Study On Bimodal Noises In Electronic System

In recent years, non-Gaussian noise is one of the popular task in advanced signal processing. And it has been receiving more and more study all over the world. Yet, the concrete mathematical model has not been given out generally. Consequently, the study is qualitative. And the method has profound & widespread significance. However, it may be not the optimal method under some concrete circumstances. Bimodal noise is one simple, hybrid noise consisting of two kinds of additive noises. As a whole, it belongs to non-Gauussian noise. The three mathematical models of bimodal noise are typical & representative in this paper. And they cover Gaussian noise & some non-Gaussian noises. Therefore, the methods in this paper also have some general significance. For example, the general method removing interference is equalization in Gaussian noise with additive inter-symbol interference, that may lead to phenomenon of erroneous code increment. While the method in this paper is based on maximum likelihood criterion, which ensures accuracy of signal receiving.Generally, the method used to study non-Gaussian noise is using higher-order statistics(HOS). While many problems have been solved by two-order statistics in this paper.This paper gives some perfection & complement to the theory of communication & signal processing based on Gaussian noise. Main content of this paper is as follows:1. We analyze the anti-jamming of some communication systems in bimodal noise in order to illustrate the characteristics of bimodal noise.2. We analyze the statistical characteristics of bimodal noise thoroughly, that provides theoretical basis for models transformations & precise estimations of noise & signal parameters.3. Bimodal noises have three models mainly. Each of them has virtues & defects. We give out the algorithms of models transformations in order to study conveniently. And we test & verify their validity by analyzing the performance of detection & precision of estimation, thus demonstrating that the thought of bimodal noise is practicable.4. We demonstrate the existence of bimodal noise really by studying narrowband bimodal process in detail, that provides the possibility for evaluating the performance of non-coherent detection in communication system with bimodal noise.5. On the detection of signal in bimodal noise, the study of one model is advanced abroad, and the performance of detection is given out by signal-to-noise ratio; we study the three models synthetically, and give out the performance of detection by symbol error rate.6. We study the detection of undulating parameter signal in bimodal noise in detail.7. On the estimation of signal & bimodal noise parameters, some foreigners give out approximate estimation, which is applicable only on some condition; We give out the precise estimation.8. We give out the coordinative estimations of signal & noise. This method is applicable for all additive white noises. And it can be used to identify signal simultaneously. It has some general significance.9. By contrast with Gaussian noise, we give some suggestion for general methods of signaldetection & estimation, that shows bimodal noise is more representative.10. Imitation results show that the algorithms & conclusions in this paper are correct.11. We put forward tentative idea of multi-modal hybrid noise, which provides mathematical model for studying more general circumstance.