RF Distance detection between two objects.
I'm doing a project at university and need a way of detecting when an object leaves a rough perimeter.
The project is roughly akin to a house arrest anklet but it needs to be smaller - to fit around the user's wrist. It must detect when the wearer is more than x (say 10 foot) from a certain point and then flash an LED on their wrist to alert them to this.
I was thinking of solving this problem by having a small RF transmitter constantly broadcasting from the fixed point in the centre of the perimeter and having a tuned circuit built into the bracelet periodically listening for the broadcasts.
The problem is that I have little knowledge of working with RF circuits and the transmitter and received would need to be very compact (a couple of cm^2 on each end) - would it be possible for me to detect over this distance in this footprint?
Could anyone point me at resources for potential antenna / circuit designs / any other useful resources. I am trying to do this as cheaply as possible which is why I was thinking tuned circuit instead of a prefab tx/rx IC.
I realise this post is a little vague but I am really just asking for confirmation that this is a plausable approach, and if so them for a little direction in developing a solution.
Thanks in advance, all comments are appreciated!
Pete
Are you sure to have chosen the right project?
However, before bothering with RF circuit details, you should analyze the problem in terms of wave propagation and antenna properties. It sounds like you want to determine distance based on the received RF level. Unfortunately, the received level is affected by many additional parameters like transmitter and receiver antenna orientation, interference due to wave reflections and multipath reception, RF absorbing objects. As a result it's impossible to determine distance from received level with some accuracy. Unless you considerably relax the requirements, there's little chance to implement the project this way.
Things to consider: an array of multiple receivers, directional antennas, RF pulse time-of-flight. Unfortunately none of them leading to a simple "circuit".
Thanks for the reply FvM.
The project is not an optional one unfortunately, otherwise something digital would suit me better :)
I understand the issues with determining distance based on the signal strength, however, the actual distance between the objects is not of that much concern to me, if the radius before the alert is made from the fixed point is between 10 and 30 foot then that would be fine either way, does that relax the requirements enough do you think?
(What I'm trying to say is I'm not that interested in exactly how far away it is before the bracelet is triggered so long as it's more than 10 foot)
Do you have links to any sources regarding antenna design and wave propogation?
Thanks again!
Pete
FvM gives you wise advice. Signal strength is highly dependant on the signal path. For example, suppose the signal source is only a few feet away but the person wearing the bracelet reaches inside a refridgerator. The metal body would almost completely screen the signal even though the field strength around it is still strong.
I'm not sure of the circumstances this would be used in but a possibility might be to surround the perimeter (inclusion zone) with a loop antenna and use a low powered transmitter in the bracelet. It still wouldn't be foolproof but it would make it more difficult to create 'dead spots' at any point within the loop. You could also consider RFID technology, especially if there might be more than one bracelet within the detection range.
Brian.
Years back a senior project group at my university did something like this to make an automatic cornhole scoring system. The platform had a loop antenna wound around its perimeter, and the bags had small loop antennas in them to detect the strong signal emitted by the platform. So long as the bags weren't sideways, or leaning up against the side of the platform, it seemed to work pretty well. Using an array of loop coils in the platform would likely give better results. I forget what frequency they operated at, but it was definitely low enough that they could ignore wave propagation effects, which makes things a lot easier.
If you can assume the object has a fixed orientation, then that makes things much easier. Anything that reduces the degrees of freedom in the system, really.
Thanks for the replies folks :)
That's sounds like an interesting project mtwieg. Unfortunately this is not an option for me as I have no means of setting up a perimeter.
I understand that if the wrist device were placed inside a fridge or some other kind of faraday cage then it would block the signal, this doesnt concern me too much, there may be metal present but generally it is safe to assume that the transmission medium will just be though the air.
Also, I can make no assumptions about the orientation of the TX or the RX. I completely get that there are a lot of unknowns here, antenna choice, frequency choice, the orientation of the antennas and the transmission power.
I've been doing some reading around the subject and it seems to me that as I have no idea about the orientation of the two antennas I would need to use one with the lowest directivity - this would mean using a short dipole antenna to get the maximum spread of signal and using a relatively high frequency signal such that I can make the signal travel farther and the antenna short enough?
Again, I realise there are a lot of unknowns but I was hoping that I could overcome the directivity issues and possible issues relating to mediums other than air just by bumping up the TX power - this would mean that the distance from the TX to the point where the bracelet would 'leave the perimeter' would be highly variable. (again, not really an issue).
Is there a website or a book that anyone could recommend that would teach me the basics of trace antenna design and basic RX / TX design? I have a feeling I am just going to have to get some components and have a play with all of this, should be fun! :)
Again, thanks for all the comments!
Pete
My thought process on this is that, if it can be done for those tags that you can connect to your keys so that if you lose them, you can press a button and the tag will beep, then it should be possible for me to do something similar?
Pete.
THere are several methods used for Wireless perimeter guards.
Central RF transmitters rely on the Loss of Signal (LOS) and for dogs, they have a 15 to 30 m range.
The transmitter is central on a residence main level raised above the floor. Obviously buildings with metal cladding wont work.
Buried perimeter wire works on low frequency with the detection of the signal near near perimeter for effective detection but requires buried wire to become wireless but works for long perimeters more accurately.
Narrow band filtering and coding permit better false detection and misses for the RFID.
Friis Losses increase inversely with distance and proportional to frequency, while reflections can interfere with signal levels at higher frequency and antenna efficiency drops at lower frequency if too short. Thus the long buried wire works well at low RF frequency and the central transponder works better in the UHF and up band.
Here are some concepts for such tracking, and references you can look up:
http://www.technologyreview.com/news...llow-shoppers/
https://en.wikipedia.org/wiki/Indoor_positioning_system
After thinking about this, I decided the best method "might be" to use something near the AM band at low levels such as not to cause interference with FCC regulations.
It could be a 1kHz pilot tone or addressable data signal with AM or FM modulated on say a 1.5MHz carrier at 50mW. The Tx antenna would be a vertical wire centrally located while the Rx antenna would be the human body so as to receive the most signal coupled into a loop coil around the wrist band to an AM radio cct. with AGC but detection based on adequate signal to noise ratio of the pilot tone carrier with a PLL chip and lock detector. Using the body as an antenna would give the vertical polarization of the signals better omnidirectional response , better than a loop or ferrite bar AM antenna which have nulls in directivity.
The Tx power depends on the RX sensitivity , so depending on which AM design you choose, the results will vary. Distance will not be precise if there are sources of signal absorption or reflection or attenuation nearby but at least the wavelength is much longer than the path length. Coupling to a telephone cord for example may improve sensitivity further away as the phone cord may pick up the signals. Power consumption can be very low for AM receivers but a beeper or buzzer or electric pulse to the wrist may be more easily detected than an LED.
Just thinking aloud :- The instantaneous field strength tells us little, but as the wearer will only be moving at a relatively slow pace, if the field strength "dies" in 2 mS, then its value can be ignored as its the "fridge" effect. So if there is an electronic comparator that compares the expected field (based on the wearer moving at +- 5 MPH) with the actual field then spurious readings can be ignored. Another technique could be to alternatively shift the transmitted frequency and compare the to received levels as again one would expect them both to decay at the same rate. Any sudden changes must be due to a reflection.
Frank
Just use a RF trans-meter which transmit a particular digital string and receiver will receive that string. they will work only for eche other. and when trans meter reaches out of range then you can detect it.
search HT12E and HT12D. parallel to serial encoder and decoder ICs. send 12 bit string from transmitter where 8 bits used for address by which your detector can work for a particular transmitter.
Good thinking SunnySkyguy with the human body being used as an antenna! I would like the device to work if it was not on the users wrist as well though, for instance if it was placed in their pocket - so maybe its a bit of a non starter! Nice idea though!
I also like chuckeys point about the field strength decay rate as well, I could infer a lot from it!
So I've been having a look around at antennas and general approaches to solving this problem and think I've found a possible solution for the receiver at least - If I use an AM carrier wave with say a modulated 1kHz sine wave and the following circuit:
My thoughts are that I can easily use the discrete components to strip away the carrier wave, and the negative half of the modulated signal, and then use a series RC circuit to require a certain voltage across a capacitor to be present which can then be used to trigger a digital output from an OP amp.
This would give me all that I need really in terms of functionality with minimal power consumption and a very small footprint. Thoughts?
Sorry for the hand sketched circuit - I am away from my desktop PC!
Cheers,
Pete
In principle that will work but you will face problems with it. The primary one is the level of voltage you will pick up. To keep the range and potential for interfering outside low, you need to use the least transmitter power as possible, prefereably no more than about 5mW. Remember your receiver may be far less sensitive than others in the locality.
The low transmit power and need for a portable antenna will mean you will only pick up a few uV at the diode, the recovered voltage will therefore be extremely low and the comparator threshold will be barely above ground. It would make more sense to use a simple TRF receiver circuit to boost the received RF level up to at least a few hundred mV and also forget the AM, just use the RF level itself. I visualize a single transistor oscillator as the signal source and maybe a two transistor receiver feeding the comparator directly.
Brian.
some random thoughts:
if you transmit all the time, your battery power will quickly be used up. it is much easier to receive all/most of the time than to transmit all the time.
there are a couple ways to do this problem:
a) transmit from a known point, receive this on the bracelet and retransmit it, receive it back at the known point and compute the round trip transit time.
b) transmit from the bracelet in the blind. receive it at multiple known points, and triangulate where the transmitting bracelet is.
c) make an in-house "GPS" system where the bracelet receives multiple signals from known points, and calculates where it is from time of arrival. you would probably want to use a GPS chipset, or some portion thereof
d) use some sort of distance/angle of arrival receiver system to figure out where the bracelet is. For instance, the bracelet transmits for 10 milliseconds every 10 seconds. You receive it with an array of antennas and figure out the azimuth to it. then use some method, even just signal strength to estimate distance. IF you have TWO known points, you can triangulate where it is too.
e) use one of the new Bluetooth chips that have a time of flight calculator engine in it.
f) just make a "doggie fence" out of the house. run a line around the outside perimeter of the house. when the bracelet crosses that line, the bracelet receives a low frequency (maybe 100 KHz) signal and flips out with an alarm.
g) use existing wifi networks in a house to somehow figure out where you are. Maybe there is someway to transmit to the wifi some data, and using pings or data arrival times and some software running off your router, you can figure out how far away the user is (maybe the easiest since you do not need to make the hardware...just program existing WiFi chips/cards)
h) use an impulse transmitted, and a equivalent time sampler in a receiver at the known point to figure out round trip distance. the bracelet is just a bent pipe retransmitter. you would have to separate known returns from the room with moving returns from the bracelet wearer.
I) there are probably a ton more ideas....put on your thinking cap.
Thanks Brian - very helpful!
My reasoning for having the AM signal modulated onto it was that if I do not modulate something on to it and search for that, then any signal at the same frequency would be enough to trick the device into thinking it was still within the perimeter - by modulating the signal I can be sure that only a carrier of X Hz with a modulated signal of X would cause a reset - unless I am missing a trick?
OK, I hadn't considered just how low the signal level was going to be - I guess a simple pair arrangement would be enough to boost the signal to a high enough level, am I right in thinking that this is what you're envisioning for the receiver:
As I say, I am happy with this as a design concept barring the fact that anyone transmitting at the same frequency will cause the device to malfunction.
Cheers,
Pete
The circuit has to be a little more complicated than that!
You are quite right about 'customising' the signal so it can be picked out from the background noise but to do that you need an extra stage at the transmitter and receiver. It does to some extent reduce your reliance on signal strength at the receiver but at the expense of more complicated circuitry. Basically, you need to modulate the transmitter, AM will do and you can do it with an extra transistor acting as a low frequency 'ID' oscillator. At the receiving end you still have to pick up the RF and demodulate it but then you need a selective filter to accept only the 'ID' tone and reject everything else. There are several approaches to that, in circuit terms the simplest is probably to use a PLL tone detector, the more reliable method is an active bandpass filter.
There is another similar method which is more technically complicated but cheaper to build, that is to send data instead of a tone as modulation. It means a microcontroller at each end but it can also serve to flash the LEDs so some of the other components are no longer needed. Some programming expertise would be needed. It has another advantage, it would be possible to use several systems in close proximity, even if their perimeters overlapped. You can't do that with a signal strength related detector.
Brian.
Proximity effect of the hand or body to antenna coupling is very effective for improving signal strength, so a wrist loop may not be necessary and just capacitive couple to the AM receiver.
The important thing to remember is the body will be a more reliable antenna to act as an isotropic antenna, which is critical for field strength LOS detection.
It will be better than any loop, dipole , path , stripline or whatever. which all have nulls but can be used as weak antenna , capacitively coupled to the body. ( gee I bet someone has a patent on this)
You're absolutely right, don't know quite where my head was at drawing that!
I think that for my needs it will be sufficient to amplify the signal using the mosfet pair, strip off the carrier with an envelope filter arrangement as in the circuit before and then pass the remaining signal through a low pass filter - then, assuming that my transmitter 'ID' frequency is very low (maybe 10Hz) it should be sufficient to stop any other spurious signals from causing a problem.
My thinking is that what are the chances of there being another signal at (for example) 900MHz with amplitude modulation at below 100Hz (the filter cut off)? Fairly low I would imagine - it might happen occasionally but I doubt it would be the norm.
If my assumption is true then it would mean that the component count could remain as low as possible while still functioning almost all of the time.
What are your thoughts about the assumption?
You've been very helpful Brian, thank you! Sure I will be posting sketchy hand drawn pictures of my transmitter attempts in the days to come!
Cheers,
Pete
Somehow I feel that RF is not the way to go on this project. Unless there are well-defined entry/ exit points from your "space", in which case simple RFiD type tags can be used - like in shops !
Following from this, the most important i think would be to understand & characterize the "space" where such a system would be used.
For e.g.:
1) Are there many obstacles within this space ? If any ? Like furniture, warehouse racks, walls ?
2) Does more than one 'object' need to be monitored simultaneously ?
3) What type of 'object' are we talking about ? Is it a T-shirt ? A vase ? A human ? A parakeet ? Your description implies a human person only.
4) How 'rough' can your parameter be ? Is it 10ft +/- 10% ? 50% ? 100ft +/- 20ft ?
The 10 foot stated implies a medium sized room. Or is it clear open space ?
5) ........
And lastly - what course is this ? The course title would provide some clues on the best approach to take. If it happens to be something like TA102/ An Introduction to Design, then just a description of how you analysed the problem & evaluated various options is probably what is required. You see what I'm saying ?
cheers!
It would be easier if you could do it in reverse, carry the transmitter and have a fixed receiver but that would negate the possibility of flashing LEDs on the wrist band. I wouldn't use 900MHz, if you use a frequency between 150MHz and 200MHz it would make the body itself work as an antenna. Those frequencies equate to a body height of between 1.5m and 2m. I would also suggest that if a fixed tone modulation was used, a higher frequency would be better. My reasoning for that is the background electrical field will peak at around 50Hz - 60Hz which would get through a low-pass filter and also it would be harder to design a narrow pass band filter at low frequency, especially using small components. If you go too high, say above 25KHz you run into problems with interference from CFL lighting and switch mode power supplies, somewhere between, maybe 10KHz would be a good compromise.
You could consider a super-regenerative receiver, they are simple, reliable and have a low current consumption. They are used extensively in things like pagers where battery life has to be long. They work best in the VHF range of frequencies and are very sensitive for their complexity. Their biggest drawback is being extremely noisy but if you are going to pass your modulation through a narrow filter that wouldn't be a problem.
Brian.