Empower Low-Altitude Economy: A Reliability-Aware Dynamic Weighting Allocation for Multi-modal UAV Beam Prediction

TL;DR

This paper addresses the challenge of reliable, fast beam prediction for UAV mmWave links in the Low-Altitude Economy. It introduces SaM²B, a reliability-aware dynamic weighting framework that fuses multi-modal cues (visual ROI, GPS, altitude-distance, and UAV posture) through cross-modal contrastive learning and transformer-based attention to produce semantically aligned beam representations. The approach improves robustness to modality degradation and distribution shifts, achieving near 90% Top-1 beam accuracy on real-world UAV data and outperforming fixed-weight baselines. The results demonstrate the practical potential of adaptive, semantic, multi-modal fusion for reliable UAV connectivity in dynamic LAE scenarios, with future work on online uncertainty and active beam tracking.

Abstract

The low-altitude economy (LAE) is rapidly expanding driven by urban air mobility, logistics drones, and aerial sensing, while fast and accurate beam prediction in uncrewed aerial vehicles (UAVs) communications is crucial for achieving reliable connectivity. Current research is shifting from single-signal to multi-modal collaborative approaches. However, existing multi-modal methods mostly employ fixed or empirical weights, assuming equal reliability across modalities at any given moment. Indeed, the importance of different modalities fluctuates dramatically with UAV motion scenarios, and static weighting amplifies the negative impact of degraded modalities. Furthermore, modal mismatch and weak alignment further undermine cross-scenario generalization. To this end, we propose a reliability-aware dynamic weighting scheme applied to a semantic-aware multi-modal beam prediction framework, named SaM2B. Specifically, SaM2B leverages lightweight cues such as environmental visual, flight posture, and geospatial data to adaptively allocate contributions across modalities at different time points through reliability-aware dynamic weight updates. Moreover, by utilizing cross-modal contrastive learning, we align the "multi-source representation beam semantics" associated with specific beam information to a shared semantic space, thereby enhancing discriminative power and robustness under modal noise and distribution shifts. Experiments on real-world low-altitude UAV datasets show that SaM2B achieves more satisfactory results than baseline methods.

Figures (4)

• Figure 1: UAV network in low-altitude economy and the proposed multi-modal beam prediction framework.
• Figure 2: $\text{SaM²B}$ performs semantic-aware multi-modal beam prediction by dynamically reweighting modalities under reliability cues.
• Figure 3: Performance comparison under different modalities of the proposed $\text{SaM²B}$.
• Figure 4: Task performance comparison about $\text{SaM²B}$ and other baselines.