Distributed Split Computing Using Diffusive Metrics for UAV Swarms

2025-03-20

AI Summary

To address communication bottlenecks, high latency, and poor energy efficiency in dynamic machine learning inference for large-scale heterogeneous UAV swarms, this paper proposes a fully distributed diffusion-based collaborative inference framework. Unlike centralized or topology-dependent approaches, the framework operates without global network knowledge and introduces Aggregated GFLOPs—a novel diffusion-aware metric—for adaptive, resource-driven task partitioning. It integrates local neighbor-aware information diffusion, distributed consensus, edge early-exit, and heterogeneity-aware offloading to enable real-time, dynamic inference path reconfiguration. Experimental evaluation demonstrates that, compared to baseline methods, the framework reduces end-to-end latency by 42%, cuts energy consumption by 37%, and improves task throughput by 2.1×, while significantly enhancing system fairness and topological robustness under dynamic swarm conditions.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Distributed Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

In large-scale UAV swarms, dynamically executing machine learning tasks can pose significant challenges due to network volatility and the heterogeneous resource constraints of each UAV. Traditional approaches often rely on centralized orchestration to partition tasks among nodes. However, these methods struggle with communication bottlenecks, latency, and reliability when the swarm grows or the topology shifts rapidly. To overcome these limitations, we propose a fully distributed, diffusive metric-based approach for split computing in UAV swarms. Our solution introduces a new iterative measure, termed the aggregated gigaflops, capturing each node's own computing capacity along with that of its neighbors without requiring global network knowledge. By forwarding partial inferences intelligently to underutilized nodes, we achieve improved task throughput, lower latency, and enhanced energy efficiency. Further, to handle sudden workload surges and rapidly changing node conditions, we incorporate an early-exit mechanism that can adapt the inference pathway on-the-fly. Extensive simulations demonstrate that our approach significantly outperforms baseline strategies across multiple performance indices, including latency, fairness, and energy consumption. These results highlight the feasibility of large-scale distributed intelligence in UAV swarms and provide a blueprint for deploying robust, scalable ML services in diverse aerial networks.

Problem

• Distributed split computing for UAV swarms
• Overcoming network volatility and resource constraints
• Enhancing task throughput and energy efficiency

Innovation

• Distributed split computing for UAV swarms
• Aggregated gigaflops metric for resource allocation
• Early-exit mechanism for dynamic workload adaptation