Abstract
A fixed-point approach was applied to the problem of estimating freeway origin-destination (OD) matrices given historical sequences of input and output counts. This estimator was compared with other previously tested estimators in simulation experiments that demonstrated the properties of the chosen estimator and the effect of erroneous data on the precision of the OD estimates. The simulation results indicated that the fixed-point estimator produced the most accurate OD estimates of those tested and that data with measurement error (e.g. from malfunctioning loop detectors) severely affect the precision of OD estimates.