Intelligent Downtilt Configuration for Low-Altitude Air–Ground Cellular Networks via Sampling-Assisted Evolutionary Optimization

Low-altitude cellular services are increasingly demanded by unmanned aerial vehicle (UAV)-enabled logistics, inspection, and emergency missions, yet existing terrestrial-oriented configurations often lead to severe 3D coverage overlap, interference accumulation, and spatial serving-sector ambiguity in airspace. This paper develops a practical region-oriented framework to enhance air–ground cellular performance under existing deployments by jointly tuning sector electrical and mechanical downtilts. A unified evaluation pipeline is constructed to characterize received signal strength (RSS), signal-to-interference-plus-noise ratio (SINR), and spatial serving-sector patterns, and to quantify service reliability through threshold-based coverage rates and edge performance. By treating UAV service altitude as an adjustable planning parameter, the proposed framework captures height-dependent coverage behavior and provides altitude-specific downtilt optimization results for representative low-altitude service heights. A sampling-based regional evaluation scheme is further designed to approximate area-level performance with reduced computational cost, enabling repeated fitness evaluation within a genetic search procedure. Extensive evaluations show that the proposed approach consistently enhances low-altitude coverage reliability and edge experience while keeping the coupled impact on terrestrial performance within a controlled range.