Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Wireless project needs help

Status
Not open for further replies.

Dangerous

New Member
Hi,

I have a pretty ambitious project in mind, and 13 years of Electronics Repair experience to assist. Unfortunately, I have no design experience :(

I want to be able to produce a large number of transmitters, each with a small range (say 5 metres radius), and a number of receivers (up to 100).

The transmitters are fixed at least 10 metres apart, thus not interfering with each other (Yes I know they will anyway, but I hope to randomise their transmission intervals so that the crosstalk will be minimal), and transmit a unique signal. The frequency should be common, though the data transmitted should not.

The receivers are carried around the area, and pass through the areas transmitted in.

So, the receiver should know the ID of the last transmitter it passed by.

On top of this, the receiving station needs to transmit a small amount of data back to a central point that collects all of the information. The range of this transmission is around 100 - 200 metres. The information transmitted will be ID, Location, and Status.

The required end result is a mapped out area, where on a PC screen you know the last location of each node as it moves around the sensor area.

The nodes will move at a maximum rate of 5 metres per second, but only in short bursts of up to 30 seconds.

All I have at this stage is the concept, and the idea that it will involve PIC's and some Holtek https://www.commlinx.com.au/Holtek_chips.htm and tx/rx chips similar to https://www.commlinx.com.au/RLP433A.htm

Is it feasible? I think so.

Any assistance in this regard is most welcome.

PS, I live in Australia, so the frequencies available from the above URL's are my best bet.
 

crust

Member
One of the ways to handle the data collisions is to use the same algorithm that regualar ethernet uses. It is called Carrier sense multiple access with collision detection (CSMA/CD). It is pretty simple, but it requires each transmitter to also have an associated receiver with it. If you have each of the transmitters connected to a centralized controller, they can also "learn" where the receivers are. Basically, when the mobile units transmit their data (or periodically), their ID is transmitted. The stationary units can report to the controller when they see an ID. The central controller knows the stationary locations, so it can make a cluster and the maximum likelihood location of the mobile unit was the center of the cluster.
 

Dangerous

New Member
Thanks crust. I am familiar with CSMA/CD, and considered it, but was wondering for simplicity and cost reasons if it was required.

For further info, i was thinking of placing the stationary nodes inside one of those solar powered garden lights eg:
BRI82240000-1.jpg


Yes, this project is outside, and basically will be used to track people carrying a small transmitter.
 

crust

Member
It certainly isn't required, depending on the transmit frequency of the little moving beacons, you might be able to have several of them in the area without much trouble (presumably, their transmit time is randomized enough). The more often they transmit and the smaller the physical area, the less devices you will be able to sense. On the other hand, if there is a collision, you just throw away that data and wait for the next transmission. As an added measure of robustness, you could also add some error correcting codes to the transmission. One more item is that in many real systems, channel equalization is used to take what looks like a bad signal and extract the digital data from them. I think this is beyond the scope of this project though.
 

bmcculla

New Member
Take a look a bluetooth, its designed to operate in an environment just like you're talking about. Bluetooth transmitters frequency hop. Every ms or so they switch to a different frequency. If two devices try to use the same frequency at the same time they detect the collision and hop to a new frequency. All of this is handled by the bluetooth IC. A simple UART is all that is required to send and receive data through the IC. Class 1 Bluetooth devices have a 100m range. Bluetooth operates in the ISM band(2.5GHz) which is legal almost everywhere.

You can buy modules complete with antenna for ~35$ in quantities of 1. Avnet.com has one made by Taiyo Yuden with built in antenna.
 

McGuinn

New Member
I assume that if you are far into this project, you will already have come up on this problem...

The Holtec controllers are good for what they are, but you will need to find another method of passing an address (or unique identifier) as they only cater for 4 bit addresses (Max 16 'unique' addresses). I am looking to do something which requires multiple transmitters and a single receiver, and what I am considering is, to leave the address section set to a universal standard on all transmitters, and then pass a unique ID from a PIC or similar device.

The other problem, as you have discussed, is using collision detect.
There could be another solution. Instead of adding an additional receiver to the transmitter circuitry to check for a collision, you could employ a different strategy, assume collision...
Let me explain: Firstly, make the assumption that there will be collisions, then develop a method to overcome that. That would be to use multiple sends at random intervals to hopefully overcome the collision. This would be very dependent on the length of the data stream, i.e. if it goes on for 5 seconds lets say, then the risk of collision is too great.
So the idea being that a transmitter sends a stream, waits a random time and then resends the signal, then waits the same random time, and transmits again. This could be repeated quite a few times if you feel the need. The randomness of the time would have to be seeded with a value greater than the maximum transmit time of any unit to allow any other device in the same cell to complete transmission.

Another solution would be to have three transmission frequencies. You could to do two sends, each on a different frequency and randomise the choice of the frequency.

Dunno if any of this helps...
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top