There have been multiple discussions in the threads of various ArduStation posts in the last few weeks about form, function, capabilities both in conjunction with and without a GCS or planner. I have joined in on some of those discussions and what came out of many of them is that we all have different approaches to what we want out of an ArduStation-like component. Some people want a simple mechanism to just decode the LAT/LON, calculate a direction and point the antenna. Some people want it to point the antenna and pass along the data stream via another communication protocol. Still others like having the screen and buttons to tune PIDs, etc. Some wanted the whole enchilada PLUS a touchscreen to do away with the old text screen and buttons...needless to say there was not a lot of consensus - so that got us talking and a modular format became a part of the discussion.
So what I would like to propose is a new project called the ArduStation HUB. A completely modular unit that will allow people to build upon it and add each component they require. So chime in, tell me what you think. Do we even need anything like this? Is my design brilliant? lunacy? Let's crowdsource the design and I wil get them built - if we come up with something truly awesome, perhaps we can get it added to the store. I know MANY of you have been working on various versions of hardware / software etc. Can we come up with the modular software to match?
Comments
@Suayb - that was actually going to be my suggestion, that you use a very small form factor PC with a VGA output and a USB video capture. Then you could use the OSD built into the GCS (I just saw it running this morning here) If you resized the left hand panel to be full screen (which may not be possible??) it would make a great HUD for FPV and have really nice resolution.
@Scott
Thats what I mean VGA output with OSD. We are using portable monitors or goggles. All of them digital devices but we are still using analog video output for monitors and goggles. Yes I know our video feed from video RX is analog but analog video output (lite composite) give very bad text and OSD details.
I want to use headplays vga input for better fpv experiment. Or 10-15" lcd screen with colored real maps and much smooter icons and OSD feed. But also I know VGA output will be expensive. But It can be separate module.
Mine is just dream. An mini-itx pc system with touch screen assembled in small suitcase with all extra electronic modules. :)
@Andrew - Actually it was your post, along with discussions on several other posts about peoples own iterations of the ArduStation that got me looking into upgrading the station again. I was a part of the team that was building a "new" ArduStation Mega over a year ago, but we never got beyond the hardware because we didn't have the software chops to get the graphic LCD up and running correctly. Now I believe we can get it done with this modular approach. We WILL be including a version that uses the 20x4 screen, so we will likely be diving into your code. It is truly impressive work. Also, to clarify, we're not talking about using a standard Arduino shield, but building a custom shield (like the oilpan IMU) for the APM; the obvious difference being that this one will be designed for a ground station so it will have a lot of sockets for people to plug things in, and no sensors of its own.
I know the 3DR team is hard at work on getting out an XBee<->Bluetooth Bridge we have been collaborating on and that should make the use of a Playbook an easier thing to accomplish, along with the great work that Bart has done on the Android side for those devices.
I like the modular approach. My (more primitive) try at this is described at http://diydrones.com/profiles/blogs/ardustation-as-a-terminal. It is based on an APM board with an Xbee for the data - works fine running in parallel to an Xbee to the mission planner. The UI is via a traditional Ardustation board, but this is not required to be connected for the other features to work. OSD is from a Remzibi driven by one of the serial ports. I would really prefer an OSD with graphics capabilities rather than only characters, but haven't had much luck in finding one. Antenna tracker code is (still) in work - the core is done, but am working on pages to support calibration of the two axis. The box will be attached to my tracking antenna, and it has the two video RX and the Oracle diversity switcher integrated, so basically it is one power lead that goes in to the tracker and video/audio that come out. All the video stuff is on the rotating part of the azimuth stage and the power in/video out are on the fixed stage, with wires going through a hollow drive shaft, so should be quicker to setup than my old toolbox filled with individual components and lots of cables.
I don't know if the APM board will accept the regular Arduino shields, but if not there is nothing critical about using the APM - it could just as easily be based on a standard Arduino Mega card (which are described as accepting the normal shields, although I have not tried this). SD cards, or whatever, could be added that way.
I noticed that RIM has just released the native SDK for the Playbook with OpenGL support for the upcoming OS 2.0. That would be nice for the GUI using a Bluetooth link (from an Arduino shield). The recent 40% drop makes this more palatable than it was a month ago (as long as it lasts) . Presumably Android has something similar.
@CW I agree the SD Card slot is a must have, but many people might not want/ need it so modularity is the answer there. We are trying to provide a platform for building on top of some versions of existing hardware (AP, APM 1280 / 2560) so upgrading this to handle a framework like that is likely out of scope.
An SDcard is need for those that need a real blackbox recorder mechanism.
The hub is a great idea. Though one should consider what the mavlink folks are doing to support modularity as well as the ROS folks (their architecture is basically a hub, aka bus, problem is that it's a very heavyweighted framework).
If you guys have every programmed Java's JMS or even JMX, we should be looking into that for a starting architecture.
And use the video link as a watchdog [timer], we've done it many times here for other projects.
@Andy - You may have seen that TV show called "Hoarders" - that is me with data. I try to have a logger of some sort on as many of my flights as I can, and REALLY like the idea of having that data collected on a simple device, even if I don't have my tablet PC available. I think we design it to be modular so those that don't want it don't have to spend the extra dollars, but I certainly want it!
@Alexander - VERY cool work, I'd love to work with you and see where we can join forces.
@Jeff E - I agree that our current ArduStation is a powerful piece of hardware, and that we can do a LOT with it. I think there are a couple of points that we need to bear in mind. The reality is we may be running up against its limitations, so the thought was to do something more powerful, based on recycling some older APM hardware as people move toward new ARM based hardware. I am really excited for what the migration to ARM means! I think you are exactly right when you say "what is the target?" and I think our answer is "modular." I want people to design their own target, and for this HUB to be able to act as the connection point for whatever they need to do. That may prove to be an insurmountable task, but I'd really like for someone to make this as simple OR as complex as THEY want to make it. I see this as being a board with a lot of connectors, clean power handling, and a USB interface - pretty inexpensive, then people simply add on the modules they want and load the software modules they need to make it work, and they are good to go.
- So at risk of sounding like yoda on a zen trip "the target is no target."
@Scott - Initially, I didn't see real value to an SD card. But there have recently been several airplane recoveries using the GPS, heading, speed data from recorded OSD video. Saving Mavlink nav data to SD logs would be a backup/alternative to GCS logs. Together with the battery state and current draw, you could even predict the landing area if telemetry contact was lost while in flight. Not perfect but better than hiring a 172.
-
1
-
2
-
3
-
4
of 4 Next