Multimodal Environment-Aware 3D Adaptive Scheduling for UAV-Enabled Fluid Antenna Systems

To mitigate 3D spatial blockages and channel uncertainty in VHF/low-UHF UAV emergency networks, this paper presents a multimodal environment-aware framework for 3D virtual fluid antenna port scheduling within an Integrated Sensing, Computing, and Communication (ISCCC) architecture. Under rigorously verified spatial resolution and channel stationarity conditions, UAV micro-mobility is mapped onto a discrete 3D virtual port array, transforming continuous flight space into a controllable fluid antenna system (FAS). We define a spatial efficiency metric that quantifies the Pareto trade-off between spatial degrees of freedom and estimation error, parameterized by an error-sensitivity index, and prove the existence of a unique optimal flight scale. Utilizing a joint spatio-temporal channel model, we derive the irreducible entropy lower bound of channel uncertainty, demonstrating that intrinsic environmental randomness constitutes a fundamental predictability limit regardless of port density—a benchmark independent of any specific scheduling strategy. To ensure real-time viability, we introduce an ISCCC-inspired computation-and-caching strategy that leverages pre-calculated stationary probabilities to drive a multidimensional scoring mechanism incorporating channel entropy-based stability, predictive SNR, and load balancing. The suboptimality gap relative to a perfect-CSI oracle is analytically bounded, and proven to narrow significantly under the high temporal correlation inherent in VHF bands. Numerical results confirm that the proposed strategy attains 10.36 bps/Hz effective throughput and 10.5% outage probability, consistently outperforming rule-based, learning-based, and 2D spatial baselines, particularly under prolonged structural obstructions.