Wireless/RF suggestions for a new product?

Hi,

I stumbled upon this site searching for RF/wireles related informations. I’m building a new product that needs a RF component but I’m unfortunately not very knowledgeable in this area.

Below is a partial spec sheet that describes the requirements in greater details. If you know of specific products that may be a good fit, please let me know! I’d rather hear suggestions from hobbyist than from salesmen!

Thanks,

Max

Operations

· There shall be a central controller (probably a PC - personal computer) located indoors at a fixed location.

· There shall be remote machines outdoors that will roam around the controller in a 1-3km radius (the “field”).

· The field shall be largely unobstructed by trees or buildings. (There will be trees and small buildings nonetheless.)

· There shall be between 20 to 100 remotes.

· Bidirectional communication between the controller and remotes shall be provided.

· Remotes shall report their position (GPS) every minute. It is estimated the ‘position’ message will take 1 kilobyte.

· Remotes shall report some data every hour. It is estimated the ‘data’ message will take 10 kb.

· From time to time, typically once or twice a day, configuration/operating parameters for the remotes shall be changed on the controller.

o The controller shall then send all remotes their new configuration.

o It is estimated the ‘configuration’ message will take 100 kb.

o It shall take no more than 10 minutes to update every remote in the field.

· All data shall be encrypted.

Prototyping/Production

· Two sets of components shall be provided. One for prototyping and one for production.

· During prototyping, components shall be easily integrated together. For instance, no custom PCBs shall be required and connections between components shall be made with off-the-shelf connectors and cables.

o Remotes shall have a full onboard computer with USB and RS-232 interfaces (thus the wireless/RF prototype component shall have USB or RS-232 IO).

o Prototypes shall be easy to debug and shall serve as a test beds for future enhancements.

· Production systems shall be optimized for cost. It follows that the cost of remotes shall be minimized the most.

o Components will be better integrated together (e.g., by using custom PCBs, wireless/RF modules or board instead of stand-alone components, etc.)

o It is acceptable to install any number of boosters and/or repeaters in the field if this results in lower costs.

o Production remotes may not have a full onboard computer (but instead a microcontroller with accompanying logic, etc.)

· The proposed solution shall allow for an easy migration from prototyping to production.

o In particular, software/firmware, drivers, communication protocols, etc. shall work seamlessly on both prototypes and production systems.

zigbee

Zuigbee’s range is typically 30 meters. That’s a little short of 3 kms…

max3000:
Zuigbee’s range is typically 30 meters. That’s a little short of 3 kms…

Have you considered mesh networking? If your remote devices are distributed then it may be a way of reducing the transmitting power at the expense of some extra software complexity.

zigbee IS mesh networking and you can get transceivers that do better than 30 M.

ZigBee modules (not chips) are plentiful from 10 or more manufacturers. Some have 60mW high power and really long range. Of course, meshing (AODV) in ZigBee allows cooperating nodes to relay traffic for long distances or through RF propagation impaired locales.

Hobby use: MaxStream XBee and XBeePro are really easy to use and low cost. Sold by Digikey/Mouser and all.

Philba:
zigbee IS mesh networking and you can get transceivers that do better than 30 M.

Ok ok, I know that zigbee can do mesh networking but it is not an exclusive feature of zigbee.

Mesh networking can be implimented as a software layer on top of a lot of different tx/rx schemes.

I did not know that they did high power zigbee modules, it does seem to defeat some of the point of using zigbee tho.

Meshing in ZigBee adapts the IETF RFC for AODV but adds wireless link quality to the route selection strategy, unlike wireline AODV. I know of no other standards-based meshing today with interoperability among many vendors. 802.11s would like to do this but it’s been in committee for years. There are plenty of sole-source proprietary meshed network products around today. OK if you can take the risk.

high TX power because most ZigBee applications have a transmitter duty cycle of a fraction of a percent. So battery life is a function of the sleeping strategy for the receiver and required latency. The higher Tx power affects cost more than battery life. Example: The MaxStream XBee 1mW is $20@1ea but the XBeePro with 60mW is about $45 as I recall - these are full plug and play modules. The power amp is a 2nd chip.

For those that don’t know, the ZigBee alliance defines an optional network layer on top of the MAC and PHY defined by IEEE 802.15.4. ZigBee is not an IEEE or other formal standard. ZigBee is to 802.15.4 what WiFi is to 802.11.

see this chat thread…

http://www.avrfreaks.net/index.php?name … ic&t=41860

how would zigbee cope with the nodes wandering around? Having followed it a bit in the home automation world many folks seem to have a lot of problems when changing nodes on the network if there are multiple hops.

My impression from glancing at some of the lower level docs is that routing automatically could be supported, but just isn’t handled by the software currently available for the home automation controllers. I don’t know what the state of the support in individual modules is or how difficult it would be to have it discover that it needs to hop and to find a node inbetween itself and the computer. It may be that a non-trivial amount of work would need to be done to keep the nodes in communication if they were randomly moving around.

rod_vdb:

Philba:
zigbee IS mesh networking and you can get transceivers that do better than 30 M.

Ok ok, I know that zigbee can do mesh networking but it is not an exclusive feature of zigbee.

Mesh networking can be implimented as a software layer on top of a lot of different tx/rx schemes.

I did not know that they did high power zigbee modules, it does seem to defeat some of the point of using zigbee tho.

then I suggest you study some more. sorry for being disagreeable but the info is out there.

Of course zigbee doesn’t have an exclusive on mesh. I thought that rather obvious. You should look at tinyos and the various products around that.

transmitter power is completely seperate from the networking stratagy.

JamesS:
how would zigbee cope with the nodes wandering around? Having followed it a bit in the home automation world many folks seem to have a lot of problems when changing nodes on the network if there are multiple hops.

My impression from glancing at some of the lower level docs is that routing automatically could be supported, but just isn’t handled by the software currently available for the home automation controllers. I don’t know what the state of the support in individual modules is or how difficult it would be to have it discover that it needs to hop and to find a node inbetween itself and the computer. It may be that a non-trivial amount of work would need to be done to keep the nodes in communication if they were randomly moving around.

The only shipping meshed stuff in Home Automation that I know of is ZenSys’ Z-Wave, licesensed to a Intermatic and a couple of others. (People on Homeseer.com’s forum talk about it). ZenSys is not ZigBee nor IEEE 802.15.4. I think it’s 1mW and 900MHz.

Home Automation and Industrial controls with ZigBee hasn’t happened yet. Probably late 2006.

The concept in any true self-forming, self-healing mesh (ignoring initial setup), is that when a node cannot communicate with its prevously selected parent or peer node, it does a limited broadcast to find another route. Each node keeps a list of neighbors (candidate routes) and a list of active peer-associations for use in forward peers’ traffic.

So when a node moves, it should follow these and many other rules to find a new association. Not unlike what happens in WiFi where there are many access points as in a large building with ubiquituous coverage. But in WiFi, every link is at most one hop. Not so in ZigBee. Today’s products limit the hop count to about 5.

I’ve found that ZigBee 1.0 which uses AODV for the routing strategy works quite well at reforming when a nodem moves, or more importantly, when a node drops off the network due to power off. So long as there’s sufficient transmitter power, the routing can re-form to offset the loss of the node which had been routing.

stevech:
The only shipping meshed stuff in Home Automation that I know of is ZenSys’ Z-Wave, licesensed to a Intermatic and a couple of others. (People on Homeseer.com’s forum talk about it). ZenSys is not ZigBee nor IEEE 802.15.4. I think it’s 1mW and 900MHz.

OK, that clears up a lot for me. I was at the Electronic House Expo this past year and I chatted with at least a dozen different companies working on “Z-Wave” stuff and they were all as confused as I was

Home Automation and Industrial controls with ZigBee hasn’t happened yet. Probably late 2006.

Yup, thats what I would have to guess too. Given the size of some of the companies involved and their glacial pace towards actually releasing any product it may be even longer…

I’ve found that ZigBee 1.0 which uses AODV for the routing strategy works quite well at reforming when a nodem moves, or more importantly, when a node drops off the network due to power off. So long as there’s sufficient transmitter power, the routing can re-form to offset the loss of the node which had been routing.

Excellent! So the ZigBee stuff actually works, thats very good to know. I may have to rethink some projects… Thanks!

We have had a lot of success with OLSR in wireless mesh environments. There is a lot of documentation and a lot of source code around to assist you.

you guys realize that motion in 2 dimensions really doesn’t mean a whole lot to a radio transmission so long as you stay in range, bluetooth or your own 2.4 GHz networking would be your best solution, or even WiFi. just make sure your wanderers are designed with radial geometry in mind when it comes to your signal, so you’re in phase with your ‘hub’ or main system, otherwise you might drop connections a lot when moving.

You also might want to use UDP with a robust layer 5 protocol. As much as we all like to pretend TCP is robust, you will find it breaks frequently in lossy, intermitent environments.

The more robust your UDP-based protocol is, the less throughput you will see on your system, but since we are only talking a couple kB, it shouldnt be a problem.

If you are savvy enough with mesh routing protocol design, you could even build mesh routing into your application.