Reconfigurable Modular Robot Distributed Control

Along with the development of science and technology, the application field of robots becomes wider and wider, the assignment and work environment cannot be forecast, reconfigurable modular robot emerges as the times require robot to work at such unknown assignment and unknown work environment. Reconfigurable modular robot makes up with a set of standards connected interface link module and joint module, these link modules and joint modules can be fleetly buildup in different kinematics parameter and dynamics action to fit different assignment and work environment. Reconfigurable modular robot has can accommodate more work request and has the advantages of lower cost and easy to maintain. The development of reconfigurable modular robot technology expend the application of robot in agile manufacturing, military, aerospace, industry. In this dissertation, the contribution is the distributed control method of reconfigurable modular robot.Module is the reconfigurable modular robot ultimate composition cell, the design of module is one of the most important problem in reconfigurable modular robot research, the design of module directly influence the performance of reconfigurable robot. The research of module comprise module modality, module mechanical structure, control circuit, mechanical & electrical connect interface . Modules can fast connect, and have different length link modules and different mechanical interface, the connection of different modules can satisfy the request of robot configuration different kinematics and dynamic. For simplification the robot system structure, link modules and joint modules concept models are designed in this paper, and built module configurable depict coordinate system.Research in the robot kinematics and dynamics direct relation of the robot design and control: the research aim to rapid calculation the robot dynamics model. Robot kinematics depict the motion relationship of the joint and the rigid body which buildup the robot, that's a very important problem in robotics. Once known the robot kinematics equation, the inverse kinematics is also known, hence ascertain the driving force to get along the setting contrail. If already establish the robot dynamics model, hence the force moment driving each joint is ascertain.Reconfigurable modular robot control is a very challenges research problem. From superior century the research of reconfigurable modular robot began, while realize the mechanical structure quick reconfiguration, the research in control system has done a lot. On the one hand base traditional controller and build fast reconfigurable control software & algorithm, this method still adopt central control. On the other hand increase the function of module, consequently the module can do more assignment alone. So people begin to research part and complete distributed control method.This dissertation introduce distributed control method, each joint directly communicate with each other, transfer the information of each other(joint speed, force), because of there exit time delay etc. uncertainty factor, cause each joint submodel exist uncertainty, Variable Structure Control is a good idea to solve this problem. This control method is different from other control methods for its system structure is unfixed, its structure can base system's current state(deviation & derived number)change itself state in dynamic procedure, compel system according to the trajectory of destine slide modal move, so it is called Sliding Mode Control. Because variable structure control system is absolutely adaptable to disturbance, it is attached important, and use it in the robot control. SMC control method has the advantage of controller simple, easy achieve, reduced order, decoupling, ensure system state error astringency to nil, and has robustness & inflexibility to modeling error & external disturbance. Because of such advantage, design slide adaptive controller, achieve three joints reconfigurable modular robot contrail control simulation, and identify the control algorithm. For the uncertainty in the submodel, using RBF NN can approach random continuum function at will, in both assume know the structure error of NN and assume unknown the structure error of NN, design RBF NN controller, realized estimate the uncertainty in the submodel, and using two articular reconfigurable modular robot numerical value emulator, and identify the control algorithm.