DAG-Based QoS-Aware Dynamic Task Placement for Networked Multi-Stage Control Pipelines
This paper proposes a new framework for dynamic task placement in networked robotics, specifically for multi-stage control pipelines. The framework uses a directed acyclic graph (DAG) to model the pipeline, considering attributes like compute cost and communication delay. It aims to optimize task placement by balancing factors such as latency, hardware utilization, and the cost of switching placements, with a focus on reducing chatter in industrial automation settings. AI
IMPACT Introduces a novel framework for optimizing AI task placement in robotics, potentially improving efficiency and reducing latency in industrial automation.