AdaSprite: Resource-efficient Online Co-Adaptation for V2I Systems Under Large-scale Data Drifts
2026-08-17 • Operating Systems
Operating Systems
AI summaryⓘ
The authors address challenges in processing many vehicle-to-infrastructure (V2I) video streams on limited edge servers using vision-language models with Vision Mixture-of-Experts (V-MoE). They identify problems like memory fragmentation, slow data access, and task switching delays that reduce efficiency and accuracy. To fix this, they propose AdaSprite, which smartly manages resources and schedules tasks to improve how many V2I tasks can run at once and how fast they complete. Their system supports nearly triple the concurrent tasks compared to existing methods while balancing accuracy and speed.
Vehicle-to-Infrastructure (V2I)Vision-Language Models (VLMs)Vision Mixture-of-Experts (V-MoE)Edge ComputingDRAM FragmentationSparse Expert RoutingTask SchedulingCo-adaptationResource ConstraintsLatency
Authors
Lehao Wang, Zhiwen Yu, Sicong Liu, Kefan Chen, Fengmin Wu, Bin Guo
Abstract
The rise of vehicle-infrastructure (V2I) collaboration enables safer and broader perception. To process large-scale V2I video streams, vision-language models (VLMs) are promising as they unify multi-view vision into end-to-end task grounding, reducing handcrafted design. We use Vision Mixture-of-Experts (V-MoE) as the distributed visual backbone of VLMs, leveraging sparse expert routing to enable conditional computation across diverse viewpoints under resource constraints. Yet, V-MoEs face a critical challenge: large-scale data shifts over minutes to hours in V2I systems, amplified by agnostic participants and biased features propagating through experts. To maintain accuracy efficiently, we find it beneficial to co-adapt multiple V-MoEs on edge servers, avoiding the latency and privacy risks of cloud offloading and the accuracy sacrifices of on-device methods. However, the resource-constrained edge poses challenges for efficient co-adaptation: i) DRAM fragmentation and imbalance limit expert parallelism, ii) memory-I/O bottlenecks restrict computation reuse, and iii) asynchronous adaptation increases task-switch overhead. Also, prior work rarely explores the upper bound of concurrent tasks under limited edge resources, a critical factor for practical V2I deployment. To address these, we present AdaSprite. By combining cooperative elastic scaling with multi-level multiplexing, AdaSprite optimizes expert lifespans to reduce DRAM fragmentation, exploits predictable activation patterns for efficient I/O reuse, and employs twin-buffer scheduling to leverage sparsity. On a weak edge, AdaSprite supports up to 17 concurrent V2I tasks (vs. up to 6 for baselines), improving SLO attainment by 1.6x and throughput by 2.1x. Also, it allows users to trade accuracy and concurrency for second-level adaptation.