This academic article discusses a novel utility-aware dynamic task offloading strategy for multi-edge-robot systems. The strategy aims to optimize resource utilization and minimize task latency by considering computation, communication, and task execution load. This approach involves continuous device, network, and task profiling, and assigns tasks to an edge with maximum utility using a weighted utility maximization technique. The proposed approach is shown to outperform conventional offloading methods in terms of optimizing resource utilization and minimizing task latency. The article also discusses the advantages and challenges of using cloud and edge computing for task offloading in robotics.
Publication date: 20 Nov 2023
Project Page: unknown
Paper: https://arxiv.org/pdf/2311.10196
