Multimodal Mixture-of-Experts for ISAC in Low-Altitude Wireless Networks

2025-12-01

AI Summary

Existing static multimodal fusion methods underperform in Low-Altitude Wireless Networks (LAWNs) due to dynamic channel heterogeneity and time-varying modality reliability. To address this, we propose a dynamic adaptive multimodal fusion framework tailored for Integrated Sensing and Communication (ISAC). Built upon a Mixture-of-Experts (MoE) architecture, the framework employs dedicated expert networks for visual, radar, LiDAR, and positioning modalities, coupled with a lightweight, information-driven gating mechanism that dynamically assesses modality reliability and informativeness—enabling sparse activation and real-time weight allocation. Evaluated across three representative ISAC tasks, our method significantly improves sensing accuracy and communication reliability while reducing onboard computational load and energy consumption. Experimental results demonstrate its effectiveness, efficiency, and scalability in dynamic low-altitude environments.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Integrated sensing and communication (ISAC) is a key enabler for low-altitude wireless networks (LAWNs), providing simultaneous environmental perception and data transmission in complex aerial scenarios. By combining heterogeneous sensing modalities such as visual, radar, lidar, and positional information, multimodal ISAC can improve both situational awareness and robustness of LAWNs. However, most existing multimodal fusion approaches use static fusion strategies that treat all modalities equally and cannot adapt to channel heterogeneity or time-varying modality reliability in dynamic low-altitude environments. To address this fundamental limitation, we propose a mixture-of-experts (MoE) framework for multimodal ISAC in LAWNs. Each modality is processed by a dedicated expert network, and a lightweight gating module adaptively assigns fusion weights according to the instantaneous informativeness and reliability of each modality. To improve scalability under the stringent energy constraints of aerial platforms, we further develop a sparse MoE variant that selectively activates only a subset of experts, thereby reducing computation overhead while preserving the benefits of adaptive fusion. Comprehensive simulations on three typical ISAC tasks in LAWNs demonstrate that the proposed frameworks consistently outperform conventional multimodal fusion baselines in terms of learning performance and training sample efficiency.

Problem

• Adaptive multimodal fusion for dynamic low-altitude ISAC
• Overcoming static fusion limitations in heterogeneous aerial channels
• Enabling scalable, efficient ISAC under aerial platform constraints

Innovation

• Mixture-of-experts framework for adaptive multimodal fusion
• Lightweight gating module assigns weights based on modality reliability
• Sparse MoE variant reduces computation by selectively activating experts