A Challenge: RF propagation between cargo containers

How far RF signal (2.4 GHz, 802.15.4 standard) can travel between containers? Average cargo container size: Width 8 feet (2.5 M), Height 8.5 feet (2.6 M), Length 20 feet (6 M) or 40 feet (12 M). Containers may be stacked up to 6 levels vertically, up to 12 rows wide and 15 rows long. The horizontal gap between containers is about 2 inches (5 cm), same distance minimum between the walls (vertical gap). A good picture of cargo ship is here: **broken link removed**
Low profile (dipole) antenna will be placed on the top of container, close (about 20 cm) to the corner, about 2 cm from container surface. What minimum transmitter power we may need to support container-to-container communication, if our “smart” containers may be 5-8 containers away of each other in any direction? These containers should be able to create a network, to relay their ID’s to the control point, or to “main” container, which becomes a (low speed) communication hub for all “smart” containers.
I don’t have RF simulation tools, so any ideas what can be used (frequency, power, polarization etc.), will be greatly appreciated.

Sounds to me like you want to build Cantenna systems

https://www.google.com/search?sourceid=navclient&ie=UTF-8&rls=RNWE,RNWE:2004-40,RNWE:en&q=Cantenna


Robert A. Patterson
https://quantumgravitics.tripod.com

This is a real problem because the parallel metal walls are closer than a wavelength at most operating frequencies and also you have no control over where the container with the sophisticated electronics is placed in the three dimensional stacking.

The other problem is that you will be amongst the containers from several owners so that if your containers have "windows" in the metal with plastic covers they still may not be able to communicate if they are on the inside of the stacking.

The best, but expensive, solution is to have all of your containers with the 'windows' and complex electronics. One of them will be able to communicate to the satellite and from that you will be able to tell generally where all of the containers are. Once they are offloaded from the ship each will report its location and when they are on motor vehicles they will not be stacked and they can continue to transmit their locations.

Your system will prevent accidentally warehousing a container which will be forgotten. Sophisticated thieves will be able to defeat your system by offloading the containers at night when the solar cells will have no power and they can put cloth or other items over the cells to prevent your system working during the day.

About solar panel issue, there will be a rechargeable battery, supporting at least 30 days of independent operation (no sun exposure). This should be good enough for one shipment cycle. Also, we can’t have a RF-transparent “windows” in container, we just place our antennas on the top (epoxy-filled “brick”), and electronic box inside container.
What if we will try to use 5.8 GHz, or will go down to low frequency (HF or even lower), where container itself may act as an antenna… Probably a stupid idea, since stacked vertically containers make a good electrical connection to each other.

This is going to be expensive. All of the bricks will have to be fully functional and be able to communicate to the satellite and to the other bricks. The one on the top of the stack will know it when the solar cells charge the battery. It is on during all hours of daylight. The other bricks send out, at a very low bit rate, their serial number on 5.8 GHz every hour or so. All units have 1 ppm accuracy reference oscillators. The brick that is on all of the time will receive the serial numbers and report them through the satellite.

As I said, this is a very expensive system.

What is this system for? Do you need to define the possition of every container in everey period of time during shipment or what?
If you want to do RF comunication between containers, which are separated with another containers, I don't thing it is possible. The containers between yours are just like the electric wall and thay will reflect every RF signal.

The container metal walls are the major problem of your RF transmission. It will just change your antenna characteristics and draw you in a madness of poor performance....

More is, if it is to located outdoors, the weather will also a critical point of the system.

The problem is pretty trivial, but you must take a system perspective to see it. Sure there are many metal containers spaced close together. So you will have "parallel plate" type waveguides formed all over the place. These will propagate the 2.4 Ghz just fine (because the waveguide "width" is bigger than λ/2), but there will standing waves. The solution is to have one lowpower transciever in each cargo container, but at least two antennas in a diversity arrangement spaced apart on the container sidewall. If the data rate is low, like I would guess it is, you do not need any fancy diversity software, just switch in one antenna for a few minutes, and then switch to the other antenna for the next few minutes. Eventually all messages will be shared among the network.

As far as contacting an orbiting satellite, you are going to need something like 5-20 watts or so of power to do so. That is a costly part, and takes a lot of battery juice to run, so you want to NOT put one of those into every container. I would put one or two uplinks onboard the ship that the thousands of containers would share to talk to the satellite. You can have some uplinks on each dock that the containers are offloaded onto. If you really want coverage in transit, you can have one uplink on each truck or a few on each train, one at each end recipients loading dock, etc.

I do not think that you lose any security by having the shared uplink, as a smart theif would simply take a crowbar to the transmitter antenna on a container wall if he wanted to steal it before moving it.

I imagine you could devise such a system so that the battery life was around a 1/2 to 1 year with 4 "D" cells, depending on the data length per container.

You can get a glancing result by 10*n*log(d)(dB) power loss at a distance of d. n is 2~5 depend on the radio path(from LOS to diffraction).

But it has a big problem to match the antenna. The adjacent container is in the near field region (<1 wavelength) of the antenna, and will effect the impedance of the antenna remarkably. Any change in position will lead to remarkably change in impedance too.

Thanks everybody for deep analysis. Yes we are trying to track (position and various sensors alert) EACH container ANYWHERE on Earth, on land (truck, train, yard) and in the sea. So far nobody could do this with guaranteed coverage and stable back link. There is a huge demand in this feature, considering security issues. I am sure this post will have lots of views. Some companies are trying to use Iridium and Orbcomm (LEO) satellites for uplink; we are using Inmarsat D+ with good coverage (no blind spots) and low equipment cost. One of the companies rushed to post their sea trial results claiming they achieved a container mesh network using 802.15.4 standard (pending for release ZigBee, 2.4 GHz, 0-3 dBm?) architecture:
http://www.raesystems.com/whitepaper
If you look at the PDF presentation on this page, all “traceable” containers are stored in one location (not a real life), they have multiple “sensors” inside each container, plus “network access point” next to their containers, and then “Satellite Communicator”. Quite messy and depends on many factors. Not an elegant solution, if usable today at all.
Eventually (in 10-15 years) all containers will become “smart” and will handle the network issues, but we need to protect the world now. Technology is in the middle of research. It is an outstanding subject for brainstorming; please continue generating ANY related ideas.


Added after 22 minutes:

We are trying to put a unified set of equipment on each container, capable of independent and in-network operation any time. Most likely we will end up placing at least one stand-alone unit on some location of the cargo ship permanently, to use it as a guaranteed satellite communication hub. 30 W solar panel will support a maintenance-free long time operation. Weather (salt, moisture, dust) will affect antennas and solar panel effectiveness, but hopefully not so badly.

You do not really need the local area network. If the containers are traced to the harbor and then some of the signals disappear and yet the container on the top of the stack shows that the ship is going where it is claimed, you can deduce that all of the containers are on the ship and going to the proper destination. Once they are offloaded at the destination, all of them can communicate to the satllite and indicate that they are traveling on a road or railroad by your knowing the locations.

This system can still be defeated by knowledgeable thieves who place a metal housing over your bricks which keep them from communicating to the satellite.

Another option for tracking on land would be to have mobile telephone units that telephone messages to you.

You got it, Flatulent. We already duplicated the satellite uplink by GSM/GPRS quad band transceiver for on-land tracking. It works fine on stand-alone (or on-top of stack) container. The units will have the “sleep mode” for battery saving. Alarm or timer will awake the processor. How often it will report the position/condition, the geo-fence (not planned route changes), and switching on/off different sensors, all this we would like to be able to program remotely via back-to-container communication. The data packets are short, and transmitted in a 5-10 sec low-rate burst. If the container skips the scheduled report several times (CTC network or via satellite), it should raise the alarm. THAT is the level of security we need. Then the ground control center will send a request to the ship/land authorities, to inspect the “silenced” container. If container looses all communication for any reason, it switches to the log-only mode, with periodic attempts to establish communication. As a luxury service, you can even request a satellite image (if sky is clear) of the truck or ship with your container via internet. The resolution is high enough to see an individual container or even a human body next to it.
Yes the project is ambitious and challenging, a lot of testing to do. Yet Mr. Murphy with his Laws is always around…

If we want to put 2 antennas to fight dead zones, can we operate 2 antennas simultaneously (share of power, unbalanced impedance)? If loss of power is an issue and we still want to switch, then how we can switch 2 antennas without mechanical relay (battery life etc.). Are there available solid state UHF (2.4 GHz) switches?

If we can switch 2 antennas, then next question is about a network algorithm. How to pick the best antenna, should it then stay in the “best RX” chosen antenna, or it need to be switched periodically all the time (“good RX” antenna for container 1 may not mean the same for container 2. And we are not talking about container 3 yet).
What if “neighbor” containers keep switching antennas with the same rate and by accident there is never enough time overlap for the best combination of antennas between them to exchange with data? Looks too complex to me.

The diversity receiving antennas are commonly used on 802.11 devices. The receiver tries all of the available receiving antennas during the start of another stations's transmission and stops trying when it finds an antenna that has sufficient signal.

Your other idea is commonly called 'mesh' in which all stations relay between each other.

Rokor, just one idea.

Think about placing antenna not at the center of wall, but in some up corner - then you will have good radiation in 3 planes instead of one. Or on the rib - 2 planes.

The antenna module in the form of a "brick" filled with RF-transparent epoxy is about 20cm X 15cm X 2cm size, and placed on the top front corner of container, about 20cm from the front and 50cm from the side. So position is pretty much in the corner, and this is most protected zone from the mechanical damage during container manipulation and transportation. We don’t want to separate antennas (CTC, GPS, Satellite) physically in different modules. Fewer modules = easier installation and more damage protection. We can not change this position, also because solar panel should be located right next to it and cover all cables and cable hole in container.
By the way, antenna "brick" and solar panel should withstand the direct metal (crane or other container) hit, and still be operable. Solar panel is flexible, mounted by self-adhesive tape and may just loose some effectiveness if damaged.

So we have 2.4 GHz dipole for container-to-container Mesh network, and power limit is 100 mW (European regulations). We still are not sure about network protocol (algorithm) to use. Potential solutions could be a ZigBee, or RF232 from AeroComm. But they all need a designated Server, it is not flexible as we wanted originally. This means we may need to put a stand-alone unit on the ship (main communicator) to assign ALL containers as clients only.