next »

[Ben]

I'm looking at this as a project to teach myself a few things.  I would probably get more benefit out of adding a Bluetooth interface to the XO, but I want to play with GPS, so that's where I'm going.

  How about combing the two and install Bluetooth and using a Bluetooth GPS?

[pgf]

@matt -- the micro i'm thinking of is literally only 8 pins, and has no uart.  it would do nothing but monitor transitions on the TX line, and power up the GPS when it saw enough of them.  it would have no smarts, nor any expansion capabilities.  and it rwould be tiny.  something to run an LCD, or other stuff, would have to be reprogrammable (since the code would be complex enough to have bugs), and it would quickly turn into a separate project.

@ben -- bluetooth implies using USB, i think.  quite possible, but again, maybe that's another project.  and it also implies an external gps unit, something i'm personally trying to avoid.

paul

[mattd]

something to run an LCD .... would quickly turn into a separate project.

Where do you think I'm trying to steer this?! 

Really, though, the LCD and the other flights of fancy would require different micros and development. My point was that by persuing this down to the point where we could start development - and maybe even implement it on an XO - would open the door to dozens upon dozens of add ons.

But for this, if we really really wanted to follow KISS, you'd just get the tiniest white two pole pushbutton switch that you can find and mount it somewhere discrete to kill power manually.

sloe, since you're the only one with an XO, is there somewhere on it that is flexible or able to be pressed that you could mount a PB switch UNDER the plastic somewhere? That'd do away with the hole.

[pgf]

someone on the atmel avr forum pointed me at the right part of the datasheet.  we can't just parallel the i/o ports to supply power, because the sum of current from the pins can only be 60mA.  so scratch that idea.

paul

[Tortuga]

Endgadget has an article on a USB TiltStick [accelerometer]
http://www.engadget.com/2008/03/10/usb-tiltstick-opens-up-new-world-of-accelerometer-projects/

This might be a possible alternate solution for the mini joystick ?

"Reportedly, the minuscule device features a two axis acceleration sensor and emulates a USB joystick in order to play nice with any USB-equipped host."

[pgf]

Endgadget has an article on a USB TiltStick [accelerometer]

interesting!  thanks for the pointer...

paul

[Patrick]

Well here I am chiming on at the present end of the thread. I have an XO and a USB GPS and am ready and willing to start trying code(s). I am not intereseted in hacking the XO box but am willing to beta test. I have acces to at least two different GPS USB units and have a serial/bluetooth unit I could use as well. Unfortunately I am windows savvy only.

I am very curious about the added possibility of leveraging GEOPDFs as basemaping as well. USGS has created 24k quads as geopdfs and terrago has created a free reader plugin for geopdfs - juat need to get the native XO PDF reader to read GEOPDFs.

I work in GIS so mapdata and GPS toys are relatively easy to come by.

Anxiously awaiting the *.xo gps solution.

Thanks,

Patrick

[pgf]

your bluetooth thing won't be much good until someone builds bluetooth support into the kernel.

i have a program called RoadMap running on the XO.  i'm waiting for my new, hopefully reliable, usb gps unit to arrive from amazon so i can finish testing  it runs as an activity, and displays maps from the census bureau's Tiger data or maps created from the OpenStreetMap data.  i'm sure you could also get something like GPSDrive running on the XO pretty easily.

i'm not going to have much time in the next week to publish my RoadMap as an activity, but i do intend to.

paul

[Sarah Elkins]

Paul, I'm looking forward to your RoadMap activity!

So far I've just used a Garmin and (just yesterday) a Magellan.  A friend of mine says Tom-toms are Linux-friendlier than the other car GPS's out there.  I don't know if I'll be able to get my new Magellan working with RoadMap -- I'm not sure I understand whether RoadMap is supposed to work *with* other GPSs, or all on its own -- but it'll be fun to play with.

[pgf]

hi sarah -- it works on its own.  you can see the full project at http://roadmap.sourceforge.net

it doesn't compete with commercial gps units, not by a long shot -- 5 years ago, it might have, back before auto-routing was still on the horizon.  but it displays a nice moving map, and can keep track of waypoints.  you can even load up a route that you've downloaded from google, and it will help you follow it -- but no "you made a wrong turn, recalculating".

paul

[tangomike]

I've installed GPSDrive on my XO with Ubuntu. Works great. I attached a Garmin Rino and Magellan Color using a 9pin serial to usb adapter. GPSDrive worked with both (one at a time) of them, using NMEA protocol. Lotsa free maps, and the XO screen is a treat for my old eyes.

[arterich]

Very interesting discussion but I personally don't see what purpose is served by mounting the GPSr in the XO?

What is the intended application of the GPS enabled XO?  It's too large and bulky to be used as a handheld GPS while hiking so the only application I see for it is for auto navigation.  If you mount the receiver in the XO, won't that mean the XO will need to ride on the dash in an auto?  Many autos don't have enough room on the dash for an XO.

While reading this thread I got to brainstorming about how an integrated GPSr can be incorporated into various projects on the machine.

Theft Recovery
The first thing that popped into my head is a LoJack (http://en.wikipedia.org/wiki/LoJack) style application which you can use to track your XO in the event that it is stolen.  If the XO is running it will turn on the GPS and check it's location every 10 minutes.  That info would be appended to a file.  When the XO senses that it is connected to the internet it would send this data to a pre-determined email address or use an HTTP POST to send the info to a web repository.  It can be hard to track a machine if the info is only sent if the machine is online and a GPS lock is available, but with the method above you can not only see where the machine currently is but also where it has been.  For instance the coordinates available when connected to the internet might place it at a library or cyber-cafe, but the coordinates that are recorded when offline place it at a person's home.  The tricky part is how the machine would know that it's stolen given that there are no passwords.  I haven't read up on BitFrost but I figure it could somehow be integrated with that.  Alternately the machine could be setup to always email the coordinate info.  That info would take up very little space so it wouldn't fill up a person's inbox.

Network Mapping
The current Network View shows Wi-Fi access points and other XOs which are nearby.  This view isn't accurate in the sense of physical locality.  With built-in GPSr you could know where other XOs are physically located relative to your machine.  And if three or more XOs are in your area they maybe be able to triangular the location of the Wi-Fi access points relative to the XOs.

Open Network Database
When sitting at home you have easy access to the internet via your local Wi-Fi connection.  Likewise you may have a favorite local coffee shop that also provides free Wi-Fi.  But what if you are traveling out of town or out of the country.  There are often plenty of companies or organizations providing public Wi-Fi but the hard part is knowing where.  Each time your machine connects to a public access point which doesn't require a password, the XO would make a note of the network SSID and the XOs current location.  That info would then be uploaded to a database along with a "Notes:" field that the user could fill out if they wanted to.  The field would contain stuff like who owns the access point and whether there are any restrictions.  Any other XO owners who use this app would have the database synced to their machine on-command or automatically when they have accees to the internet.  Then the next time they are somewhere unfamiliar they can see their location and on a map where all the local public access points are.

Only the Theft Recovery idea would rely on the GPSr being inside the machine as the other two ideas would work just as well with an external USB-based unit.  Also all of these activities are dependent on a relatively mobile society.  It's important to remember that those of us who participated in the G1G1 program move around quite a bit thanks to cheap transportation whereas the target audience for this device, school children in developing countries, probably don't travel nearly as much.  As someone mentioned before, the XO could handle GeoCaching and general mapping pretty well thanks to the high-res screen and low power requirements.  However, the XO would be a device that you take out every so often to check your progress vs. something you would be looking at all the time.  As for the car navigation, it's really note economical.  You can pick up an in-expensive portable navigation system for about $200 which already has tailored software and maps.

Omar

[Gabey8]

The tricky part is how the machine would know that it's stolen given that there are no passwords.

How about when the machine goes online, it quietly checks a particular site to see if it's been registered as stolen? (Insert joke here about looking for one's own name in the obituaries first thing in the morning, to make sure one is still alive. ) If it sees itself listed as stolen, then the GPS/theft-recovery tracking can begin. Ideally, this check will be completely transparent to the user, so that an unauthorized user won't even know that the process exists, nor will they attempt to disable it.

Each user of this anti-theft system can have their own account where they register their XO. If the XO goes missing, then they can log onto their anti-theft account and report it stolen. Then the next time the XO goes online and checks the site, it'll see itself listed and that will trigger it to start recording its location.

[pgf]

regarding "integrated GPS" -- i'm becoming convinced, having now tried two different GPS units (one USB connected, with a sirf III chipset, and the other bluetooth-connected, with an mtk (i think) chipset), that a built-in GPS unit won't work.  neither of my GPS units will properly pick up a full complement of satellites unless the receiver is moved some distance from the laptop.  this might be fixable with a big external antenna mounted on the outside of the laptop case, somehow, but for me that defeats some of the purpose of "integrating" it.

i think the next best thing (from a convenience point of view) might be to build in the USB bluetooth dongle instead, and carry the GPS unit separately.  (but i'm currently experiencing very long latencies (lag times) between moving the unit and getting the actual report.  i don't get these delays, which are about 3 or 4 seconds -- a long time when trying to figure out what turn to make -- with the direct connected unit -- just the bluetooth one.  i haven't had time to figure out where the delay is coming from.)

paul

[Tortuga]

Not knowing much about this, but have you considered that the mesh networking may be a complicating factor? It seems to be easy to turn off, might be worth trying?