Spectral Estimation to Quantify Road/Track Surface Degradation

Track testing methods are utilized in the automotive industry for emissions and fuel economy certification. These track tests are performed on smooth road surfaces which deteriorate over time due to wear and weather effects, hence warranting regular track repaves. The study focuses on the impact of repaving on track quality and surface degradation due to weather effects. 1D surface profiles and 2D surface images at different spatial frequencies were measured at different times over a span of two years using various devices to study the repave and degradation effects. Data from coastdown tests was also collected over a span of two years and is used to demonstrate the impact of track degradation and repaving on road load characterization parameters that are used for vehicle certification tests. Kernel density estimation and non-parametric spectral estimation methods are used to visualize the characteristic features of the track at different times. In the pre-processing stage, outliers are removed from the surface profile scans by using a method based on kernel density estimation. Gaussianity and non-stationarity in data is also verified while pre-processing by using a strided convolution-based mean filter. Signal power obtained from power spectral density estimate is hypothesized as a metric to compare track roughness for various measurements in order to develop a useful indicator to justify a track repave.