This correspondence introduces nonlinear regression techniques to estimate
gain and phase mismatches between the in-phase(1) and quadrature (Q) branch
es of a quadrature receiver. Under modest assumptions, the system intrinsic
ally follows a nonlinear regression model. The algorithm is effective, easi
ly implemented, customizable, and requires few assumptions. Large-sample, j
ackknife, and bootstrap techniques provide on-line error assessment and par
ameter inference.