The Research On Modeling And Control Of The Flexible-joint Robot Arm

Compared with the traditional rigid manipulator, flexible joint armcharacterized with the light weight, small volume, low energyconsumption, and it is increasely used widely. In some areas, such as: spaceexploration, high-precision robots and severe environment application.Both modeling and control method flexible joint manipulator have takenattention of many researchers not only in China but in overseas due to thepractical significance.A number of control methods have been purposed after mathematicalmodels of flexible joint manipulator are built at the start pages of this thesis.All is to provide a theoretical basis for flexible joint manipulator inindustrial applications. Core research and main conclusions are as follows:Firstly, using Lagrange dynamics equation to construct a generalmathematical model of flexible joint manipulators, and deducingmathematical model equation of multi-flexible joint manipulators as theformer. All of this study based on relevant domestic and internationalreferences support and analysis the model of Spong.Secondly, backstepping control is used to the model which mentionabove. A brief description of the development of this method and its basicprinciples, and analysis the advantages of flexible joint backsteppingmethod is used, on the basis of the use of backstepping controller isdesigned to control the position of the flexible joint robot research andsimulation validation. Then, for the shortcomings of the previousbackstepping controller design, namely the controlled parameter changesvery sensitive issue, the neural network algorithm method introduced incontroller design and optimize the controller.Thirdly, the state variables of the flexible joint manipulator system isdifficult to measure,to track this problem,this thesis purpose a stateobserver to well observed various state variables of the system. Thenthrough the state variables to carry out flexible joint manipulator nonlineartransformation, the complex nonlinear system control problem is converted into a linear system control problem. Finally using sliding mode controllerdesigned to control the equivalent linear system, and get a good controlperformance.Finally, the use of dynamic analysis software ADAMS mechanicalsystem to establish a virtual prototype model flexible joint manipulator.Then using ADAMS and MATLAB co-simulation, by analyzing the motionparameters and experimental results verify the backstepping method can beused for effective control of the flexible joint manipulator and control togood performance.