Need some advice in building a simple position system!

Wireless & IoT
1

Hey, guys

I want to build a simple position system for a project for my class. It sounds like it is impossible to build a gps tracker for indoor use, so I want to build a small scale system could be used indoors.

The thing I am building is more like a indoor position system. The way how it works is there are three transceivers and one transmitter.

The transmitter is sending out signals periodically. The transceivers would get the time the transmitter sent the signal and calculated the distance or directly forwarding the time information to 8051 chip.

The microcontroller chip I have to use is intel’s 8051 for the base.

That’s an ugly drawing of the idea. It should position a moving object in a small range like a space of a floor or just a classroom.

http://i44.tinypic.com/wtvssx.png

I feel like other transceiver also need microcontrollers like 8051 or arduino to handle the setup of the transceivers and send correct info of distance to the base.

The transceivers I feel I could use are ADF7902 and MAX7030H. Anybody has an idea of other transceivers or transmit, receiver pairs could be easier to implemented? The infomation I need to transmit to the base would either be calculated distance or the time it takes to receive the signal from transmitter.

Thanks!!!

2

I doubt GPS will give you the position resolution you want in that small a space. Which now begs the question … how accurately must you know the position ? RE: your diagram … how are you determining distance ? Can you tell us what (TT) and (CC) are supposed to represent ? You can attach graphics here so do something in MS Paint if you want to illustrate your concept. Where is the object to be “positioned” ? Are you trying to use RF to measure time delays or IR to measure angles or ??? What’s the underlying working basis for your concept ?

ps - don’t try to replicate a GPS constellation in a classroom. That’s doomed to failure unless you’ve got beaucoup $$s. You might demonstrate the concept of GPS trilateration using acoustics though.

3

king000:
…The way how it works is there are three transceivers and one transmitter.

The transmitter is sending out signals periodically. The transceivers would get the time the transmitter sent the signal and calculated the distance …

I would suggest having 3 transmitters and a single transceiver (vehicle) that calculates it’s position. Also, you may run into issues with having accurate enough clocks for timing over such a short distance. The speed of light is very fast and you will only be measuring the difference in the time it takes for a signal ( traveling the speed of light) to cover a few meters.

You may consider using constant transmissions on close, but different frequencies for each transmitter. Onboard your vehicle, build a filter and amp for the frequency group of the transmitters and judge distance based on signal strength. The strength (amplitude) varies with the inverse of the square of the distance. This is relatively easy to build and the power of the signal will vary enough to measure a change. Once you calibrate each transmitter’s signal strength to a real-world distance, it should be accurate to less than a meter, maybe even significantly better if you have a good ground plane under a good omni antenna.

4

OK, good update on the diagram ! So your concept is to measure the position of the transmitter by measuring the time differences between the time-of-arrival at each of the transceivers ? Think about what that means in terms of how finely you must measure time. Consider that EM waves travel about 1 foot every nanosecond. Go 1 foot closer to one transceiver (and 1 ft away from another) and you’ve got a 2 ns time difference. What resolution do you want in position ? Can you measure time to ns resolution ? This is why I suggested acoustics. Same principle (if that’s what you’re trying to demonstrate) but the speed of sound is soooooo much slower than c. In very rough numbers again about 1 millisecond per foot. Can you measure time to better than ms resolution ? Yes, and easily so !

5

Do not want to dampen your enthusiasm, but radio signal TDOA (time difference of arrival) system are notoriously difficult to create in small areas. And by small I mean “dozens of miles”. Mee_n_Mac only mentioned the first problem you are going to have (timing). The next problem will be to generate a pulse with sharp enough edges that the rise time will be much much less than the time differences. Next you need to design a receiver for the pulse and be able to characterize it well enough to subtract out inherent latencies. And the last problem that is often the death blow to system like this is multipath. A RF signal does not simply go from source to destination. It radiates out in all directions reflecting off countless objects. The result is a cacophony of overlapping pulses at the receiver and little way to know what the right signal is. If you go this way, down conversion of the signal, digitization, and statistical pattern matching with the reference signal is in your future.

Been there, done that. Spent several years on the problem as part of a 10 person team and have little to show for it including the lack of a shirt.

You will still have many of the above problems with an audio or ultrasonic system, but at least with normal test gear (scope, and signal generator) you can track down the problems. In the RF domain, you will need a high quality spectrum analyzer in the US$50K+ and up price range.

6

The March 2012 issue of Circuit Cellar carried an article describing an indoor robot navigation system that uses acoustic triangulation. The design involved four fixed ultrasonic beacons at the corners of a square. The author goes through all the math required for positional location, using integer arithmetic on an eight bit processor. It seems pretty straightforward, if possibly oversimplified.

7

Mee_n_Mac:
OK, good update on the diagram ! So your concept is to measure the position of the transmitter by measuring the time differences between the time-of-arrival at each of the transceivers ? Think about what that means in terms of how finely you must measure time. Consider that EM waves travel about 1 foot every nanosecond. Go 1 foot closer to one transceiver (and 1 ft away from another) and you’ve got a 2 ns time difference. What resolution do you want in position ? Can you measure time to ns resolution ? This is why I suggested acoustics. Same principle (if that’s what you’re trying to demonstrate) but the speed of sound is soooooo much slower than c. In very rough numbers again about 1 millisecond per foot. Can you measure time to better than ms resolution ? Yes, and easily so !

Thanks for all of your advice!

The acoustic idea sounds awesome !

I found one really interesting acoustic sensing device.

http://www.knowles.com/search/prods_pdf/SPM0404UD5.PDF

But I feel like the spec told me nothing about how to connect the device and what the output would be…

Anybody has used a similar one before and know how to use them?

Also, it is hard to find a ultrasonic transmitter, range from 10 khz to 65khz. Anybody happens to know one?

8

No, it isn’t hard to find an ultrasonic transmitter or transducer. Spark Fun carries them.

http://www.youtube.com/watch?v=ueqRp_XYau8

9

Remember when picking a sound generator, you will likely want to be able to generate a single sharp ‘loud’ pulse not a series of pulses. Not all ultrasonic sensors will be capable of a “one ping, one ping only” mode. If you run a series of pulses, it may become impossible to differentiate the first ping at a long distance away to the second ping much closer. Your receivers need to match fairly well the frequency of the transmitter.

In order to process the arrival times, you will either need to time stamp the arrivals to a common clock or simply forward a pulse to one node that will do the computations. That interconnect has to minimize the differences in latency. The absolute amount of latency is not overly important, but they must all be the same. I would suggest copper wire connections. A common clock is a possibility, but keeping them all at the same time can be a pain in the but.

10

king000:
The acoustic idea sounds awesome !

I found one really interesting acoustic sensing device.

http://www.knowles.com/search/prods_pdf/SPM0404UD5.PDF

But I feel like the spec told me nothing about how to connect the device and what the output would be…

Anybody has used a similar one before and know how to use them?

Also, it is hard to find a ultrasonic transmitter, range from 10 khz to 65khz. Anybody happens to know one?

I think you’re putting the cart in front of the horse. Actually I’m wondering if the cart isn’t down the road while the horse is still sleeping in the pasture. Before you start buying parts, why don’t you figure out how it’s going to work, in general and then in specifics and then buy the parts needed that do the specific things they need to do. The first set of questions are what kind of resolution in position are you looking for ? 1 yard, 1 foot, 1 inch ? How often are you going to get a position fix ? Every second, every minute, every ??? Is your “thing” to be positioned moving ? If so how fast ? What’s the environment going to be like ? A open floor or cluttered with chairs and desks and such ? How much are you willing to spend in time and $$s to get this working ? Have you looked for other similar systems to see how they were done and what problems they had. Would you be better off using 4 acoustic transmitters and 1 receiver instead of how you have it above ? Do your homework first and you’ll have a much easier time and higher probability of getting something to actually work.

To give you an example … How much acoustic power will you need to get a good T-o-A measurement at the transceivers ? How far away might the transmitter be ? What does that say about how much acoustic power the speaker at the transmitter will have to output ? Over what kind of angle must this power be spread ? (somewhat obvious per your diagram 180 deg) Is there a part that actually does this or will you have to have multiple speakers ? If so what does this mean for matching one to another ? Will reflections of the floor be a problem ? Are you running this off a battery ? If so for how long ? What does that say about how efficient any amp + speaker must be ? Or how often you’ll be able to determine a position ?

Before you go picking parts, you need to have some understanding of the inter-relationships of the differing parts. A very sensitive microphone (at the transceiver) might mean you can use a low power speaker (at the transmitter). Or visaversa. Depending on other factors (ie - availability of all the parts) you might favor one set of choices over the other. Get a design “working” on paper before you go buying stuff.

11

Mee_n_Mac:

king000:
The acoustic idea sounds awesome !

I found one really interesting acoustic sensing device.

http://www.knowles.com/search/prods_pdf/SPM0404UD5.PDF

But I feel like the spec told me nothing about how to connect the device and what the output would be…

Anybody has used a similar one before and know how to use them?

Also, it is hard to find a ultrasonic transmitter, range from 10 khz to 65khz. Anybody happens to know one?

I think you’re putting the cart in front of the horse. Actually I’m wondering if the cart isn’t down the road while the horse is still sleeping in the pasture. Before you start buying parts, why don’t you figure out how it’s going to work, in general and then in specifics and then buy the parts needed that do the specific things they need to do. The first set of questions are what kind of resolution in position are you looking for ? 1 yard, 1 foot, 1 inch ? How often are you going to get a position fix ? Every second, every minute, every ??? Is your “thing” to be positioned moving ? If so how fast ? What’s the environment going to be like ? A open floor or cluttered with chairs and desks and such ? How much are you willing to spend in time and $$s to get this working ? Have you looked for other similar systems to see how they were done and what problems they had. Would you be better off using 4 acoustic transmitters and 1 receiver instead of how you have it above ? Do your homework first and you’ll have a much easier time and higher probability of getting something to actually work.

To give you an example … How much acoustic power will you need to get a good T-o-A measurement at the transceivers ? How far away might the transmitter be ? What does that say about how much acoustic power the speaker at the transmitter will have to output ? Over what kind of angle must this power be spread ? (somewhat obvious per your diagram 180 deg) Is there a part that actually does this or will you have to have multiple speakers ? If so what does this mean for matching one to another ? Will reflections of the floor be a problem ? Are you running this off a battery ? If so for how long ? What does that say about how efficient any amp + speaker must be ? Or how often you’ll be able to determine a position ?

Before you go picking parts, you need to have some understanding of the inter-relationships of the differing parts. A very sensitive microphone (at the transceiver) might mean you can use a low power speaker (at the transmitter). Or visaversa. Depending on other factors (ie - availability of all the parts) you might favor one set of choices over the other. Get a design “working” on paper before you go buying stuff.

I feel like the scope of the project would be smaller if I use acoustic transmitters and receivers, because if I try to measure the power of signals in order to find distance, the place has to be empty to let the signals pass through not being blocked or reflecting most of the energy. I don’t have an empty room, therefore, the distance from transmitter to receivers would be about 1meter or in that range of circle.

Then the original design becomes easier to implement , since receivers do not need to be located far apart and communicate with each other by sending signals, the receivers could be all connected to the main microcontroller, so I can skip the stage of finding a way to transmit information correctly to the microcontroller.

Also, I found two interesting paper about a similar design, based on its decription, I believe a short distance position system would work. I’ve already bought the ultrasonic microphone in that paper, but I cannot find a ultrasonic transmitter in sparkfun or digikey…the transmitter described in that paper is only sold in Japan… Most of the transmitters I’ve found are currently in China and it is going to take about 2 to 3 weeks to arrive… Although I’ve bought the transmitter sold in China, but it takes tooo long to arrive, does anybody happen to know an ultrasonic transmitter (about 10kz to 65khz ) sold in US?

12

king000:
… the distance from transmitter to receivers would be about 1meter or in that range of circle.

So what resolution are you looking to achieve ?

FWIW I wasn’t advocating you try estimating distance via power, I don’t know how you read that into what I wrote. I was trying to get you to think about what power levels would be needed. I asked that you consider “reversing” the roles to get you to think about the radiation pattern of the transmitter vs that of the receiver. It may be easier to find an onmi-directional microphone than an onmi-directional speaker. A speaker, in the corner of a room, “only” has to have 90 deg beamwidth.

13

Mee_n_Mac:

king000:
… the distance from transmitter to receivers would be about 1meter or in that range of circle.

So what resolution are you looking to achieve ?

FWIW I wasn’t advocating you try estimating distance via power, I don’t know how you read that into what I wrote. I was trying to get you to think about what power levels would be needed. I asked that you consider “reversing” the roles to get you to think about the radiation pattern of the transmitter vs that of the receiver. It may be easier to find an onmi-directional microphone than an onmi-directional speaker. A speaker, in the corner of a room, “only” has to have 90 deg beamwidth.

I’v changed the idea of the project to make it easier and practical to implement thanks to all of your help, especially Mee_n_Mac !

It might be my unclear expression of my last reply. I narrowed down the scope of the project, there is no transceiver in the corners of the room. I am going to build 3 or 4 acoustic receivers on the board with microcontroller and use a seperate transmitter in front of the receivers.

I think the resolution I am looking for is about 10 cm in the range of 1m to 3m and I am going to use a one dimentional speaker in front of the microphones. Based on the information I collected so far, I believe it will work ! Thanks for your help again!