Design And Implementation Of Software Module For Synthesizing Custom Distributed Noise Signal

Digital noise signal synthesis methods have good scalability and flexibility,and can be divided into look-up table and transformation methods according to the implementation method.The accuracy of the noise distribution synthesized by the look-up table method is limited by the storage depth,which makes it difficult to cope with scenarios requiring high output accuracy.On the other hand,the maximum synthesis speed of the transformation method is limited by the hardware computation speed,which makes it difficult to cope with scenarios requiring high output speed.To solve these problems,this thesis proposes a new digital noise signal synthesis method,which divides the output of noise into high probability interval and low probabil-ity interval according to the probability interval.High probability interval noise genera-tion is based on look-up table method to ensure output speed.The low probability interval noise generation is calculated in real time based on the transformation method to ensure the output accuracy.Based on this method,a custom distributed noise signal synthesis software module is designed and implemented.The main work is as follows:1.On the basis of the analysis and summary of traditional digital noise signal syn-thesis methods such as look-up table method and transformation method,a new digital noise signal synthesis method for dividing high/low probability interval is proposed,and the implementation scheme of this method is designed.2.According to the implementation scheme,discuss the division method of proba-bility interval,calculate the ultimate output speed of the scheme,analyze the influence of storage space,DAC resolution,custom noise distribution and other factors in the imple-mentation scheme on the final output,and establish the output model for the quantitative analysis of errors.3.Digital noise signal synthesis software module design:designed the parameter setting interface responsible for data display and user input acquisition;The waveform editing interface is designed to edit the custom distributed noise.A data processing module is designed which is responsible for the validity check,logical judgment,conversion and calculation of data.The data model module is designed to store the underlying data model and manage the data and configuration of the software.The driver program is designed to control the instrument hardware.4.The test environment is built,and the function and performance of the software module are tested.The test results show that under the resource occupancy of 12 bits of linear feedback shift register,16 bits of DAC,and 2¹⁵bytes of storage space,based on Gaussian noise with mean value 0 and standard deviation 0.2,using the method of dividing high/low probability intervals can effectively generate low-probability noise sig-nals deviating from 5 times the standard deviation range（5δ）,and the estimation error of the standard deviation is controlled within 2%of its relative value.Based on Gaussian noise with mean 0 and standard deviation 1,with the standard of controlling the standard deviation estimation error within a relative error of 0.05%,Compared with look-up table method,this method saves 64 times of storage space and improves 6 times of output speed compared with transformation method.