Modeling Edge-to-Cloud Offloading Workloads for Autonomous Vehicles

2026-03-24

AI Summary

Existing general-purpose traffic models struggle to accurately capture the workload characteristics of data offloading from autonomous vehicles to the cloud. To address this gap, this work proposes the first fine-grained, data-driven modeling framework tailored specifically for edge–cloud offloading in autonomous driving scenarios. The framework categorizes offloading traffic into three distinct types: telemetry, event-driven fleet learning, and high-definition map updates. Parameterized using real-world urban vehicle trajectories from Munich and empirical measurements, the model enables realistic simulation and analysis. The study reveals that this workload exhibits distinct spatiotemporal dynamics—scaling linearly with vehicle penetration rates, demonstrating strong temporal regularity, and showing highly uneven spatial access patterns—features that markedly diverge from those assumed by conventional traffic models.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Application Domains: Mobility, Driving & Flight
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Cloud, edge and content delivery systems for the Web
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Autonomous vehicles generate large volumes of data for applications such as fleet monitoring, model retraining, and high-definition map updates. Existing studies often rely on generic traffic traces, which do not capture the characteristics of autonomous driving workloads. This paper proposes a system-level workload modeling framework for vehicle-to-cloud data. We classify offloaded data into three types: telemetry, event-driven fleet learning, and high-definition map updates, while we model their generation using a parameterized formulation based on empirical data. Using a real-world mobility trace from Munich, we analyze the resulting workloads over time and space. The results show that workload scales with vehicle penetration, exhibits temporal structure and spatial imbalance across access points, and is distinguished from baseline traffic models.

Problem

• autonomous vehicles
• edge-to-cloud offloading
• workload modeling
• telemetry data
• spatiotemporal traffic characteristics

Innovation

• workload modeling
• autonomous vehicles
• edge-to-cloud offloading
• parameterized formulation
• spatiotemporal analysis