Safety-Critical Edge Robotics Architecture with Bounded End-to-End Latency

• DOI: 10.48550/arxiv.2406.14391
• Publication Date: 2024-06-20
• Authors: Gautam Gala; Tilmann Unte; Luiz Faustino dos...; Johannes Kühbacher; Isser Kadusale; Mohammad Ibrahim ...; Gerhard Fohler; Sebastian Altmeyer
• Keywords: End-to-end principle

Edge computing processes data near its source, reducing latency and enhancing security compared to traditional cloud while providing benefits. This paper explores edge for migrating an existing safety-critical robotics use case from onboard dedicated hardware solution. We propose architecture based on Linux, Docker containers, Kubernetes, a local wireless area network the TTWiFi protocol. Inspired by previous work real-time cloud, we complement with resource management orchestration layer help Linux manage, Kubernetes orchestrate system-wide shared resources (e.g., caches, memory bandwidth, network). Our aims ensure fault-tolerant predictable execution of robotic applications path planning) upper-bounding end-to-end ensuring best possible quality service without jeopardizing safety security.