A Novel Calibration Method of Gain and Time-skew Mismatches for Time-interleaved ADCs Based on Neural Network

This paper proposes a novel calibration method for time-interleaved analog-to-digital converter (TI-ADC). Based on the polyphase derived operation, the spurious signals due to the gain and time-skew mismatches could be compensated with a simple BP neural network (NN). In the proposed calibration structure, a pilot signal is sent to train the neural network and then the sampled signals could be directly calibrated with the trained network. The main advantage is that the proposed method could achieve higher performance without requiring a feedback loop in comparison with the prior works. Numerical simulations have been presented to demonstrate the effectiveness of the proposed structure, which shows the proposed method could significantly improve the SNR performance of the TI-ADCs.