Karshipta update: multi-vehicle fleets, calibration, flight logs - looking for real PX4/ArduPilot hardware testers

Update on Karshipta since the July post - a lot’s changed, and now I actually need real hardware to test it properly, not just SITL.

Since v0.1.0:

  • One gateway can now manage multiple real vehicles at once, not just one - this touched the gateway, our desktop pairing app, and the relay layer, a bigger change than it sounds (v0.2).
  • Enforced geofence, pushed to the flight controller, not just drawn on the map
  • Custom home position + RTL altitude control
  • Parameter read/write for safety-critical and PID tuning fields
  • Flight log download straight from the console
  • Guided pre-flight sensor calibration (gyro/accel/mag/level horizon), with live progress

Everything above has been tested against SITL. Real flight controllers, properly, are the actual gap right now - this hasn’t been field-tested at scale, so treat it accordingly.

If you’ve got a Pixhawk or similar around, here’s what would genuinely help, lowest-risk first:

  1. Props off: pair it with our gateway over USB or a telemetry radio; try arm/disarm, parameter read, calibration. Zero flight risk, 15-20 minutes.
  2. If you’re already flying something small anyway: try commanding a flight you’d be making regardless - takeoff, land, a basic mission - through Karshipta alongside your usual GCS.

Use your own judgment and your normal safety practices either way - don’t do anything with it you wouldn’t otherwise do with the aircraft.

Live demo: karshipta.com (spin up simulated vehicles in your browser)
Repo: github.com/NIKX-Tech/karshipta - open source, AGPL-3.0, self-hosted
Changelog: link

Feedback helps too, not just hardware time - open to anyone poking at the code or the UI either way.

Hi Erfan,

Great progress since the original announcement. The multi-vehicle gateway, parameter management, calibration and log-download capabilities are particularly interesting.

I reviewed the v0.3.0 changelog and issue #268, which you opened following our earlier discussion about telemetry timestamps and tracker staleness.

From our side at PAMILANGA, I think there are two useful areas for collaboration:

  1. A reproducible validation matrix for telemetry freshness, delayed messages, disconnected trackers and recovery behavior.

  2. Multi-vehicle SITL regression scenarios, particularly simultaneous connections, reconnects and gateway behavior under partial link loss.

These could be useful before moving to real flight controllers.

I would be happy to contribute to the test design and review results as the implementation progresses.

For real-hardware testing, we would first need to confirm a suitable board and setup. I would prefer to begin with non-flight tests, with propellers removed, before considering any further integration work.

Congratulations on the progress, and looking forward to seeing how Karshipta develops.

Best regards,
Yiğitcan Öztürk
PAMILANGA LTD

Hi Yigitcan,

Appreciate the detailed read-through. Opened a tracked issue for the multi-vehicle SITL scenarios too: #270 - simultaneous connections, reconnects, and partial link loss are a real gap I hadn’t scoped out yet, separate from the timestamp/staleness work in #268.

Both of those are exactly the kind of structured testing I don’t have the bandwidth to design well alone, so the offer to help with test design and review is genuinely useful, not just a nice gesture.

Agreed on sequencing - gateway/SITL first, real hardware once there’s a board and setup worth committing to. I’ll keep both issues updated as the implementation moves.