Use of Terrain and Obstacle Databases to Support UAS Detect-and-Avoid in Complex Urban Environments

Abstract

Recent developments in detect-and-avoid (DAA) technologies have focused on onboard sensing and cooperative surveillance, yet their performance is tightly coupled to the underlying representation of terrain and obstacles used to constrain trajectory planning and interpret sensor outputs. This paper investigates the use of high-resolution terrain and obstacle databases as integral components of DAA architectures for small and medium UAS navigating complex urban environments. The study considers static man-made structures, vegetation, moving ground traffic proxies, and topological information such as road corridors and building facades, and examines how their representation influences conflict prediction, sensor fusion, and maneuver generation. A modeling framework is developed in which terrain and obstacle databases support probabilistic occupancy prediction, occlusion-aware surveillance, risk-based separation buffers, and constraint-aware trajectory optimization. The approach is analyzed with respect to map resolution, latency, integrity risk, and robustness to registration errors between geolocation, navigation, and database coordinates. Simulation experiments in canonical downtown scenarios are used to evaluate detectability of cooperative and non-cooperative intruders, feasibility of safe maneuvers within regulatory altitude layers, and sensitivity to degraded database quality. The results highlight the conditions under which terrain and obstacle information significantly influences achievable safety margins and operational efficiency for urban UAS DAA, and identify modeling choices that are compatible with scalable implementation.