Browsing by Subject "road departure mitigation system"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Contrast Between Road and Roadside Material For Road Edge Detection In Vehicle Road Departure Mitigation System
    (National Highway Traffic Safety Administration, 2019) Yi, Qiang; Chien, Stanley; Chen, Yaobin; Sherony, Rini; Electrical and Computer Engineering, School of Engineering and Technology
    Vehicle roadway departure crashes results in a large number of fatalities in the U.S. Road departure mitigation (RDM) systems rely on the road edge and road boundary identification. Cameras are widely used in RDMS for identifying road edges. The contrast between road and road boundary objects is one of the key image features used by the camera to detect road edges. This paper analyzes and compares the contrasts between various road surfaces. and road edges.
  • Thumbnail Image
    Item
    Determine characteristics requirement for the surrogate road edge objects for road departure mitigation testing
    (2019) Chien, Stanley; Yi, Qiang; Lin, Jun; Saha, Abir; Li, Lin; Chen, Yaobin; Chen, Chi-Chih; Sherony, Rini; Electrical and Computer Engineering, School of Engineering and Technology
    Road departure mitigation system (RDMS), a vehicle active safety feature, uses road edge objects to determine potential road departure. In the U.S., 45%, 16%, and 15% of car-mile (traffic flow * miles) roads have grass, metal guardrail, and concrete divider as road edge, respectively. It is difficult to test RDMS with real roadside objects. Lightweight and crashable surrogate roadside objects that have representative radar, LIDAR and camera characteristics of real objects have been developed for testing. This paper describes the identification of automotive radar, LIDAR, and visual characteristics of metal guardrail, concrete divider, and grass. These characteristics will be referenced for designing and fabricating the representative surrogate objects for RDMS testing. Colors and types of the roadside objects were identified from 24,735 randomly sampled locations in the US using Google street view images. The radar and LIDAR parameters were measured using 24GHz/77GHz radar and 350-2500nm IR spectrometer.