Model calibration of non-ideal lowpass filter in modulated wideband converter for compressive sampling

Abstract

Purpose – The purpose of this paper is to present a model calibration technique for modulated wideband converter (MWC) with non-ideal lowpass filter. Without making any change to the system architecture, at the cost of a moderate oversampling, the calibrated system can perform as the system with ideal lowpass filter. Design/methodology/approach – A known test sparse signal is used to approximate the finite impulse response (FIR) of the practical non-ideal lowpass filter. Based on the approximated FIR filter, a digital compensation filter is designed to calibrate the practical filter. The calibrated filter can meet the perfect reconstruction condition. The non-ideal sub-Nyquist samples are filtered by a compensation filter. Findings – Experimental results indicate that, by calibrating the MWC with the proposed algorithm, the impaction of non-ideal lowpass filter could be avoided. The performance of signal reconstruction could be improved significantly. Originality/value – Without making any change to the MWC architecture, the proposed algorithm can calibrated the non-ideal lowpass filter. By filtering the non-ideal sub-Nyquist samples with the designed compensation filter, the original signal could be reconstructed with high accuracy.