Research On Multi-motor Synchronization Control Strategy Based On Adaptive Neural Network

With the continuous development of basic science and application technology,synchronous control system composed of multiple motors has broad application prospects.In this paper,the synchronous control strategy of multi-motor system is studied.Firstly,the research background and significance of multi-motor control are introduced,and the important role of multi-motor control technology in production and life is pointed out.The existing motor control technology is introduced,and the advantages and disadvantages of these methods and their application in the actual scene are analyzed.Then,the control problem of a single motor is studied,and the internal structure and driving principle of the motor are analyzed.The control of the motor is divided into two parts: electric control and motion control.At the same time,it establishes reasonable and necessary assumptions combined with practical application scenarios and establishes state equation since these assumptions.The state equation of multi-motor with multi-input and multi-output is further established by matrix expression.At the same time,the application of neural network in motor control is analyzed,and the basic structure of neural network is introduced.Aiming at the coupling problem in multi-motor control,various synchronization control strategies are introduced and analyzed.In the controller design,the unknown non-linear part of the motor model is estimated by using the neural network.By designing the adaptive law of the neural network,the estimated weight matrix can approximate its ideal value.On the premise that the neural network can estimate the unknown non-linear part of the motor,each order of the state equation is regarded as a simple first-order system by using the design method of inverse control.Each of these first-order systems is designed with a state variable as a virtual control variable,so that each simple first-order system can take the outputs according to the pre-designed state.Aiming at the problem that the approximation effect of the adaptive law of the ordinary neural network is not ideal,the error between the estimated value of the weight matrix of the neural network and its ideal value is estimated by designing auxiliary variables,which guarantees the adaptive speed and accuracy of the neural network from another angle.In order to solve the problem of differential explosion in high order inversion control method,the design method of dynamic surface controller is adopted.The derivative of the virtual controller can be expressed in a concise way by means of first-order differential filtering,which eliminates the problem of differential explosion and simplifies the design of control on the premise of ensuring the stability of the system.Aiming at the dead-time problem in system input,inequalities and alternative controllers are designed to deal with it,so that the design method of dynamic surface controller can still control the multi-motor system to run stably and effectively according to the desired trajectory under the condition of dead-time input.At the same time,by designing a new tracking error state,the real tracking error of the system can fluctuate within the preset error range,which greatly improves the controllability of the system,and makes the performance of the system be limited in advance at the controller design steps.Finally,these methods are verified by simulation one by one,which proves the effectiveness and feasibility of these methods.