A computer program based on statistics and signal process theory was developed to automatically detect peaks and valleys from sensor response signals obtained during live heavy truck and falling-weight deflectometer testing. Statistics are applied to each signal to characterize the nature of the response signal and to make the detection of maxima and minima more efficient. Noise effects are treated by applying filtering techniques including Fast Fourier Transform and time domain filtering.
The Procedure was found to work effectively and is now being used to process pavement response data that has been collected at the Minnesota Road Research Project (Mn/ROAD) over the past three years. The output file from the program is readily loaded into the Mn/ROAD database.
