Research On PSD Nonlinear Correction Algorithm Based On BP Neural Network

PSD is a new type of light spot position sensitive sensor,it has been widely used because of its fast response,high position resolution,without spot constraints,simultaneous measurement of position and intensity,high reliability and matching of the signal processing circuit structure is relatively simple.However,due to the PSD in the manufacturing process of the surface diffusion layer and the bottom layer resistivity uniformity,effective sensitive area of factors such as nonlinear change causes the output of the PSD device,this defect makes the whole device measuring confidence drop data.To solve this problem,a new PSD nonlinear correction algorithm based on improved BP neural network is proposed in this paper.Firstly,this paper studies the structure and characteristics of PSD,introduces its advantages as a new type of photoelectric position sensor,and analysis of the working principle of PSD,explains the causes of nonlinear and necessity of nonlinear correction.Then the commonly used PSD nonlinear correction algorithm is studied,including the interpolation algorithm and BP neural network algorithm,the comparative analysis of the pros and cons.After that,This paper focus on the principle of two-dimensional PSD nonlinear correction algorithm based on BP neural network,analyzes its disadvantages,from home and abroad for the latest dynamic improvement of standard BP algorithm to explore a practical strong model of improved BP neural network correction algorithm,which is based on thinking about improving convergence speed and reducing the complexity of the trigger.The PSD position measurement platform is built including signal processing circuit and control system based on ARM and CPLD.Experiment data and simulation results by computer programming to complete the validation of the new algorithm through the experiment platform,and with the existing nonlinear correction algorithm are compared,results show that the improved BP neural network calibration algorithm has fast convergence speed and high precision,and the calculation method of the program is simple and easy to fulfill,the new algorithm is more practical.