John Wiseman and Pat Hickey of the Mavelous team have been making great strides with this HTML-based MAVLink GCS, which can run in any browser. You can see a very cool live demo of it here, and follow the project here. 3DR is supporting the project, so look for products from us soon that will allow you to use any device, including phones and tablets, as your GCS for ArduCopter/Plane in the field.
Comments
Totally agree. I've found two major issues, rendering speed of gauge changes (specially for the complex ones like attitude) and coping with the amount of data. I'm supporting 8 gauges right now and initially found that SVG transformations work nicely on the desktop but mobile devices just don't have the grunt at this point in time (for webkit/html5), CSS manipulation for images works far better but again, still some performance issues. There are also some inconsistencies at this point between SVG handling for browsers alas...
I've yet to do some testing to see if the issue is to do with the frequency of events or the css transforms. I want to try doing the gauges as a canvas next, as I know that will be pretty slick. I'm sure it won't affect the drag and drop of the gauges, just need to experiment when I get time.
In regards to data, I use websockets to communicate with a node.js based server. At the server I filter out too many updates which increases performance, but again, I'm not 100% certain if the issue is rendering or event handling. When I have time I'll sort it out.
I also tend to think that there isn't much that a ground control station needs to do that requires super high performance.
No worries Dennis, thanks for the heads up, though I'm aware of the limitations vs native. I'm just happy to use HTML5, CSS3 and run in webkit (which is all phonegap is in all honesty, with some extra bits for hardware access).
Right now I prefer to run it in a chrome browser as I find that nicer than the phonegap implementation.
@James Jenner:
I've tried making a simple HTML5/JQuery mobile app with phonegap. Just a warning: compared to native code it is painfully slow. Rewriting the app to java made an incredible difference. So, just to share my experience: if you at all can, go native on the mobile development.
Interesting approach. Funnily enough I've been working on a similar project but with a different bent for the last 9 months. My approach has been to be non protocol specific automated drone support with a multi client touch based interface for monitoring and interaction. The first stage was to support the Parrot AR Drone (v1), providing telemetry and interaction (as the parrot v1 isn't autonomous at the least). I'm in the middle of the next stage which is incorporating mavlink support. The goal is to be AV agnostic (not specific for UAVs).
Right now mission planning (limited to geographical commands), multi drone management and telemetry monitoring is the limit of the software.
One other goal was to use PhoneGap to build android and iOS apps so it could be run on phones natively. As such I've limited the project to HMTL5 which requires latest versions of various browsers.
Would love to collaborate and see if we can find some common ground to share stuff :-)
I had an idea in the shower the other morning ( all my best ideas arrive while I shampoo ). I thought it would be cool to send a Drone either across the US or around the world.
GeoCache people do this all the time. They take a little toy pirate or something, put a note on it "Send me to Australia", stick it in a cache and it'll eventually hitchhike around the world. (Geocache)
So, back to drones.
To do a cross country flight, it would take a lot of people to retrieve, charge, reprogram, and launch the drone. But if there were a website where people could watch the flight and know where to go to help the drone do another leg (and sign the drone log book), it might just work.
I have a "dumb" Slinger I would donate.... Never flown. I have no idea how far a drone can fly. I've seen some mentions of 200k, can that be true? I would think even a 100k would be stunning.
Any ways, that's what I would do with the Mavlink web browser.
Ive been looking at the smaller of the 2 verizon 4g LTE Jetpacks. It is certainly small enough to fit within a bixler, and could easily be mounted to a drone. If the web server board just had Wifi, so it could talk to the jetpack, it could easily be accessed by any Internet connected tablet.
Is anyone currently working on a way to hook a mavlink module to Android tablet? Pretty much all of the tablets have some sort of adapter available to from there proprietary ports to USB so you could probably use the same module on most tablets. And although keeping up with all the new tablets would be time consuming, I for one would be willing to pay for the app to keep up with supporting new devices.
@John Wiseman:
Yup, that too ..
But think of simply uploading some log files to a site, and then later be able to replay the flights. That would be a nice and convenient way to document flights, I think.
Maybe find a way to sync the output with recorded video and such would be another cool thing.
That is awesome. Great work
-
1
-
2
-
3
of 3 Next