Research On Analog Circuit And PID Parameter Tuning Based On Deep Learning

Deep learning is a machine learning process of multiple levels of deep network structure based on the sample data.With the era of big data coming,deep learning technology has become a hotspot research in the field of artificial intelligence.It is widely used in image recognition,speech recognition,natural language processing,behavior recognition and so on.In this paper,the application of deep learning in analog circuit fault diagnosis,chaotic circuit control and PID parameter tuning is taken as the research object.Because the application objects of different deep learning models are different,VGG16 and DBN models are adopted in this paper.In order to solve the problem that the analog circuit is prone to failure and difficult to diagnose,the fault diagnosis algorithm for analog circuits based on deep learning was proposed.A chaotic circuit tracking control method based on deep learning is proposed for the problem that the circuit is prone to chaos and difficult to control.In addition,for the problem that PID parameters are difficult to be fixed,a PID parameter tuning method based on deep learning and improved firefly algorithm is proposed.So,the main efforts of this paper are as follows:(1)Propose a fault diagnosis algorithm for analog circuits based on deep learning.The problem of traditional analog circuit fault diagnosis method,which its fault identification classification accuracy is not high or the number of identifiable types is small,was improved by propose a fault diagnosis algorithm for analog circuits based on VGG16.In the algorithm,the sampled raw data is converted into a phonetic form,and then transformed into speech spectrum by time-frequency domain change,finally sent into VGG16 model for training and testing.And Sallen-Key low pass filter will be used to test effects.The experimental results show that the algorithm can identify nine kinds of fault types with 100% accuracy,which is proven to have a strong capability in fault diagnosis.(2)Propose a chaos track and control method based on deep learning.In order to remove the limitation of the traditional chaos control method that the system mathematical model must be known clearly,a control method based on Deep Belief Network is proposed.In this method,the function of the hyperchaotic system is realized by DBN and a high precision fitting function is obtained.Finally,a controller which is composed of the fitting function and the tracking reference signal is designed to achieve the tracking control of hyperchaotic systems.And the new six-dimensional chaotic system proposed by this paper as a test circuit for experiments.The results of the simulation experiments show that this control method can effectively realize the tracking control of six dimensional hyperchaotic system for any reference signal.(3)Propose a PID parameter tuning method based on deep learning and improved firefly algorithm.In order to solve the problem that PID parameters are difficult to be fixed,a new self-tuning algorithm is proposed by combining DBN and LOGFA.The asynchronous motor is used as the simulation model for experiments.And the results show that the method is more stable than the GA and FA.Final,a GUI interface of PID parameter tuning method based on DBN and LOGFA was designed,and the application possibilities of the algorithm are enhanced.