The modified equations provide reliable results with average absolute errors between 7.86% and 8.56%, and inferior dispersion around the averages.
This showed average absolute errors around ±4% for the brake power, and ±5% for the brake efficiency at different frequencies.
For thicker layers (from 11 to 13 cm), the average absolute errors for both methods were similar (approximately 7%).
The CMP method presented lower average absolute errors for the thinner asphalt layers (about 6 cm) in comparison with the SRM method (6% vs 18%).
The average absolute errors (AAE) of training set (366 measurements) and test set (92 measurements) are 0.344 v % and 0.304 v %, respectively.
The average absolute errors of loss of load expectation (LOLE) and loss of energy expectation (LOEE) are 0.22% and 0.39%, respectively.
Figures 15 and 16 compare the average absolute errors between the correlation (Glaso 1985) and the GRNN model predictions for the training and testing data sets, respectively.
First column shows the average absolute error calculated for translation, and the average absolute errors for rotation and scaling are given in second and third columns, respectively.
