Integrate ROS into EdgeX - Startover of Edge Computing System
Talkat Open Source Summit Europe + Embedded Linux Conference Europe 2019 · on Wednesday, 2019-10-30· at 12:20· Forum 6·
summary
Shows a concrete integration path: ROS device nodes as EdgeX device services register callbacks that translate ROS topics into EdgeX readings, using a local MQTT bridge to pass sensor data, and OpenCV-based inference with hardware-accelerator model conversion to run vision models on an attached Google NCS-style accelerator. Uses a minimal container-friendly Linux image and explores preempt-rt / tiny VM deployment for lower jitter. Demo is a camera-equipped robot pipeline; lab runs show acceptable inference latency while office-networked deployments suffer noticeable delays.
AI-written summary · about
chapters (18)
- 0:00Intro
- 2:31What is ROS?
- 3:38ROS Filesystem Level
- 4:32ROS Computation Graph Level
- 6:09ROS Ecosystem
- 6:45Why is EdgeX?
- 9:34Edge X Architectures
- 10:54EdgeX Enables Tiered Deployments
- 11:50A Global Partner Ecosystem
- 12:00Why to build ROS into EdgeX
- 12:58Approaches of Integrating ROS into Edgex
- 13:34ROS as a device service layer of Edge
- 16:31Privileged Device Service Solution
- 18:31Enable ML Inference framework III
- 22:10Built-in RT: Enable preempt-RT
- 22:35Built-in RT: Preempt-RT Linuxkit
- 24:26Vision of Intelligent Edge Computing
- 27:40HW Setup Environment
related talks
30mRobot Operating System (ROS): current and future capabilities on embedded systems
22mKubernetes Robotics Edge Cluster System
57mThe Convergence of Robotics, the Web, and IoT
34mMicroROS, combining advanced robotics and low-cost embedded systems
48mBuilding Robotics Applications at Scale using Open Source from Zero to Hero
43mPatterns for cloud-enabled robots
01h 16mKeynote: Yes! Robotics for the C++ Developer
57mIntroduction to the Robot Operating System (ROS) Middleware