You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • Just to add to the previous discussion, both autopilots are fine. In fact, all autopilots are fine, all of them are quite similar. All autopilots have control, guidance, and navigation blocks. All of them are implemented quite similarly - linear PID loops, waypoint navigation, 3-4 control axes. But devil is in details. Reliability is a big question. Robustness is a big question. All of this takes time and code space. Mature autopilots, such as commercial Kestrel or free Paparazzi, are in development for more than 10 years. The basic structure is simple and similar to all others, but some code works in 99.9% scenarios, and some other code works in 80% scenarios. Would you like a malfunction once in 1000 flights or 1 in 5? This matters. Also, I am sure that Kestrel does multi-point calibration of all sensorsand test-flights their autopilots, which adds costs, but increases reliability, reduces burn-in period.

     

    Quite a big rant :) Sorry for a wall of text.

  • No, of course they don't use any ArduCopter code. I remember BYU university showing one of the first versions of Kestrel autopilot in 2003 or 2004 on IMAV hosted by us in the University of Arizona and even at that time they were going very strong. Of course, it didn't fly that time :) But at IMAV in Florida two guys from BYU showed up 1 hour before the flight, unpacked, flew beautifully without taking any prep time or test flight; then packed, and went back to Utah.

  • Just curious, have any of you flown the DJI S800 with Wookong and Zenmuse?  I was lucky enough to try one and I would put money on it if you did a multicopter Procerus shootout.

    -dave

  • T3

    I just visited Procerus web site and their Autopilot board called the Kestrel Autopilot cost $5000!  That's just the board, no frame, no power distribution, no motors, no ESC, no camera mount, just the board! 

    ArduPilot Mega 2.5 Fully Assembled System cost just $199!

    http://www.procerus.com/productPricing.php

  • Dealing with a little Fluid dynamics. One has to consider there are pressures that cooincides with area flow. All together with time it makes work in the form of Horse Power. Hp. If this efficent machine can do what it says it would be amazing.

  • 3D Robotics

    Ariley: Any chance you're in the SF Bay Area? I'd love to do a proper head-to-head comparison between the two. You may be right about the performance gap, but there's only one way to really know.  It would be good for us to know if there's a significant deficiency in what we're doing, or, who knows, it might turn out that we're now closer than people think.

  • Simul-post...

    So, Ariley, you've used the Procerus?  You mention hover performance.  What exactly does it do better?  What is the measureable?

  • I'm also curious what Ariley *knows* about the Procerus and what it's capable?  Is it just assumption?  Or is there more info available somewhere I haven't seen.

  • Chris, I've used both products. I've watched both closely over the past five years or so. You guys have a long ways to go, particularly on the hover side of things. I don't say this to be mean, and I sincerely hope you prove me wrong sooner rather than later. Procerus' products are priced absurdly, but that doesn't change the fact that they're the gold standard for small UAS.

    You guys may prefer "visionary to delusional" ... I prefer humility to unearned arrogance. I suspect it might do you good to see an Arducopter go head-to-head with the new Procerus system in some real-world missions, but I won't hold my breath on seeing that happen. Keep up the good work. Seriously.
  • 3D Robotics

    Ariley: You may be right, but how do you know? Have you seen the Procerus code? Last time we interviewed Reed Christensen from Procesus about the Kestral autopilot, he said the first version of that autopilot just used PID loops (not full sensor fusion algorithms). The more recent version introduced the standard Kalman/DCM style sensor fusion, but it was interesting to learn how similar the techniques were to what were already used by Paparazzi, APM, UAVDevBoard, GluonPilot and other open source autopilots.  We've been at this a long time, too ;-)

This reply was deleted.