Coordinating Control Of The Dual-stage Actuator System

As the development of economy and science, the requirement on the quality and accuracyof the products with high technologies are increasing. Thus, the single-stage actuator systemgradually has been replaced by the dual-stage actuator (DSA) system. The DSA system canperform well in solving the contradiction between the large stroke and high precision.The first actuator (coarse actuator) is of long traveling range but has poor accuracy and slowresponse time. On the contrary, the secondary actuator (fine actuator) is of great higher precisionand faster response but has a limited travel range. Anyone of the two actuators can becompensated by merit of the other. They together control the system. Therefore, the DSA systemhas the characteristics of large traveling range, high accuracy about positioning and fastresponse.The control structure for DSA system we select for research in this paper is a decoupledstructure. Thus, controllers with two actuators can be separately designed and they are thegain-optimal of dynamically damped proximate time-optimal servomechanisms and PIDself-tuning controlled based on RBF neural network, respectively.The controller of the first actuator is the gain-optimal of dynamically damped proximatetime-optimal servomechanisms. The controller is based on proximate time-optimalservomechanism and a more aggressive control law is applied to replace the original control law.In this paper the Composite Nonlinear Feedback is taken as the more aggressive control law. Butthe gain in this control law will affect the control effect, so the particle swarm algorithm is usedto optimize the gain to obtain the best value to improve the control effect of the DSAsystem.The controller of the second actuator is PID self-tuning controller based on RBF neuralnetwork. For the reason that neural network has good ability of nonlinear mapping, learning,adaptation and need not mathematical model in accuracy. In addition, it has the ability to applywhat it learned in new area of knowledge and has a faster convergence. Therefore, the controllercan automatically adjust the parameters of PID to obtain the best value and improve the controleffect of the DSA system. In this paper, particle swarm algorithm is applied to optimize thecoordination factor of the DSAsystem to get the best value. Due to the usage of piezoelectric and magnetostrictive materials in the second actuator, ithas the characteristics of hysteresis affecting the control effect. In this paper the Bouc-Wenmodel is applied to describe the hysteresis. Inverse multiplicative structure is used to compensatehysteresis nonlinearity in order to improve the control effect.A coordination factor of the DSA system is designed in this paper. Performances about twoactuators can effectively be assigned by the coordination factor. When the error in system isbigger, the coordination factor tends to zero and the first actuator play an primary role in system.When the error in system is smaller, the coordination factor tends to one, the second actuatormake an primary role in system. In the paper, a best adjustable parameter in the coordinationfactor could be obtained by a trial and error method.