AC13 Brookstone Rover

Hi, this is a great thread! I just got one of the AC13 Rovers and am curious to interact with it from a computer UI and more-so over WAN.  Is there any update on this? I am unsure how to send commands to the wifi_car_control.cgi.  I used the standard format (?command=4, etc.) and netcat to no avail.  I've looked at the RoverOpen Android project and tried to look at some of the source, but am not too familiar with java/C++.

Keep up the good work!

Hello, I see that you guys are trying to get into the Foscam (or similar) camera. My colleague and I have tried everything we can think of to try to extract the audio from the camera onto our iphones or androids. We know its in the MO_V, its G.726 but we cannot seem to extract the data. We wiresharked it and still are having a tough time pulling the  audio out. Please HELP!

I don't know about the rest, but #2 on your list (connect to router) seems to be possible with this url:

http://192.168.1.100/wireless.htm

I've gotten it to connect to my WAP, but it doesn't seem to have an option to enable DHCP.  Has anyone else tried this?

Very interesting, I've played around with it and looked at the source on the page and seems there is a ap_mode=infra option.  However, I haven't been able to connect to my AP yet. Any tips?

Hi all. Sorry for the near abandonment of the thread. Apparently my subscription to the thread stopped emailing me so I figured there was a notable lack of interest.

Right now I am fed up with the cobbled together nature of the camera application and have started down the road of replacing the main application wholesale with a light HTTP server (lighttpd, klone, etc.) with CGI support and necessary configuration options. To start off I am currently trying to get a cross-compiling environment set up under Arch linux but am running into trouble at the moment.

I have the development SDK installed but am unhappy with the build toolchain it provides, so I will be looking into using standard arm-elf-gcc tools to compile the executables and manually remix the rootfs partition. All told this will take a long time and I haven't a lot of time to devote, but I will update with any progress as well as put any information both here and source on Github.

Let me know if you're interested in this effort.

Seeing as its pretty much identical hardware, you may as well look at the existing stuff.

There is a 2.6 kernel, and a 2.4 kernel available, and compile tools + instructions for both.
Those are in the uCLinux forum, and in the wiki.

What SDK do you have?

If its different, suggest PM me, and I'll upload to the files section.

Yeah, I ended up reneging on my goals and backtracked into shoehorning the standard W90N745_PR BSP into an Arch Linux install. I was able to get Mathopd compiled using the ARM tools and I will be looking into reflashing the board with a few more tools in order to run some tests.

I am interested in getting a more recent kernel on the board than the standard 2.4 that comes with the BSP, but from what I can see on the forums so far the 2.6 and 3.0 kernels are currently having trouble with the wireless. Unfortunately this is a showstopper for me as this board does not appear to have an ethernet connection on it and I will need the wireless to be functional in order to communicate with it remotely.

I will report back with any more successes.

Progress is slow due to both my lack of knowledge and the available tools for OSX. One problem is the issue of finding a terminal program that has proper XMODEM send support- as most have already discovered the best bet is using minicom or ckermit with lrzsz which is itself properly borked due to flushing issues.

Oh a whim I went ahead and whipped up a quick XMODEM/XMODEM-CRC implementation in Python based off of pySerial and can be accessed/forked at https://github.com/ArminTamzarian/txmodem. I can now upload to the rover's memory space using the non-invasive mx 0x8000 but all I end up with so far is a g 0x8000 execution that can only be described as a giant ball of crashfail or silenthang.

So now it's a matter of backing things out and getting to a mode where I can upload and execute something useful before forging too much further ahead. More updates as they come...

Hi,

• The power leads are connected directly from the battery compartment (6AA -> ~9.0V) and soldered onto the main board instead of using a plug or connector. Also, I didn't look too closely but I did not notice any power regulation circuity, so be careful what you feed this device.

The switching regulator is nearby the elkos. The elkos themself and the coil are part of the 3.3V regulation circuit.

• There is a decent amount of space in the shell behind the battery compartment. Probably enough to hold a properly size Li-Po battery cell and charging circuitry.

The space is definitely big enough to keep 3*2 flat Li-Ion cells. I added an additional regulator for reducing the voltage to ~10V to prevent the motor drivers from to high currents. This is located on the left motor/gear-block. The battery spce is now to hold the cables for charging the Li-Ion cells with an external RC-charger.

BR,

slarti

Thanks for the info slarti. Didn't get too much of a chance to trace a lot so good find on the regulator. Do you have any pics of the guts with the batteries in place and additional regulators? Would be nice to get a schematic with pins specified along with part lists for others looking to mod in this same fashion.

What li-ion cells are you using or are you just leveraging a standard rc batter pack with normal connector and charger?

I used mathopd too

Actually compiling tips for that are in the forum.

I'm working on the 2.4 side a bit more these days, as I hired a student developer to work on things.
I think I'm learning more about how much I actually know showing him the ropes, so thats good

Right now doing some GPIO related stuff, I'll probably release that soonish up on github, as its generically useful.

Also looks like I'm finally getting compatible with the "original" linux image, as I can get everything up using that now with my own binaries finally.  Took me a while, and quite a bit of tracing through disassembler though

I used mathopd too

Actually compiling tips for that are in the forum.

Hah! Never thought to do a search on the forum for that specific software. Although now I see that multiple people have come to the conclusion that mathopd is a good solution for the web server side of things, so I guess we've all independently verified its usefulness. I think I already have it compIling to a useful state, but when I get some time I will actually try uploading it to the rover and see his it performs. However, part of that will require getting the wireless networking functioning with my router. So small steps.

Good to we are all in the same page thought will be very interested in the code for the GPIO as I will need to use that via cgi for wheel control.

Armin and others - great job with this thread. Just saw it today and I love it...

I worked on decoding the binary communication protocol described on the android forum and now look fwd to diving deeper with renewed enthusiasm.

My goal is to have complete control over the AC13 and the ArDrone and them making them talk to each other someday - hardware wise they both are very similar technically - except one is a ground vehicle and the other flys!!!

I came across your project and wanted to ask a couple of questions... I have the AC13 rover also and have pulled it apart to use it in another project.  It is working great!  I'm not experienced with modifying code, electronics, etc.  Mechanical modifications are easy though.  I have two questions...

1)can the Loftek camera be used to make a "rover"?  Actually, can I use either of the motors that pan and tilt to turn a single wheel, or could the control signal for pan or tilt be used to run a new servo or stepper motor? 

2)can the Loftek camera connect directly to a device like a smart phone without going through the internet (like the rover does)?

Sorry for the really basic questions.  I can't afford to waste money on a camera that I can't use.

My project...
I need to find a way to view camera feed and turn a simple wheel when I'm out in the middle of no where without internet connectivity (I'd still have access to my smart phone's wifi capability though, just no network to connect to unless the camera broadcasts one).  I can do this using the AC13 parts, but I'm trying to build a second cheap prototype with parts that are readily available and relatively cheap (less than the rover).  I don't have experience to write my own android or iphone app or to alter code or scripts.  I'd be willing to pay someone that could create this type of system for me though.  I know I could do the same type of thing using an RF camera and monitor and an R/C RF receiver to run a servo, but that would require an R/C controller.  Using a smart phone is much easier - no need for another "monitor" or controller.  I'm making something that my brother needs to be able to use out on the farm (no cell coverage).  He has a smart phone and an ipod touch, but not R/C stuff.

Thanks for any help you can give!

Admin, Armin Tamzarian,

Would be grateful if you could please provide information on which port/connector on the ac13 was used with the Bus Pirate and what protocol was used to extract info? I have some cycles to dig into other stuff that may be of interest.

Webhansen,

My interpretation of Armin's first post is that AC13 is basically a clever copy of the Loftex camera to drive left/right motors instead of tilt/pan motors. All inside a new tank casing. So you could theoretically convert the ac13 to an ip webcam and webcam to a tank


Thx, nanu