Apparent distance measurements during a mutual approximation (blue dots on top panels). The polynomial used to fit these observations is typically a fourth order one (left panel). A second order polynomial was also tested, for the whole duration of the event (middle panels) and over a reduced time interval (30 min) centred on the central instant t_c. For each of the three case, the residuals between the fitted polynomial and the true apparent distance history are displayed in the bottom panels.