Study On Adaptive Fuzzy Position/Force Control Schemefor Rheonomically Constrained Manipulators

With the development of the robotics, the application in each field of the robotics is more extensive. The motion of robots can be classfied into two general types, free motion and constrained motion. Typical free motion includes spray-paint and spot- welding which position control will be only required. Constrained motion includes wiping-glass, screwing-bolt, polishing, and deburring. Because the end-effector is required to make contact with environment, it is necessary to control not only position but also the interaction force between the robot end-effector and environment. The constraint can also be classified into: constraint to be known and still, constraint to be unknown and still, constraint to be known and rheonomical, constraint to be unknown and rheonomical. According to current research status about compliance control, the research about compliance control of the constraint to be known and still have already achieved certain results.However, in most industry yield, for enhancing work efficiency and making the most of time and energy sources, cooperative method is used. For example, most assembling, work pieces machining, etc, are operated on dynamic platform, which means work pieces are handled on moving assembling line instead of on static platform. So the manipulators should adaptive this dynamic constrained initiatively to do the task accurately. In other cases, when the end-effector is compelled by the force, the manipulator should be able to track according to the force, the force also should be controlled in a certain range. For example, robots shake hands with human whom force the robot, and robots should impose the force and track the motion of human. When the force is changed along with the change of the environment, robots should identify the force through the sensor and layout the on-line target track, then robots track along the moving environment with the force is commendably controlled. It's significative for exploituring the entertainment robots. The research on the manipulators that constrained by rheonomically environment in this paper can solve the problem proposed above.The work of this dissertation is combined with"Study on system dynamics and control of manipulators cooperating a flexible payload", which is supported by China National Natural Science. Based on system dynamics modeling, reducing order dynamics and adaptive fuzzy, this dissertation mainly research on position/force control algorithm for rheonomically constrained manipulators.The dissertation first established the n dimension dynamic model of manipulator, the relationship between the contact force and Lagrange multiplier was educed. Choosing the independent coordinates to reduce the order of dynamics, a new model that only contains independent coordinates and Lagrange multiplier was derived.Secondly, based on reduced order model, according to dynamic model which is totally unknown, combined with the basis knowledge of fuzzy theory, designed the position/force controller and the stability was also proved. But considering the approximate matrix N? 1 may be singular, furthermore, in practical cases, certain information can be gained mainly. So, we can introduce the nominal model, the nominal model can make sure that all the matrix is nonsingular, the other nonlinear part of the model is approached by the adaptive fuzzy. The stability is also proved finally.Finally, simulations were done. The sign function was replaced by saturation function, so the chattering phenomena caused of sign function were solved. The results have confirmed the effectiveness of the controllers.