Noise-Aware UAV Routing for Urban Logistics: A Data-Driven Decision-Support Framework

Urban UAV route planning requires an explicit connection between spatial inputs, route restrictions, and evaluation criteria. This study presents a single-case demonstration of risk-map-based adaptive ant colony optimization (RMA-ACO) within an OD-based spatial routing framework. Building geometry and origin–destination activity supply receiver-related inputs under a common trip-generation-rate assumption. The method combines spatial exclusions, a local distance–noise heuristic, backtracking, early stopping, and geometric postprocessing. One origin–destination pair in a residential area of Changzhou, China, is examined at a fixed altitude of 10 m. History-dependent local scores guide route selection, whereas final routes are evaluated by geometric length and an unscaled noise-related model index on a common field. Across 30 fixed-seed runs at 5 m, RMA-ACO versus ACO means are 2763.6 versus 3190.1 m for route length, 143.2 versus 202.9 s for runtime, and 59.2 versus 60.9 for the noise index (paired difference: −1.7; 95% CI: −2.26 to −1.14). These results compare the complete tested configurations under the prescribed proxy settings. They do not isolate component effects or establish measured exposure reductions, transferability to other delivery connections, or operational flight feasibility.