Abstract
The Indiana Department of Transportation (INDOT) traditionally conducts annual network inventory friction testing using locked-wheel trailers to mitigate wet pavement crashes. However, this method presents safety risks on high-speed facilities and raises concerns regarding data credibility. While macrotexture can be measured at highway speeds, microtexture—a critical component of friction—has historically required low-speed surrogate devices. This pilot study investigates the use of high-frequency laser sensors to capture both macro- and microtexture. Findings indicate that microtexture metrics (MPD, RMS, and SV) stabilize at a baseline length of 12.75 mm. Furthermore, Slope Variance (SV) proved as effective as Root Mean Square (RMS) in estimating friction. Correlation analysis revealed that wet pavement friction is highly sensitive to both macrotexture and microtexture, with microtexture SV playing a dominant role. This research suggests that high-speed laser scanning can effectively estimate pavement friction, enhancing both operational safety and data quality.
Introduction: The Need for Non-Contact Testing
Maintaining adequate pavement friction is a cornerstone of highway safety, particularly for preventing wet-weather accidents. For years, the Indiana Department of Transportation (INDOT) has relied on the locked-wheel trailer method for friction inventory. While reliable, this approach has significant drawbacks: it requires testing at 40 mph on interstates where traffic flows at 70 mph, creating hazardous speed differentials, and the physical contact accelerates tire and pavement wear.
Engineers have long sought to replace physical friction testing with non-contact laser measurements. However, previous attempts focused primarily on macrotexture (wavelengths 0.5mm to 50mm), which facilitates water drainage but is a poor sole predictor of friction. The missing variable is microtexture (wavelengths < 0.5mm), the fine-scale grit that provides tire adhesion. This study explores a framework for capturing both scales at highway speeds using advanced laser sensors.
Methodology and Sampling Requirements
To bridge the gap between lab-scale accuracy and field-scale speed, the research team evaluated two laser systems: a stationary 1 kHz high-resolution scanner for baseline "ground truth" and a 62.5 kHz vehicle-mounted sensor for high-speed data acquisition.
A critical component of the methodology involved the Nyquist sampling theorem. To accurately capture the peaks and troughs of pavement texture at speeds of 30–50 mph, the team determined that a laser must pulse at a frequency between 54 kHz and 90 kHz. This ensures that the microtexture profile is not "aliased" or lost in the noise of high-speed travel.
Beyond the standard Mean Profile Depth (MPD), the study utilized two additional metrics to characterize the surface:
- Root Mean Square (RMS): A measure of the overall magnitude of surface asperities.
- Slope Variance (SV): A measure of the "sharpness" or angularity of the pavement grit.
Experimental Setup and Datasets
Testing was conducted at the INDOT Friction Test Track, utilizing three distinct surfaces to represent the spectrum of road conditions:
- Slick Concrete: Finished to a high smoothness to simulate worn, low-friction areas.
- 9.5-mm Hot Mix Asphalt (HMA): A standard SuperPave mix representing typical highway surfaces.
- Tined Concrete: Featuring 3mm deep transverse grooves to simulate high-drainage, high-friction environments.
Field validation also included in-service asphalt testing using both smooth and ribbed tires to correlate laser data with traditional Friction Numbers (FN).
Key Insights: The Role of "Sharpness"
The study's most significant finding was the importance of Slope Variance (SV). While depth (MPD) is the traditional metric for texture, the "sharpness" of the microtexture (SV) was found to be a more critical driver of wet pavement friction. Essentially, the angularity of the aggregate particles matters as much as the depth of the gaps between them. Notably, the study found that SV and RMS are redundant when used together; either metric is sufficient on its own to estimate friction from microtexture.
Additionally, the researchers identified a "stability threshold" for microtexture. Metrics like MPD, RMS, and SV tended to fluctuate at very short measurement lengths but remained constant once the baseline exceeded 12.75 mm. This finding provides a concrete standard for future sensor calibration and data processing algorithms.
Results and Correlation Analysis
The correlation between laser-derived metrics and physical friction measurements was remarkably strong. Wet pavement friction showed a Pearson correlation of 0.9722 with macrotexture MPD and 0.9159 with microtexture SV.
Regression analysis further demonstrated that:
- Dry Friction: Almost entirely dependent on microtexture adhesion.
- Wet Friction: A complex interaction of macrotexture (for bulk water drainage) and microtexture (for breaking the remaining water film).
- Predictive Power: The study successfully developed linear regression models capable of predicting Friction Numbers (FN) using only laser data, achieving a high degree of accuracy within the test environment.
Conclusion and Future Directions
This pilot study confirms that high-speed laser sensors can capture the microtexture data necessary for accurate friction estimation. By moving toward a non-contact, high-speed testing regime, DOTs can significantly improve the safety of field operations while obtaining more comprehensive data on pavement health.
However, challenges remain. The 1 kHz stationary laser proved highly sensitive to ambient light, requiring physical shielding in the field. Furthermore, while 62.5 kHz is sufficient for macrotexture, capturing the finest microtextures on extremely slick surfaces may require even higher frequencies or the integration of high-resolution image processing. Future research will focus on expanding the dataset to more diverse real-world pavements and exploring 21MP+ camera systems as a potential alternative to laser scanning.
Summary updated