Design And Error Compensation Of Magnetic Encoder

As a position sensor,magnetic encoders are more and more adopted for industrial control,aerospace,medicine and other fields.The high precision and miniaturization of the magnetic encoder have become the focus and development trend of the magnetic encoder industry research.Emphasising the tendency of improving the accuracy and streamlining of the magnetic encoder,this paper designs a magnetic encoder system and studies its error compensation strategy.This paper first designs the physical structure of the magnetic encoder,studies the space magnetic field theory and the decoding algorithm of the magnetic grating,and uses Ansoft-Maxwell for simulation research.Acting in accordance with the application requirements of the system,the hardware of the magnetic encoder is designed.The microprocessor uses STM32F103,and the magnetic sensitive element is i C-MU150.The hardware design scheme of the high precision and miniaturized magnetic encoder system is completed and complete the data test scheme design.According to the design scheme of the system hardware and decoding algorithm,the modular idea is adopted in the software design,and the overall system is divided into a data acquisition module,a data calculation and storage module and a host computer module,and the detailed software design of each module is completed.Software implementation of the decoding algorithm.In this paper,a human-computer interaction interface of the upper computer based on Labview is designed,and the magnetic encoder system sends data to the upper computer interface for display through various communication methods.The hardware circuit,system software and communication with the host computer are carried out during the magnetic encoder to realize that the magnetic encoder can output the absolute angle position stably.At the same time,the data output by the magnetic encoder is used to help determine the performance of speed and accuracy.The experimental results show that the magnetic encoder can achieve stable output at different speed,and the original output accuracy is0.23°.This paper proposes using the PSO(Particle Swarm Optimization)algorithm to optimize the DBN(Deep Neural Network)network strategy to compensate for the magnetic coding error,and use the simulation software to simulate the data set.The results are compared with the traditional BP(Back Propagation)neural network,LSTM(Long Short-Term Memory)network,and DBN deep belief network compensation results are analyzed and compared.The experimental results show that the compensation effect of the PSO-DBN network strategy is the best,and the accuracy after compensation can reach 0.012′,which is greatly improved compared with the original accuracy,which meets the design requirements of the magnetic encoder in this subject.