Sensor Network

I hope all of you are doing well.
I have come across a challenging project and I need your advice.

Its a security application with the following scenario:

Doors with computerized locks (biometric/swipecard/pincode), the locks
access a central server for access authentication. The catch is that
the communication between the lock (embedded computer/PCB etc) and the
server has to be wireless and the solution need to be extremely low
priced, you can't have a very expensive lock, as this is for a very
large deployment (that why wifi won't work).

Any suggestions regarding the design of the embedded computer in the
lock and the communication method between the lock and the server
would be really helpful.

I appologize is this the wrong group for this query, a pointer to the
correct group would be hihgly appreciated.

Thank you, and waiting eagerly for your response.

Regards,
A H

Adnan wrote:
> I hope all of you are doing well.
> I have come across a challenging project and I need your advice.
> 
> Its a security application with the following scenario:
> 
> Doors with computerized locks (biometric/swipecard/pincode), the locks
> access a central server for access authentication. The catch is that
> the communication between the lock (embedded computer/PCB etc) and the
> server has to be wireless and the solution need to be extremely low
> priced, you can't have a very expensive lock, as this is for a very
> large deployment (that why wifi won't work).
> 
> Any suggestions regarding the design of the embedded computer in the
> lock and the communication method between the lock and the server
> would be really helpful.

Without knowing A)distance B)target price per door
and C)level of security desired, it would be absurd
to comment.

"Adnan" <ahrizvee@yahoo.com> wrote in message
news:adcf748d.0403261113.385852b1@posting.google.com...
> I hope all of you are doing well.
> I have come across a challenging project and I need your advice.
>
> Its a security application with the following scenario:
>
> Doors with computerized locks (biometric/swipecard/pincode), the locks
> access a central server for access authentication. The catch is that
> the communication between the lock (embedded computer/PCB etc) and the
> server has to be wireless and the solution need to be extremely low
> priced, you can't have a very expensive lock, as this is for a very
> large deployment (that why wifi won't work).

Are all locks battery-powered? If not, what's the point of wireless when you
have to feed the locks anyway?

Meindert

Hi Adnan!

If the locks are battery powered, I would prefer TIs MSP430 micro. I 
work on a sensor network project, too and love the MSP! Current 
consumption ~300�A/MHz or only a few �A in different power down modes.

For the network you could consider a multi-hop-technology and 
communicate from door to door (your hops). This should work, if you are 
in big buildings with many doors on each floor. You could use a wireless 
link, there are cheap and smart devices available:

http://www.chipcon.com
http://www.rfm.com
http://www.nvlsi.no

Regards,

	Andreas

> Any suggestions regarding the design of the embedded computer in the
> lock and the communication method between the lock and the server
> would be really helpful.

As others have said, you haven't given us enough information for
really meaningful suggestions. But, do you really mean "wireless", or
merely that you can't run ADDITIONAL wires for signals? Power-line
modulation protocols can use the wires that you already had to run to
power this thing.

If it really has to be radio, then there are about a hundred "popular"
protocols that might or might not be appropriate depending on a dozen
factors you haven't disclosed to us; pick one, ZigBee, Wireless USB,
Bluetooth, miscellaneous FSK modem protocols operated in the ISM
bands, etc etc etc etc etc.

"Andreas" <andi@andreas-bellgardt.com> wrote in message
news:c427uk$9rl$06$1@news.t-online.com...
> Hi Adnan!
>
> If the locks are battery powered, I would prefer TIs MSP430 micro. I
> work on a sensor network project, too and love the MSP! Current
> consumption ~300&#4294967295;A/MHz or only a few &#4294967295;A in different power down modes.

You forgot the necessary current to drive the electromagnetic lock.....
Apart from that, in many countries it is mandatory (fire regulations) to
have electromagnetic locks that are locked while powered. So in case the
power fails, the locks open. This security measure is hardly achievable with
battery operated locks.

Meindert

I think i can answer the "Why Wireless?" question. To get all the
cabling thru the floors and ceiling is too invasive and costs a lot of
money.

As for your wireless solution, I'm still waiting for ZIGBEE to get
released. It's 802.14.5 (or is that .15.4). It's low power, and low
data rates.

Otherwise, the only solutions that look feasible now would be to go
use a WLAN. You'll probably be able to get away with the security that
802.11b/g offers, but if not, then you'll need SSL (??). See, it opens
a whole can of worms.

Then the next question would be how to operate your wireless card from
your embedded system.

I'm looking into the same thing for security access (but not for
biometric) stuff. There is a PIC solution for this wifi card at
www.iosoft.com
Or you can hold off and wait for www.netburner.com 's wlan solution. 
Or you can get an 802.11b embedded solution for Linux on a x186.
If 802.11b is overkill, then consider 900MHz transmitters. I think
they have a far range.

So many OR's in my explanation.
-Mike


"Meindert Sprang" <mhsprang@NOcustomSPAMware.nl> wrote in message news:<40648b72$1@news.nb.nu>...
> "Adnan" <ahrizvee@yahoo.com> wrote in message
> news:adcf748d.0403261113.385852b1@posting.google.com...
> > I hope all of you are doing well.
> > I have come across a challenging project and I need your advice.
> >
> > Its a security application with the following scenario:
> >
> > Doors with computerized locks (biometric/swipecard/pincode), the locks
> > access a central server for access authentication. The catch is that
> > the communication between the lock (embedded computer/PCB etc) and the
> > server has to be wireless and the solution need to be extremely low
> > priced, you can't have a very expensive lock, as this is for a very
> > large deployment (that why wifi won't work).
> 
> Are all locks battery-powered? If not, what's the point of wireless when you
> have to feed the locks anyway?
> 
> Meindert

Oops, in my earlier response, i missed that you said you can't use
Wifi. Looks like your're stuck with a 900MHz solution then. You can
also buy wifi extenders.

Zigbee supposedly can be like a wireless mesh, with each packet
wirelessly hopping to the next device til it gets to the destination.
It's just too bad that these companies in the "Zigbee alliance"
haven't delivered. All these standards meetings and they drag their
feet. This is where you can look into the startup companies that
already have their own proprietary stack. Costs $6.50 to put into your
hardware (but the development system costs more than $10K!!)

-Mike



ahrizvee@yahoo.com (Adnan) wrote in message news:<adcf748d.0403261113.385852b1@posting.google.com>...
> I hope all of you are doing well.
> I have come across a challenging project and I need your advice.
> 
> Its a security application with the following scenario:
> 
> Doors with computerized locks (biometric/swipecard/pincode), the locks
> access a central server for access authentication. The catch is that
> the communication between the lock (embedded computer/PCB etc) and the
> server has to be wireless and the solution need to be extremely low
> priced, you can't have a very expensive lock, as this is for a very
> large deployment (that why wifi won't work).
> 
> Any suggestions regarding the design of the embedded computer in the
> lock and the communication method between the lock and the server
> would be really helpful.
> 
> I appologize is this the wrong group for this query, a pointer to the
> correct group would be hihgly appreciated.
> 
> Thank you, and waiting eagerly for your response.
> 
> Regards,
> A H

Lewin,
The past few weeks, Zigbee has been a buzzword that's caught my
attention lately. In fact, quite a few of the electronics parts
distributors and their FAEs are now visiting potential customers. It
seems they are just looking for guinea pigs and to see who is
interested, and are targeting companies who sell units by the
thousands. The company i'm contracting at is interested. It seems that
Layers 1 and 2 are pretty much cast in stone, but i haven't been able
to get straight answers regarding how the darn thing is programmed.

Like i said earlier, the Zigbee alliance doesn't seem to have a mature
stack nor routing protocols, but if you look at some their individual
members who are startups, they do have their own proprietary stacks
which do cost a bit of money though. (Maybe i'm wrong, perhaps Zigbee
is mature, but it's not public news.) I'm waiting for the Atmels and
Motorolas to start selling their transceivers and microcontrollers
that have zigbee stuff integrated, but such press releases are several
months old and overdue.

There are quite a few skeptics out there that think Zigbee won't make
it, wireless USB won't make it, just like Bluetooth is having problems
getting adopted by the masses. Who knows, maybe it's the past two
years of the tech recession that has frozen innovation (relative to
the way it was 3-4 years ago). I do know that the ESP magazines have
gotten very thin the past several months.

larwe@larwe.com (Lewin A.R.W. Edwards) wrote in message news:<608b6569.0403261410.46df135@posting.google.com>...
> > Any suggestions regarding the design of the embedded computer in the
> > lock and the communication method between the lock and the server
> > would be really helpful.
> 
> As others have said, you haven't given us enough information for
> really meaningful suggestions. But, do you really mean "wireless", or
> merely that you can't run ADDITIONAL wires for signals? Power-line
> modulation protocols can use the wires that you already had to run to
> power this thing.
> 
> If it really has to be radio, then there are about a hundred "popular"
> protocols that might or might not be appropriate depending on a dozen
> factors you haven't disclosed to us; pick one, ZigBee, Wireless USB,
> Bluetooth, miscellaneous FSK modem protocols operated in the ISM
> bands, etc etc etc etc etc.

Thank you for your responses. Following are some more details:

Price: Under US$100 / lock, that inclues the lock, the biometric
scanner, radio transciever, embeded system. (I know, its tough)

Power Requirements: Not an issue. This is for a hotel environment,
where this lock would be installed in every door. The power to operate
the electronic lock would be taken from within the room. The reason we
can't go for a "wired" comm. link between the server and the lock is
that doing the wiring for 500 locks is just too invasive, expensive,
disruptive and is prone to higher maintenance cost.

Distance: The distance between the server and the lock (embeded system
if you may), will be approx. 200-300 metres (may be across floors).

No. of Devices: 500. With such a higher no. of devices, interference
would be a concern. Even narrowband RF transcievers might pose a
problem. Please note that these devices are spread across different
floors, so could we use a master node/hub on every floor, and use
devices with a smaller range (only floor wide) to overcome the
interfence issue.

Multihop technology sounds interesting, would it do the job?

Thanks again for your responses.

Best Regards,
A H

"Meindert Sprang" <mhsprang@NOcustomSPAMware.nl> wrote in message news:<40648b72$1@news.nb.nu>...
> "Adnan" <ahrizvee@yahoo.com> wrote in message
> news:adcf748d.0403261113.385852b1@posting.google.com...
> > I hope all of you are doing well.
> > I have come across a challenging project and I need your advice.
> >
> > Its a security application with the following scenario:
> >
> > Doors with computerized locks (biometric/swipecard/pincode), the locks
> > access a central server for access authentication. The catch is that
> > the communication between the lock (embedded computer/PCB etc) and the
> > server has to be wireless and the solution need to be extremely low
> > priced, you can't have a very expensive lock, as this is for a very
> > large deployment (that why wifi won't work).
> 
> Are all locks battery-powered? If not, what's the point of wireless when you
> have to feed the locks anyway?
> 
> Meindert