| ID | Forename | Surname | Affiliation |
|---|---|---|---|
| 1 | Vladimir | Pajic | Faculty of Technical Sciences, University of Novi Sad |
vigorous-x-ray
Serbia
Description and Rationale
Who I am. A one-person team. I hold a PhD in geoinformatics from the Faculty of Technical Sciences, University of Novi Sad, where I teach photogrammetry, laser scanning and geospatial data processing. My research has focused on managing and processing very large geospatial datasets — point-cloud data management on Apache Spark, LiDAR classification with machine learning, and spatial indexing for distributed workloads. In industry I work as a geospatial software engineer on production spatial platforms (PostGIS, raster and vector pipelines, cloud deployment). Data science, spatial processing and aviation are the three things I enjoy most, and this challenge happens to sit exactly where they meet.
Why I registered. Taxi-out time is usually modelled as a temporal problem, with the stand–runway pair treated as a categorical key. I think a good part of it is geometric: the actual taxi route from stand to runway threshold, its length, the number of turns and intersections, and the runway crossings along the way. My plan is to build a taxiway graph for all ten airports from open OpenStreetMap aeroway data, derive route-based priors from it, and measure how much they add to a solid gradient-boosting baseline with congestion and runway-configuration features. The question I want to answer, and publish either way, is whether a small, open, geodetically grounded spatial layer is worth as much as the usual temporal feature engineering.
Approach. Open data only, a reproducible pipeline, a public GitHub repository, and a short write-up for the Journal of Open Aviation Science.
Details
- Type: Academia
- Country: Serbia
- Number of team members: 1