GSOC'26 : Understanding Community Expectations for Companion-Computer Health Monitoring

Google Summer of Code
1

Hello,

I’m a BTech student currently working with ArduPilot, Pixhawk, and companion-computer integration, and I’m very interested in the “Real-Time Companion-Computer Health Monitoring & Failsafe” idea.

I’ve been going through the existing failsafe mechanisms and MAVLink handling, and I wanted to better understand what the community expects from this project in terms of scope and integration depth.

Specifically:

  • Is the expectation to primarily monitor system-level metrics (CPU, RAM, process state), or also application-level health (e.g., vision node heartbeat, MAVLink responsiveness)?

  • Should the companion-triggered failsafe logic remain fully external (via MAVLink commands), or is tighter integration into ArduPilot’s internal failsafe framework preferred?

  • Is there an existing discussion, prototype, or past attempt related to companion health monitoring that I should study first?

I’d like to align my approach early with community expectations before developing a more concrete proposal.

Thank you.

1 Like