Wireless Comm: Establishing Distant Tracking Between 2 Objects

[EEstudentNAU]

Hello, I am very new to this and don't know where to start. I am interested in establishing communication between two devices. The primary goal will be to frequently track the distance between these two objects, and if the distance takes on a certain value, execute some action.

There is more to it, but that is the basics of it. Let's say a max distance of 20-30m. The devices are portable. Communication will be in open space (essentially no obstructions). Higher the frequency, the smaller the antenna?

Any help is appreciated.

 

[chroot]

What's your budget? More specifically, describe your needs in each of the following categories:

1) How easy must it be to build and test?
2) How much can it cost?
3) How big or heavy may it be?

- Warren

 

[berkeman]

Hello, I am very new to this and don't know where to start. I am interested in establishing communication between two devices. The primary goal will be to frequently track the distance between these two objects, and if the distance takes on a certain value, execute some action.

There is more to it, but that is the basics of it. Let's say a max distance of 20-30m. The devices are portable. Communication will be in open space (essentially no obstructions). Higher the frequency, the smaller the antenna?

Any help is appreciated.

As chroot implies, many more details would be of help. You could just run a spring-loaded string between the objects and monitor the play in and out, eh? Why not use a string? Why does it have to be wireless?

How about laser rangefinding? Can you track it from some base station via robotic video tracking, and hit a cube corner on the target with a laser range finder? If your accuracy requirements are not too tight, just put GPS on the base station and the rover.

More details please...

 

[arithmetix]

Let us know more! On the basis of the problem you've described I could recommend a piece of string tied to a switch! What's it for? (string plus switch won't work reliably for a roaming dog)

 

[EEstudentNAU]

1) How easy must it be to build and test?
Must be able to be built and tested by a college student with access to labs, etc.

2) How much can it cost?
Cost is not an issue.

3) How big or heavy may it be?
Must fit in someones pocket or on their belt.

One unit ("control unit") will remain somewhat stationary, and will define the range the second unit ("roaming unit") can travel (some circle of some radius around unit 1). It has to be wireless because the second unit may travel a considerable distance out from the first unit (30 meters).

When the roaming unit exits this radius, it will alert the control unit and an alarm located on the roaming unit. Upon reentry to the zone both alarm and alert will stop. The control unit will check the distance of the roaming unit say once or twice a second. Is this something I need a microcontroller for?

 

[arithmetix]

There are dog-control devices on the market, at least there were 20 years ago, which alerted the owner when the dog went out the gate. From vague and distant memory I think that these worked on a signal/no signal basis... when the dog got too far away there was no longer a received signal and the alarm went off. Perhaps you could use one of those?

 

[skeptic2]

1) How easy must it be to build and test?
Must be able to be built and tested by a college student with access to labs, etc.

2) How much can it cost?
Cost is not an issue.

3) How big or heavy may it be?
Must fit in someones pocket or on their belt.

One unit ("control unit") will remain somewhat stationary, and will define the range the second unit ("roaming unit") can travel (some circle of some radius around unit 1). It has to be wireless because the second unit may travel a considerable distance out from the first unit (30 meters).

When the roaming unit exits this radius, it will alert the control unit and an alarm located on the roaming unit. Upon reentry to the zone both alarm and alert will stop. The control unit will check the distance of the roaming unit say once or twice a second. Is this something I need a microcontroller for?

If the control unit had a small low power transmitter and the remote unit had a receiver that could measure its signal strength, you could compare the signal strength level in a comparator in which you could set the threshold at which the alert is triggered. If your transmitter frequency were below 490 kHz, you're allowed a decent signal strength and you'd be operating in the near field of the transmitter which means the signal strength drops faster than the inverse square law and gives you a more accurate distance measurement. If you operate between 160 and 190 kHz you're allowed up to 1 watt of power which is far more than you need.

Furthermore operating at these frequencies it is easier to build transmitters and receivers. The antennas would be ferrite rods and you can even use higher speed opamps in the receivers. I would have the control unit transmit continuously and the remote unit would report back to the control unit once or twice a second whether it is inside or outside the radius, transmitting back to the control unit on a higher frequency. You don't need a microcontroller, a timer IC would be sufficient.

 

[f95toli]

If the control unit had a small low power transmitter and the remote unit had a receiver that could measure its signal strength, you could compare the signal strength level in a comparator in which you could set the threshold at which the alert is triggered.

But that signal strength will depend on many things, not only the distance. I don't think it is possible to come up with a simple scheme for measuring distances that only use the intensity of a single radio signal and one receiver. The only way to do it would be to triangulate, although which might actually be feasible if the distance is only 30m, the two receivers could placed fairly close to each other.

 

[DaveC426913]

I was going to say the same thing as f95toli. Signal strength will be a very unreliable measure of distance. You would get a high ratio of false alarms.

 

[vk6kro]

If you had the control transmitting on frequency A and the roving module receiving on this frequency, the module could then receive a pulse and retransmit it on frequency B which the control is receiving.

At 30 M distance, the time for a round trip (60 M ) would be 0.2 microseconds.

So, the control sends a pulse and starts an onboard counter counting 100 MHz clock pulses.
When the pulse gets back from its round trip, it stops the counter.
These pulses are 10 nS or 0.01 uS apart.
If the count gets to 20 or above, the roving module is too far away and this rings alarm bells.

Counters that could count at 100 MHz are readily available.

 

[berkeman]

If you had the control transmitting on frequency A and the roving module receiving on this frequency, the module could then receive a pulse and retransmit it on frequency B which the control is receiving.

At 30 M distance, the time for a round trip (60 M ) would be 0.2 microseconds.

So, the control sends a pulse and starts an onboard counter counting 100 MHz clock pulses.
When the pulse gets back from its round trip, it stops the counter.
These pulses are 10 nS or 0.01 uS apart.
If the count gets to 20 or above, the roving module is too far away and this rings alarm bells.

Counters that could count at 100 MHz are readily available.

Seems like the repeater delay would be too variable to make this accurate. There might be a way to make it more constant, but especially with varying signal strengths, the AGC delay alone would be pretty variable.

Triangulation or GPS seem like the best bets...

 

[skeptic2]

But that signal strength will depend on many things, not only the distance. I don't think it is possible to come up with a simple scheme for measuring distances that only use the intensity of a single radio signal and one receiver. The only way to do it would be to triangulate, although which might actually be feasible if the distance is only 30m, the two receivers could placed fairly close to each other.

I was going to say the same thing as f95toli. Signal strength will be a very unreliable measure of distance. You would get a high ratio of false alarms.

I worked many years designing and testing on location units that worked on this principal for use on overhead cranes in steel mills, auto plants, etc. While what you say about the signal strength may be true at higher frequencies, at these frequencies the wavelength is much longer than the distance being measured. Multipath or other effects we're accustomed to worrying about that affect signal strength have virtually no effect at these frequencies. The near field effect also enhances the accuracy because the signal strength drops off more rapidly with distance than it does at higher frequencies. Once calibrated, the method produces remarkably repeatable results. [Note: The buildings in which these units were installed were made of steel and filled with machinery made of steel, yet we still achieved very reliable operation.]

The most important factor in maintaining accuracy is keeping both both transmit and receive antennas vertical.

 

[vk6kro]

Seems like the repeater delay would be too variable to make this accurate. There might be a way to make it more constant, but especially with varying signal strengths, the AGC delay alone would be pretty variable.

Triangulation or GPS seem like the best bets...The receiver could be just an RF amplifier followed by a mixer to the other frequency. No need for AGC and no appreciable delay across a small circuit board compared with the delay of an RF signal traveling 30 metres.
This is an open area and money is no object, so signal levels could be very large. You would have a receiver that does not receive band noise.
Accuracy would be about 1.5 metres with a 100 MHz clock. (30 M / 20 counts =1.5 M per count)

GPS would be a nightmare because you would have to keep calculating which coordinates were within the designated circle and GPS is pretty inaccurate for precision work like this.

 

[berkeman]

The receiver could be just an RF amplifier followed by a mixer to the other frequency. No need for AGC and no appreciable delay across a small circuit board compared with the delay of an RF signal traveling 30 metres.
This is an open area and money is no object, so signal levels could be very large. You would have a receiver that does not receive band noise.
Accuracy would be about 1.5 metres with a 100 MHz clock. (30 M / 20 counts =1.5 M per count)

GPS would be a nightmare because you would have to keep calculating which coordinates were within the designated circle and GPS is pretty inaccurate for precision work like this.

Agreed about the GPS accuracy. Interesting idea about the straight-thru shot. So basically it would just be a squelch-qualified repeat, with the input signal strength enabling the output amplifier. Might work.

As for the power and frequency, are there any open bands in the range of where skeptic2 was talking about, or are they all licensed?

The open ISM/microwave oven band definitely has multipath and body absorption issues for amplitude...

 

[skeptic2]

As for the power and frequency, are there any open bands in the range of where skeptic2 was talking about, or are they all licensed?

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.3.241.7&idno=47

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.3.241.13&idno=47

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.1.241.12&idno=47

 

[chroot]

Forgive me for this silly question, but why do this yourself when nearly any modern cell phone can do it already?

- Warren

 

[vk6kro]

Forgive me for this silly question, but why do this yourself when nearly any modern cell phone can do it already?

- Warren

How does a cell phone measure distance?

 

[f95toli]

I guess he is referring to A-GPS.

 

[chroot]

It's very easy to write a program which will execute some action when two internet-enabled smart phones are within some distance of each other. There are even services which allow friends to share their location data with each other.

- Warren

 

[arithmetix]

No-one has suggested sound. Would that be accurate enough?

 

[vk6kro]

Sound would be very accurate. At 30 M you would be dealing with a delay of 100 mS which is huge. You could use that to measure distance down to a few milimeters.

The problem would be that the roving vehicle could be in any direction, so the vehicles need to have omnidirectional speakers and omnidirectional microphones on them and still be able to detect the other one at 30 M.

The distance is just a bit out of range for low powered gear but may be possible with big amplifiers.

Radio has the opposite problem in that it is a bit too fast, so we are dealing with very short delay times.

 

[arithmetix]

Well, tv remotes use ultrasonics at about 40 kHz. The transceiver pairs can be got (used to be anyway) through radio spares. I think you could send coded ultasonics from one end and return the signal by radio from the other. Then your arithmetic would be very simple and so would the cctry compared to some other designs.

I mean, more specifically, coded pulses, every 10,? 20? 100? milliseconds

 

[EEstudentNAU]

If you had the control transmitting on frequency A and the roving module receiving on this frequency, the module could then receive a pulse and retransmit it on frequency B which the control is receiving.

At 30 M distance, the time for a round trip (60 M ) would be 0.2 microseconds.

So, the control sends a pulse and starts an onboard counter counting 100 MHz clock pulses.
When the pulse gets back from its round trip, it stops the counter.
These pulses are 10 nS or 0.01 uS apart.
If the count gets to 20 or above, the roving module is too far away and this rings alarm bells.

Counters that could count at 100 MHz are readily available.

I am interested in this method. Could you sort of give me a brief overview of a system that could do this? What sort of hardware am I looking at for this?

GPS would be a nightmare because you would have to keep calculating which coordinates were within the designated circle and GPS is pretty inaccurate for precision work like this.

Just how inaccurate are we talking? Less than a meter, more than a meter?

 

[berkeman]

Just how inaccurate are we talking? Less than a meter, more than a meter?

More than a meter, IIRC. It has to do with military-versus-civilian use of GPS. So the other methods being offered and discussed will probably be better solutions if the associated issues can be sorted out.

 

[vk6kro]

GPS would be a nightmare because you would have to keep calculating which coordinates were within the designated circle and GPS is pretty inaccurate for precision work like this.

Just how inaccurate are we talking? Less than a meter, more than a meter?

Probably 10 metres; sometimes more than that.
I use a GPS in a car and also a handheld one. They both give an estimated circle of inaccuracy and it is alarming. It fluctuates with satellites, too. As more satellites become available, the accuracy improves, but it is never better than about 5 metres.
It is even obvious in the car when the GPS voice is telling me to turn in 50 metres when I am already at the street I need to turn at. At other times, it will be giving this instruction when there really is 50 metres before the next street.

Cell phones would have the same problem.

Now, hardware.
First step is to ask what country you are in? If you are in the USA there is only one band for radio control and that is about 72 MHz.
This limits the two band approach. Each receiver and transmitter have to be on all the time and they can't be on the same frequency or a nearby frequency or the receiver wouldn't be able to hear anything except its own nearby transmitter.
I would prefer a frequency in the 30 to 100 MHz range so that we could get reasonable efficiency from the antenna.
So, what country are you in and do you know if you have access to transmit low power in the above frequency range?
Some countries (including mine) have radio control bands on 29, 36 and 40 MHz or similar frequencies. These would be suitable for dual band operation, subject to official approval, if needed.

 

[berkeman]

They could use the ISM band around 2.4GHz world-wide, but there's a reason it's called the "microwave oven band" or the "WiFi Band"...

 

[vk6kro]

Yes, I'm avoiding that one unless existing wifi gear could be used.
It is a bit too high a frequency for homebrew techniques.

Universities here have allocated bands for this sort of thing and it isn't hard to get permission to use them.

Better GPS is available (WAAS ?) but it requires a local reference from a stationary signal. Again, that's getting too difficult.

 

[EEstudentNAU]

I am in the US, and am currently unsure of what the regulations are. Skeptic2 has posted the following links:

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.3.241.7&idno=47

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.3.241.13&idno=47

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=ff542c745811f0fae53018a9c122f2df&rgn=div8&view=text&node=47:1.0.1.1.14.1.241.12&idno=47

 

[EEstudentNAU]

Would Zigbee be a viable choice for this? What about transmitting in the Wi-Fi range? 2390 MHz to 2483 MHz

http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr;sid=2452557b2d93f16c042e7628f028500d;rgn=div8;view=text;node=47%3A1.0.1.1.14.3.241.5;idno=47;cc=ecfr

 

[vk6kro]

I have to wonder if we are going overboard a bit.

We are trying to locate a roving module which will be in plain sight all the time by spending hundreds (possibly thousands) of dollars on electronics.
Unless it is on Mars there must be a more realistic way of doing it.

What about marking a 30 M radius circle (with a piece of string if you like) and get some clear flexible plastic tubing and lay it out around the circumference of the circle?
Use tent pegs to hold it in position.

It would take 189 metres of plastic tubing to do this.
Then you put a plug in one end of the tube and a pressure sensor at the other end and ring a bell if something runs over the pipe.

What is the actual application?

 

[EEstudentNAU]

I have to wonder if we are going overboard a bit.

We are trying to locate a roving module which will be in plain sight all the time by spending hundreds (possibly thousands) of dollars on electronics.
Unless it is on Mars there must be a more realistic way of doing it.

What about marking a 30 M radius circle (with a piece of string if you like) and get some clear flexible plastic tubing and lay it out around the circumference of the circle?
Use tent pegs to hold it in position.

It would take 189 metres of plastic tubing to do this.
Then you put a plug in one end of the tube and a pressure sensor at the other end and ring a bell if something runs over the pipe.

What is the actual application?

http://www.linxtechnologies.com/

The have some cheap TX/RX modules that can output the RSSI. Can this be compared to a reference voltage and if its lower, then it is outside the allowable range and sets off an alert? Because information is being exchanged, will I need a TX and RX on both devices, or are there devices that can do both (transmit and receive)?

The application is for me to learn something about wireless communication. Eventually I want to be able to have a robot follow me or another robot using simple RF communication.

 

[skeptic2]

Those modules have frequency ranges from 260 - 470 MHz with wavelengths from about 1.15 to 0.64 meters. At these frequencies the signal may bounce off the ground and arrive either in phase or out of phase with the direct signal, thus the RSSI would indicate a higher or lower reading than would be expected from distance alone. In addition a signal at these frequencies can be affected by the proximity of a human body.

In contrast, by using frequencies in the LF band (below 490 kHz), the wavelengths, instead of being about 1 meter, would be closer to 1 km. At 30 meters you would see no multipath nor will the proximity of a human body have much effect on the signal strength. In addition 30 meters will be within the near field region in which the signal level falls off more rapidly than the inverse square law giving you more repeatable results than at higher frequencies.

It should be fairly simple to build transmitters and receivers in the LF band. The received signal strength can be compared to an adjustable reference voltage that you would adjust by going 30 meters away and setting it so that it just barely turns on. If you don't put a small amount of hysteresis in the comparator, it may "chatter" a little right at 30 meters. You also need to keep good conductors (metals) away from the ferrite antennas as they will tend to detune them.

Lastly these frequencies are covered in part 15 of the FCC regulations which means you don't need a license to operate as long as you don't use excessive power.

 

[EEstudentNAU]

Are LF transmitters and receivers readily available or as you said is this something I will have to build myself? Size is sort of an issue (smaller the better).

 

[skeptic2]

Are LF transmitters and receivers readily available or as you said is this something I will have to build myself? Size is sort of an issue (smaller the better).

I know of none that are commercially available so you can have the fun of building it yourself. For the transmitter you'll want an oscillator, perhaps tunable, and a buffer amplifier driving a tuned ferrite rod.

The receiver will take the signal from the ferrite rod and amplify it, rectify it and compare it to your set point. The easy way is to use an IF integrated circuit with an RSSI output such as the MC3372. If you need more sensitivity you can always add a low noise, high frequency opamp at the input.

 

[berkeman]

The receiver will take the signal from the ferrite rod and amplify it, rectify it and compare it to your set point. The easy way is to use an IF integrated circuit with an RSSI output such as the MC3372. If you need more sensitivity you can always add a low noise, high frequency opamp at the input.

And what else needs to be part of the ferrite rod antenna and first gain stage circuit...?