Title Research And Design Of Bp Neural Network Weak Signal Detection System Based On FPGA

With the development of science and technology,the technology of weak signal detection has been integrated into daily life.The weak signal detection has used the way of eliminating the noise and improving the noise ratio of the output signal to extract signal which is widely used in industrial production and agriculture.Weak signal detection in strong noise background is always a difficult problem in engineering application.The traditional linear detection method will produce the threshold effect when detecting the weak signal in the strong noise background,which may cause the sharp deterioration of the output signal to noise ratio and loss of the signal,so its detection methods need to be further improved.Based on the weak signal detection in the strong noise background,aiming to solve the problems with threshold effect of linear detection and signal loss,this thesis uses the nonlinear relationship between BP neural network and noise,then design the detection system of weak signal,which improves the detection effect in strong noise environment,and solves the signal distortion problem of the traditional detection method.Firstly,in order to solve the problem that the measured signal does not match with the input of BP neural network,the phase space reconstruction method is used to reconstruct the phase space of the measured signal.Then,we use the reconstructed temporal space vector as the input sample of BP neural network,and the training samples are trained by learning algorithm of BP neural network.Finally,the BP neural network algorithm is used to control the detection error,and the design of weak signal detection system based on BP neural network is finished.In this thesis,we use the VHDL language and FPGA technology to design the weak signal detection system with BP neural network.According to the top-down design idea,the system is divided into signal vector space reconstruction module,signal modulation and demodulation module and BP neural network detection module.Finally,we use the mixed signal of periodic signal and noise to verify the feasibility and practicability of the design.The experimental result shows that the system can detect the weak signal in strong noise environment.Compared with the traditional scheme,to some extent,the detection threshold can be improved and the signal loss rate can be reduced.The BP neural network weak signal detection system based on FPGA designed in this thesis not only has better detection effect on the signal under strong noise interference,but also reduces the loss of the signal The system has the advantages of high detection precision,stable detection performance and fast operation speed,which has a good application prospect in the fields of aviation,military and industrial control with high performance requirement.