Skid resistance has been modelled as an exponential function of speed involving two parameters, zero-speed intercept and percent normalised gradient. Several inconsistencies associated with the percent normalised gradient are presented, and an alternative parameter called a "speed constant", which is free of those inconsistencies, is proposed. Two methods for identifying the model parameters are evaluated using test data collected on road sections in Pennsylvania. Both methods produce relatively accurate predictions of skid number at different speeds even though the estimates of the model parameters calculated in each method are different. The differences between the parameter estimates obtained from the two methods clearly demonstrate that great care must be exercised in relating the parameter estimates to pavement micro- and macrotexture characteristics. The normalised texture index, a new measure of combined micro- and macrotexture, is introduced. Finally, the effect of speed on the model parameters is investigated first analytically and then using experimental data.
Abstract