Urs Luebbert

An adaptive method for compensating non-linear VCO characteristics using series reversion

Non-linear characteristics of voltage-controlled-oscillators (VCOs) can negatively impact system performance in continuous-wave radar if not compensated properly. Such non-linearities frequently depend on the junction temperature, as well as part-to-part variations and are therefore not explicitly known before runtime. Thus, the need for adaptive compensation algorithms arises. We present a method to find an approximation of the ideal compensation curve via a polynomial fit of the measured VCO characteristic and finding its inverse using series reversion, which results in a computationally efficient algorithm. Applying the series reversion formulae in a straight-forward manner results in undesirable error terms. We present methods to avoid such scenarios. Additionally, the problem of measurement noise and reduced data availability is addressed. Simulation results are presented, showing the efficiency of our method.