Parallel Static System SNR Research And Local Optimal Processor Design

The weak signal detection technique has potential applications and wide developmental foreground. A significant aim of weak signal detection is to improve the system output-input signal-to-noise ratio (SNR) gain. In this thesis, we focus on the stochastic resonance effect in the field of signal detection, and study the system output-input SNR gain by increasing the number of parallel array of systems.First, we assume the system input signal is weak, and deduce the expression of the system output-input SNR gain of a parallel array of static systems by the Taylor expansion. Based on this expression, it is demonstrated that the output-input SNR gain, without relation with the input signal, is only determined by the noise and the transfer function. Secondly, we simulate the array of static systems and calculate the output-input SNR gain on a local optimum processor, as the input signal is a sinusoidal signal and the internal and external noises are uniform noise or Gaussian white noise. We present a formula of local optimum processor. According to the simulation results and theoretical analyses, the output-input SNR gain can be enhanced upon increasing the number of systems and the internal noise level. The phenomenon is very meaningful in the weak signal detection. Thirdly, we design the local optimum processor with MATLAB GUI. In order to study the issue of systems output-input SNR gain, we give a graphical and visual interface that makes the observation of the output-input SNR gain more clearly.