Bandwidth Allocation for Cloud-Augmented Autonomous Driving

2025-03-26

AI Summary

Cloud-enhanced autonomous driving systems suffer from response latency under limited cellular bandwidth, resulting in no accuracy improvement despite cloud offloading. Method: This paper formulates the bandwidth constraint as a joint resource allocation problem aimed at maximizing driving utility and proposes, for the first time, a task-aware dynamic bandwidth scheduling framework—departing from conventional static cloud-edge offloading paradigms. The approach integrates integer programming optimization, adaptive quantization and sparse coding compression, a multi-granularity model pool (lightweight, medium, and full-scale models), and real-time bandwidth prediction with feedback control. Results: Evaluated on the Waymo Open Dataset, the method achieves an average model accuracy gain of 15 percentage points over baseline methods, while maintaining end-to-end latency consistently below 100 ms—satisfying stringent real-time requirements for autonomous driving.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Autonomous vehicle (AV) control systems increasingly rely on ML models for tasks such as perception and planning. Current practice is to run these models on the car's local hardware due to real-time latency constraints and reliability concerns, which limits model size and thus accuracy. Prior work has observed that we could augment current systems by running larger models in the cloud, relying on faster cloud runtimes to offset the cellular network latency. However, prior work does not account for an important practical constraint: limited cellular bandwidth. We show that, for typical bandwidth levels, proposed techniques for cloud-augmented AV models take too long to transfer data, thus mostly falling back to the on-car models and resulting in no accuracy improvement. In this work, we show that realizing cloud-augmented AV models requires intelligent use of this scarce bandwidth, i.e. carefully allocating bandwidth across tasks and providing multiple data compression and model options. We formulate this as a resource allocation problem to maximize car utility, and present our system sysname which achieves an increase in average model accuracy by up to 15 percentage points on driving scenarios from the Waymo Open Dataset.

Problem

• Optimize bandwidth allocation for cloud-augmented AV models
• Balance data compression and model options under limited bandwidth
• Maximize AV utility by improving model accuracy via cloud

Innovation

• Intelligent bandwidth allocation across tasks
• Multiple data compression and model options
• Resource allocation to maximize car utility