• December 12, 2018

    PacTrans Technology Transfer Success Story 2018 #6: Vehicle-Pedestrian Near-Miss Detection Using Onboard Monocular Video Data

    Several years ago, faculty and students from the University of Washington’s Smart Transportation Application & Research (STAR) Lab, partnered with the Washington State Transit Insurance Pool on a project cofounded by PacTrans and a TRB IDEA Grant to student collision avoidance technologies for transit vehicles. New algorithms were produced to identify near miss situations using the footage collected by the video cameras. Near miss data allows collision avoidance systems to be modified to capture situations that they may have otherwise missed, but also produces a rich data set for public agencies and academic researchers to investigate situations that arise with high frequency.

    To better advertise this new technology, STAR Lab researchers leveraged PacTrans Technology Transfer Success Story funds to develop a promotional video, which highlights its power and potential. Further, the original algorithms were developed for, and run on, desktop computers. The team is currently in the process of transferring that system to a real-time onboard near-miss detection system, taking real-time video inputs and generate warnings to drivers when any conflict is detected. This technology transfer requires some modification of the algorithms since onboard computers are generally less powerful in terms of computational power. For more information please email Yinhai Wang (yinhai@uw.edu).