Selecting Criteria for Optimizing Parameters of ADC for Digital Signal Processing
Valeriy I. Akimov, Aleksey V. Polukazakov, Nikolay Sitnikov · 2019 International Russian Automation Conference (RusAutoCon) · 2019
The paper presents the modern criteria analysis for the synthesizing the small-bit ADC and the evaluation of the ADC quantization noise dispersion with linear or nonlinear quantization step when choosing optimal values for thresholds and quantizer levels. The information minimum of the quantization noise (energy) dispersion and the maximum of the signal-to-interference ratio (s/i) parameters is the basis for calculation optimum values. The authors focus on the selection and validation of statistical models for signal and noise parameters. A generalized statistical model of signal and noise distributions in the form of a Nakagami distribution [1] is proposed. It includes the distribution parameter m, which can assume certain numerical values. Different values for the signal ms = 0.5 and for noise mP = 1 being set, the authors obtain different types of distribution for the signal ω(s) and noise ω(i). This allows simplifying the ADC quantizer optimization algorithms for a number of examined criteria. A statistical simulation made possible obtaining the optimum parameters for the nonlinear characteristic of the ADC for various statistical models of a noise and a signal. The program allows analyzing arbitrary s/i ratios at the quantizer input and evaluating this value at the quantizer output. The obtained data provide a means of improving the technical characteristics of “smart sensors” without increasing the digit capacity of their controllers.