Statistical techniques to determine of optimal and acceptable noise levels

The paper presents an approach for an application of QoS procedures, on noise impacts in communications, respectively Uniform White Noise (UWN), Gaussian White Noise (GWN), Bernoulli Noise (BN) and Poisson Noise (PN). The approach consists in experimental establishment of recommended optimal and acceptable limit levels of the noise indicator Root Mean Square based (RMS) on the processing of registered information sets for each specific noise. A set of methodological statistical procedures are applied to the experimental data with respect to the complete and individual input sets (RMS levels for each individual noise). Types of family characteristics are analyzed and evaluated about different quality indicators as "Mean Plot of multiple variables", "Normal Probability Plots", "Individual Plots", "X-Bar and R char variable", "Capability Plots" and "Capability Histograms". The approach is also associated with the detection of RMS samples with deviations outside the defined statistical levels, as well as their exclusion in order to improve the quality of the processed information sample.