Low-Altitude Channel Multipath Prediction via Panoramic Perception and Vision-Language Model

2026-07-23

AI Summary

Traditional statistical and ray-tracing channel models struggle to balance accuracy and computational efficiency in low-altitude scenarios, particularly due to inadequate modeling of small-scale multipath parameters. To address this limitation, this work proposes PanoLAMP, a novel framework that, for the first time, integrates pretrained vision-language models with panoramic RGB-D perception to directly regress key multipath parameters—including delay, power, azimuth, and elevation offsets of line-of-sight paths—from observed transmitter-receiver data. By circumventing the need for high-fidelity 3D environmental reconstruction and computationally intensive simulations, PanoLAMP achieves significantly superior performance over existing baselines on a synthetic dataset comprising 18,949 UAV-to-vehicle links, demonstrating not only enhanced prediction accuracy for multipath characteristics but also strong generalization across varying flight altitudes.

Abstract

Unmanned aerial vehicle (UAV) communication is expected to support a wide range of low-altitude applications in 6G mobile networks. However, traditional statistical channel models provide limited accuracy in specific environments, while deterministic methods such as ray tracing usually rely on accurate three-dimensional environment models and involve high computational complexity. Existing multimodal channel prediction approaches mainly focus on large-scale metrics such as path loss, and remain insufficient for modeling small-scale parameters. To address these limitations, this paper proposes PanoLAMP, a Panoramic perception and vision-language model-based Low-Altitude Multipath Prediction framework. It adopts a pretrained vision-language model as the backbone and captures the propagation environment features through panoramic RGB-D observations collected at both the transmitter and receiver to predict the delay, power, azimuth angle, and zenith angle offset relative to the line-of-sight path. Experiments are conducted on a synthetic dataset containing 18,949 UAV-vehicle links across seven UAV altitudes. Experimental results show that the proposed method consistently outperforms representative baselines in both multipath parameters and statistical metrics, and demonstrates stronger generalization across different flight heights.

Problem

• Low-altitude channel
• Multipath prediction
• UAV communication
• Small-scale parameters
• Channel modeling

Innovation

• vision-language model
• panoramic perception
• low-altitude channel prediction
• multipath modeling
• UAV communication