A Mobile Cross Slope Measurement Method Using Lidar

42 A properly measured, effective cross slope facilitates drainage on highways and prevents 43 hydroplaning. There is a need for transportation agencies to identify and measure road sections 44 that have improper and non-effective cross slopes so that timely corrective maintenance can be 45 performed. However, the traditional manual methods used by transportation agencies to measure 46 cross slope, i.e. using a digital level, are time-consuming and labor-intensive. They are not 47 feasible for conducting a network-level cross slope measurement. This research project, 48 sponsored by the United States Department of Transportation Research Innovative Technology 49 Administration (USDOT RITA) program, proposes a new mobile cross-slope measurement 50 method using emerging mobile LiDAR technology that can accurately and effectively conduct 51 network-level cross slope measurement at highway speed. The contributions of this paper 52 include the following: 1) proposing a mobile cross slope measurement method using emerging 53 LiDAR technology (LiDAR cloud calibration, data acquisition, ROI extraction, and cross slope 54 computation) ; 2) through a sensitivity study, determining the key parameter (i.e. the ROI 55 interval) for the proposed method; 3) critically validating the accuracy and the repeatability of 56 the proposed method by testing it in a controlled environment; and 4) conducting a case study to 57 demonstrate the capability of the proposed method. Experimental testing at the Georgia Tech 58 Savannah campus is first conducted to critically assess the accuracy and repeatability of 59 measuring cross slopes. When compared to the ground truth established by digital level, the 60 results from 15 cross slopes (ranging between 1.9% and 7.2%) show that the proposed method 61 can achieve desirable accuracy with an average measurement difference of 0.13% (i.e. 0.08°). 62 Results show that the proposed method can achieve a desirable level of repeatability with a 63 standard deviation of less than 0.05% (i.e. 0.03°) in three different test runs. A case study on 64 Interstate 285 is then conducted to demonstrate the capability of the proposed method for 65 achieving proper, effective, network-level cross slope measurements. Results show that the 66 proposed mobile method can be operated at highway speed and is very promising for a network67 level cross slope adequacy assessment. Finally, conclusions and recommendations are presented. 68 Tsai, Ai, Wang and Pitts 3 INTRODUCTION 69 The cross slope is a crucial roadway feature that accelerates water drainage and reduces potential 70 roadway hazards that cause hydroplaning. In addition, roadways are designed to slope in the 71 transverse direction to control lateral vehicle wandering, especially on curves, known as super72 elevation. Transportation agencies need to identify the potential for vehicle roadway departure or 73 hydroplaning by analyzing the roadway's cross slope features, along with other factors, e.g. 74 pavement surface friction, vertical grade, etc. 75 Most transportation agencies currently conduct cross slope measurement manually by 76 using a digital level (1, 2). This method requires field engineers to physically place a 4-foot 77 electronic level on the pavement surface to obtain the measurement. FIGURE 1 shows a field 78 engineer conducting manual measurement. Such manual measurement is time-consuming and 79 dangerous. More importantly, since it is not practical to conduct the manual measurement 80 continuously without traffic control, it is almost impossible for transportation agencies to collect 81 the cross slope data necessary to support a network-level analysis for identifying safety concerns, 82 such as potential hydroplaning locations, inadequate super-elevated locations, etc. 83 84 FIGURE 1 Manual cross slope measurement (3). 85 Light detection and ranging (LiDAR) technology has become popular among 86 transportation agencies for asset data collections (4, 5), traffic data collection (6), safety 87 assessment (7), etc. In recent years, mobile LiDAR has been increasingly used for roadside 88 inventory (8), roadway geometry measurement (9-11), etc. With accurate geo-referenced 3D 89 point cloud data and a high-scanning frequency, the mobile LiDAR technology has the potential 90 to achieve safer and more effective cross slope measurements at highway speed. Such 91 measurements can be extracted from a LiDAR point cloud. 92 Although some researchers have introduced other sensing technologies to facilitate a 93 mobile cross slope measurement method, such as airborne LiDAR (12), laser profiler and inertial 94 measurement unit (IMU) (13), etc., mobile LiDAR technology has never been practically used 95 for mobile cross slope measurement. This study proposes a new mobile cross slope measurement 96 method using mobile LiDAR; it critically evaluates the accuracy and repeatability of the 97 proposed method by using actual data collected on the Georgia Tech Savannah campus in 98 Savannah, Georgia. A case study is presented to demonstrate the network-level analysis 99 capability of the proposed method by using the actual data collected on Interstate 285 (I-285) in 100 Atlanta, Georgia. 101 This paper is organized as follows. The first section identifies the research need and 102 objective. The second presents the proposed mobile cross slope measurement method using 103 Tsai, Ai, Wang and Pitts 4 mobile LiDAR. The third section presents a critical assessment conducted to validate the 104 accuracy and repeatability of the proposed method. The fourth section uses a case study to 105 demonstrate the capability of the proposed method. 106 PROPOSED CROSS SLOPE MEASUREMENT METHOD USING MOBILE LIDAR 107 Proposed Method 108 In this section, a mobile cross slope measurement method is proposed. The method uses Georgia 109 Tech’s mobile sensing system, which can collect the data at highway speed. The integrated 110 system includes an emerging mobile LiDAR system (i.e. Riegl LMS-Q120i), high resolution 111 video cameras (i.e. Point Grey Gras-50S5C), and an accurate positioning system (i.e. Applanix 112 LV 210PP) composed of a global positioning system (GPS), an IMU, and a distance 113 measurement instrument (DMI). FIGURE 2(a) shows the mobile LiDAR system and the two 114 corresponding cameras; the rest of the system components are enclosed in the data collection 115 vehicle. 116 The mobile LiDAR system in this study is a line-scanning laser device that produces 117 10,000 laser points per second. As the vehicle moves in the longitudinal direction on the road, 118 the scanning line of the LiDAR system is aligned parallel to the transverse direction. The 119 scanning range is 80° in the horizontal direction, which produces an 80° fan covering the 120 pavement surface. Currently, the frequency of the LiDAR system is configured at 100 Hz with 121 100 points in each scan. FIGURE 2(b) illustrates data acquisition covering the whole driving 122 lane. 123