Signal recovery in time-interleaved analog-to-digital converters

Abstract

Calibration is a serious challenge in the design of high-speed time-interleaved analog-to-digital converters (ADC). We develop an iterative blind calibration technique for such converters. In particular, an expectation-maximize (EM) algorithm is used to estimate the associated unknown gains and time-offsets, from which the calibrated signal is reconstructed. Tradeoffs between the calibration time, reconstruction quality, and the oversampling factor are developed. The proposed algorithm can also be used in a variety of other applications, including problems of distributed sampling in sensor networks.