On Wed, 14 Jul 2004 17:03:10 -0400, jon@dakota-truck.net wrote:
> Ok, I hadn't realized that (wireless newbie) :-) - that
>"repeater" mode would sure make things easier. :-) Basically
>it would accomplish the same thing I was talking about but with
>a lot less complexity on our part, and that's a whole lot of code
>that wouldn't need to be written. :-) See, this probject is
>getting easier and easier! :-)
>
> As far as the "repeater" mode goes, is this something that
>the access point does, or can a laptop with a wireless card in
>it perform the same function? That is, can a group of laptops
>with nothing but PCMCIA 802.11B/G cards talk to each other as
>if they were connected via a hub on a LAN, or is at least one
>access point required?
In theory, this is how an "ad-hoc" network works. I'm not sure I
would trust that, though. For one thing, PCMCIA cards (except very
expensive ones) have relatively low power output. A couple of
companies make "long range" PCMCIA cards, but they are typically in
the ~$200 price range. A Linksys access point can be had for 50-60
bucks new and will do the job just fine. It's also 12-volt so it'd be
pretty easy to rig up some cigarette lighter adapters for them. You
can create a more efficient bridged network that way that would
probably allow for more data transfer.
> Speaking of doing something along those lines, I should think
>that if you had a group of say, 5-10 trucks, an intermediate step
>might be to have each one set up in the manner you described, with
>a server in one of the vehicles, accepting GPS data from the others.
>Then, later on, this data could be used to generate a "blips moving
>on a map" display to show exactly what happened and which truck went
>where. This isn't exactly our ultimate goal of a realtime display with
>'net access, but it would certainly be a fine way to work out the
>bugs and test the system. Taking the server along for the ride also
>conveniently eliminates all of the problems of extending the network
>from the trucks back to home base where the 'net connection is. Once
Yep, that's a good idea. A laptop running a basic database server
like mysql or sqlite would work just fine. I'm wondering if Linux
gpsd will accept data spewing in from multiple locations over the
network. If it did, that'd make the process of dumping the data into
a mysql database extremely simple.
The data collection in the individual trucks could be done with any
GPS unit capable of outputting to a serial cable in standard format
(basically every GPS..) and cheapass old laptops. They don't even
need functioning hard disks... I could get all of the necessary stuff
to fit on a floppy disk or a bootable CD.
This is actually becoming easier and easier as I think about it more.
I might try to test something like this out soon just to see if I can
make it work. Even if gpsd won't accept data streams over the
network, you can just use it on each of the clients and then have a
shell script that dumps that data into the database on the "server."
> Absolutely! :-) Done right, it might even be something
>other groups (offroading, racing, whatever) might be interested
>in, and may even have some market potential were someone so inclined.
>At the very least, it would make a neat thing to have on one's resume. :-)
Damn, now you've convinced me to actually spend some time on this. :-)
If there were actually demand, I could build the entire thing into a
small 12V powered box. Hmm... <light bulb>
> That's probably the best way to go. Taking this even further,
>something like this could achieve a "playback" type system.
>Think of 5-10 trucks, not networked with each other but each with
>an 802.11 equipped laptop connected to a GPS. Each truck goes about
>its merry way, and the application records all the GPS data. The
>next time the truck happens to wander close enough to DML HQ for the
>wireless connection to establish, the application feeds all the data
>its collected to the main server. Since each laptop has their clock
Come to think of it, clock synchronization could be achieved via the
GPS satellite. Darn, this is just getting too cool. Now I've got a
task for those old junker laptops in my garage.
>synchronized, this data could be assembled as you mentioned, and all
>the trucks could be seen moving around at the same time. (Speaking
>of which, doesn't the GPS signal include time data? Could that be
>used either to synchronize the computer's clock, or maybe even better,
>just include a timestamp directly from the GPS with each piece of
>stored data?) A system like this technically wouldn't need each truck
>to be wireless capable. The stored data could simply be copied onto a
>floppy and the main server updated the really old fashioned way. ;-)
Yep (see above.. heh), that would definitely work.
> Another idea I just had related to the whole idea of getting some
>sort of property-wide network-ability, is that the FRS radios we
>use are actually pretty good, range-wise, and have handled things
>like the ravine fairly well, overall. I wonder if some sort of
>"FRS modem" could be created to piggyback the network on those
>radio waves? Perhaps a couple of these FRS modems, in one or two
>trucks could act as a bridge between the wireless network and
>DML HQ?
Seems possible, but I'm sure it's a violation of FCC regs to transmit
data over the FRS/GMRS frequencies. Could also be done with amateur
packet radio.
First step is to work on the GPS data collection. Maybe by the time I
make it to a DML BBQ, we'll have come up with a solution that will
work :-)
-Bill
This archive was generated by hypermail 2b29 : Sun Aug 01 2004 - 00:46:14 EDT