ABSTRACT:
Road roughness is a key parameter for controlling pavement construction processes and for assessing ride quality of both paved and unpaved roads. This paper describes algorithms used in processing three-dimensional (3D) stationary terrestrial laser scanning (STLS) point clouds to obtain surface maps of point wise indices that characterize pavement roughness.
The backbone of the analysis is a quarter-car model simulation over a spatial 3D mesh grid representing the pavement surface. Two case studies are presented, and results show high spatial variability in the roughness indices both longitudinally and transversely (i.e., different wheel path positions).
It is proposed that road roughness characerization using a spatial framework provides more details on the severity and location of roughness features compared to the one-dimensional methods. This paper describes approaches that provide an algorithmic framework for others collecting similar STLS 3D spatial data to be used in advanced road roughness characterization.
DATA COLLECTION USING STATIONARY LASER SCANNER
LIDAR systems measure information (spatial coordinates and color) of a 3D space and store the information in a 3D point cloud. The term LIDAR is generic, and includes airborne laser scanning technologies, mobile scanners mounted on vehicles, and stationary laser scanners or stationary terrestrial laser scanners (STLS), where the laser scanner is fixed at a station with known coordinates and based on the distance between the scanner and the detected points, a geospatially referenced 3D point cloud is constructed.
Trimble CX 3D STLS system was used in two case studies to acquire 3D laser scans. The position accuracy of a single point is 4.5 mm at 30 m and drops to 7.3 mm at 50 m. The distance accuracy is 1.2 mm at 30 m and drops to 2 mm at 50 m. Figure 1 shows the scanner set-up.
Trimble CX 3D Laser Scanner Set-Up.
DATA PROCESSING
Visioning Algorithms:
To use the road surface in simulations developed algorithms require a uniformly spaced grid the longitudinal direction. To achieve that a mesh grid is formed with grid elements that have a predefined x and y edge dimensions. The grid centre elevation is calculated as the average of all cloud points falling in that grid region.
All points are rotated and translated to a local coordinate system corresponding to the longitudinal and transvers axes (Zalama et al. 2011).Transformation for processing along horizontal curves can be achieved by constructing a curvilinear local coordinate system; however, for simple geometries without curves the point cloud is rotated globally, where the x axis corresponds to the longitudinal direction and the y axis to the transverse direction.
Point Cloud Data after Processing in the Visioning Algorithm for a Pavement Section.
Roughness Evaluation Algorithms:
Evaluation of roughness described herein is based on the responses of the quarter-car model described in (ASTM E1926-08). Figure 5 presents a schematic of the quarter-car model. Where Ms and Mu are the sprung and unsprung masses respectively, ks and kt are the suspension and tire spring coefficients respectively, and cs is the suspension damping rate.
CASE STUDIES
Road roughness of two road sections was evaluated, the sections include a 55 m long rural unpaved road and a 58.8 m long HMA overlay over a jointed plain concrete pavement (JPCP) pavement. Figures 6a and 6b show the rectified slope map for rural road and the paved road respectively.
Absolute Rectified Slope Surface Maps for (a) Unpaved and (b) Paved roads.
SUMMARY AND CONCLUSIONS
This paper introduces the frame work of quantitative techniques for evaluating the surface roughness of paved and unpaved roads. Methods for evaluating the surface roughness can be set to two main categories, finite difference simulation and frequency domain analysis. The key findings from this research are:
- Terrestrial laser scanning is a promising technology to assess a range of surface
conditions for unpaved roads. - 2-D Surface roughness maps were developed using the information obtained from the laser scanner.
- Algorithms used in producing 2-D roughness maps are semi-automated, and further
developments are expected to introduce fully automated algorithms that can process the data directly after scanning. - IRI values are highly variable across the road section, thus it is hard to define the
appropriate profile to be used as the representative profile. - The proposed analysis technique can be used to identify localized rough features.
- Finite difference simulations are not suitable for short profiles.
- Filtering of the profiles in Fourier space as a tool to get the spatial roughness maps are more suitable compared to finite difference algorithms for analyzing short profiles, however wavelet analysis provide a more robust approach to analyze short profiles and identify localized features.
- At this stage, high resolution terrestrial laser scans are time consuming and require
trained personnel; however, newer terrestrial laser scanners will be able to reduce the data acquisition time significantly due to faster scanning rates and longer ranges.
Source: Iowa State University
Authors: Ahmad A. Alhasan | David J. White