Simulation-driven control of UAV smoke-munition release for dynamic line-of-sight obscuration

Abstract

This paper studies smoke-munition release control for unmanned aerial vehicles in a reproducible three-dimensional simulation environment. The simulator integrates constant-altitude UAV flight, free-fall munition dynamics, post-detonation smoke-cloud settling, incoming-object motion, and a strict line-of-sight obscuration evaluator. Instead of relying on a target-center test, the protected cylindrical target is discretized into edge and central-axis sampling points, and the shortest distance from the smoke-cloud center to each incoming-object–target-sample line segment is calculated. The maximum of these distances is adopted as the global obscuration metric so that effective coverage is recognized only when all sampled lines of sight are covered by the smoke envelope. The supplied simulation yields an effective interval from 2.96 s to 7.48 s, corresponding to a total effective obscuration duration of 4.54 s. The resulting framework establishes a transparent deterministic baseline and an RL-ready environment interface for future policy learning. The study is explicitly positioned as a simulation platform and deterministic two-stage optimization baseline rather than as a completed learned controller; its applicable scope, target-sampling sensitivity, search-method comparison, and parameter settings are specified to make the algorithmic contribution and reliability boundary clear. The reported 4.54 s effective duration is therefore interpreted only as a scenario-specific geometric result under the fixed simulator inputs, not as a field-validated operational guarantee.