A calibration with an adaptive data selection based on Bayes estimation for a successive stochastic approximation ADC system

A novel low-power high-precision analog-to-digital converter (ADC) called the successive stochastic approximation ADC (SSA-ADC) has been recently proposed, which has two kinds of outputs from a stochastic flash ADC (SF-ADC) and a successive approximation register ADC (SAR-ADC) modes, respectively. In this paper, we propose a software-level calibration and encoding based on the machine learning for the two outputs to generate an total output. In addition, from the practical viewpoint, we propose an incremental learning which selects additional data by keeping a balance between the uniform random and preferential selections based on the Bayesian estimation at each learning.