward kinematic model is derived from the concept of modular, and a computation program for this method is designed on the platform of MATLAB. The inverse kinematic model is obtained through the differential kinematics equations and the product-of-exponential formulas, and the Newton-Raphson iteration method is employed for numeric solution.Finally, by considering each basic module as a link of traditional robot, each fixed connection between basic modules as a rotary joint which's rotation angle maintain constant zero, dynamic model is recursive established from Newton-Euler equations, and a computation program for this method is designed on the platform of MATLAB.