Question: Does anyone know about a platform to manage Edge+Robots

I’m looking for a platform that can manage Edge devices and Robots connected to these Edge devices.

  1. Robot is connected only to the Edge - not to the public Internet
  2. Edge is connected to the Internet
  3. Both of these can be monitored, i.e. metrics (cpu, memory, etc) is collected and then can be visualized through graphs (alerting is a plus)
  4. You can do filesystem operations i.e. upload and download of files to Edge and Robot

Do you know of any such platform that provides these? Are you too looking for such a solution?

4 Likes

Interesting. I am not sure what needs to be secured from or exposed as API to the public network. but here are my thoughts, hope this would help a bit.

Both of these can be monitored, i.e. metrics (cpu, memory, etc) is collected and then can be visualized through graphs (alerting is a plus)

I would use Prometheus Node Exporter to collect these statistics, or probably Open Telemetry. We can also use Grafana for monitor and visualizer and AlertManager to detect the alerting based on the statistics.
If you want to collect ROS 2 statistics, you can refer to Topic-Statistics-Tutorial — ROS 2 Documentation: Foxy documentation, be advised that statistics are available in C++ only. and these statistics cannot be collected on Prometheus unless we develop the ROS 2 exporter implementation for Prometheus.

tomoya,

2 Likes

Hi @shivamMg,

As Robolaunch team, we are providing a platform that capable of exactly what you’re looking for and also some other functionalities as well.

Our main focus is to eliminate the challenges of building robotics application and operating the physical robots through out Robolaunch Cloud Platform. Our robot deployment model is based on git repos and it generates robots from it’s ROS code. In this scope, users can develop, deploy and operate their ROS & ROS2 based robots’ life-cycle & FCAPS management decleratively.

Unified robotics operations with Roboulaunch:

  1. Design: Create 3D robots and worldviews for different use cases
  2. Develop: robotics software development through Cloud based IDE without any manual software and library installation
  3. Simulate&Train: GPU accelerated re-inforcement learning and hyper-realistic simulations
  4. Deliver: Deploy robotics application to the physical& virtual robot(s) and their digital twins decleratively
  5. Operate: Configure, manage and monitor multiple robots and fleets in run-time

About Robolaunch’s network isolation and edge computing architecture:

You can also learn more about Robolaunch and the updates here.

Kind regards,
Özgecan

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.