Comparative study of freeway incident detection algorithms.

Four incident detection algorithms have been selected for this study: California, University of California, Berkeley, and Tokyo Metropolitan Expressways (MEX) freeway incident detection algorithms and artificial neural network model. These algorithms are calibrated and evaluated using 34 real-life incident data collected on the Tokyo Metropolitan Expressway. A performance index equation was introduced to assist in selecting optimal parameter values. Results of the fully calibrated algorithms are presented. (A)