Frog and Blade Based Branching Direction Detection in LiDAR Data

Train-borne localization plays an important role in future railway applications which require quick track-selective positioning. In this paper, a branching direction detection approach based on frog and blade is presented to address the ambiguity of which track is chosen when a turnout is passed. The reference positions (i.e. the tip of the blade and the tip of the frog) on turnouts are detected as well. Since track detection is the prerequisite, rail and track sections are extracted using an onboard LiDAR sensor considering features related to the rail profile and the longitudinal geometry. The relative position between the ego track and branching tracks are considered as track events, from which the branching directions are derived. The approach is evaluated with datasets collected in a railway test ground. The results demonstrate that the branching directions are detected accurately with a recall of 97.6% and a precision of 98.8%. Moreover, the proposed approach makes determination of the branching direction 22.4m and 16.4m earlier than the frog- or blade- based one does in facing and trailing directions respectively.