Positioning System

Discussion in 'The Projects Forum' started by tornado151, Aug 21, 2010.

  1. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    Hi,
    I want to develop a system which can very accurately track the position of someone or an object on a one floor building.

    The purpose of this is because I want to make a 3D simulation of what is going on in real life, and then develop a game which uses this information.

    I have come to the conclusion that a time based positioning system like GPS may work.

    I would like to tell you my plans and would appreciate it if you could tell me if you think it is feasible and how accurate you think it would be.

    Ok the plan is I have my PC as a time server, a number of RF transmitters located around the building which connect to the computer via an RS232 interface. Software on the computer sends out the current computer time to all RS232 devices which transmit the time.

    A number of RF receiver units are placed inside a box, this is connected to a small PC which it's time has been synchronized with the transmitter PC. It does a calculation to locate the device based on the time difference between each transmitter.

    Do you think this could work? Do you think it would be accurate? How many transmitters do you think I would need?

    Thanks
     
  2. Papabravo

    Expert

    Feb 24, 2006
    10,179
    1,800
    The first problem is that the PC is rapidly losing the RS-232 Interface. The second problem is multi-channel multi-path interference. What does that mean?

    1. The multi-channel part means that either you have to have all transmitters on the same channel and find a way to make sure that two transmitters don't transmit at the same time, or you have a separate channel for each transmitter meaning that you need a receiver for each transmitter. On top of that closer transmitters will desensitize a receiver that is not even tuned to the correct frequency so that it becomes deaf.
    2. Multi-path means that the receiver need to sort out two copies of the identical signal that arrive via different paths and have a phase difference.
    I would be surprised if you could get such a system to work.
     
  3. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    I would use an RF module connected to an RS232 to USB circuit.

    That was the plan.
     
  4. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
  5. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    Ok, I have just been thinking, how well would the reverse setup work?

    The person being tracked has one transmitter on them, transmitting a time which is synchronized with a server.

    Four recivers are placed around the area being tracked, the position of the person is calculated based on the time it took for the signal to reach receiver.

    Do you think it would work well?
     
  6. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    Yes. That would work. It is know as TOF (Time Of Flight). If you have all of the receivers receiving the signal you can measure the distance to each by measuring the time the signal takes to reach each receiver. Thats the easy bit. Now you have to actually do all the timings, calculations and decodings. Microcontrollers would be the easiest in my view. You will need a common timing signal so they are all synced together.

    Thats about the limit of my abilities.

    PS. Consider ultrasonic transducers as well. You may have mullti path problems but in a straight line they are much easier to time.
     
  7. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    What advantages would there be of using ultrasonic transmitter/receivers over an RF transmitter/receiver?

    Thanks
     
  8. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    Ok just been thinking further.....

    A server is connected by a wire to four receivers and one transmitter (Being located).

    [​IMG]

    The server sends a signal to the transmitter which sends a very short pulse, as each receiver receives the pulse it notifies the server which then records the time.

    It then does the maths to locate the transmitter.

    Do you think this would work well?
     
  9. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    What degree of precision do you need? Your timebase is going to have to be very precise for the idea to work - http://en.wikipedia.org/wiki/Light-second

    Correlating the time of reception for all four receivers is going to be interesting.
     
  10. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    Personally, I can't see you pulling it off. Not only would you require a precise reference common to all, your phasing would have to be locked amoungst them. Slowing things way down into the audio spectrum might be a more managable project.
     
    tornado151 likes this.
  11. tornado151

    Thread Starter Member

    Apr 15, 2010
    34
    0
    Good idea.

    I could do a test with audio and move to RF if successful.

    So I can get four microphones and a speaker, easy to connect up to my PC.

    I will run some test and see how it goes, I will need to buy for microphones of the same spec first though.
     
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You might try ultrasonic transducers, but your range will be limited. If you get into the audible spectrum, you'll drive everyone bonkers.

    3D spatial orientation using four IR transmitters and receivers has been done since the 70's, but the equipment was quite complex, had limited range, and quite expensive.

    Trying to do something like that for a floor plan of an area greater than perhaps 20'x20' would be quite challenging.
     
  13. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    If all the subjects would cooperate it would be easier to outfit them with transmitters then place receivers above the ceiling.

    I was friends with the CEO or a corporation down in San Antonio that, during a growth phase, ended up occupying various areas spread across a larger complex. If they needed to locate someone they just clciked his name ontier PC and it showed exactly where he was.

    This was of course before mobile phones were smaller than a breadbox. :)
     
  14. BillB3857

    Senior Member

    Feb 28, 2009
    2,400
    348
    If my memory serves me, a Radio Frequency signal travels 1 mile in 6.18 MICROSECONDS. Converting that to feet results in extremely small time increments. I know processors are fast, but that fast??
     
  15. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    Time didn't enter into it, just location determined by the strength of the signal at the nearby transmitters.
     
  16. BillB3857

    Senior Member

    Feb 28, 2009
    2,400
    348
    Wouldn't the position of the human body with respect to the antenna have an effect on signal strength? Standing in one spot and turning around would look like moving across and back in the room.
     
  17. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    You got me, all I know is that it worked for them. Precision wasn't important, just knowing what phone they were near was. It may not have resolved down any closer than to what room they were in.
     
  18. windoze killa

    AAC Fanatic!

    Feb 23, 2006
    605
    24
    As far as range of ultrasonics a previous job I had I designed a system for positioning a chute over a hopper. It had an operational range of about 15M. Worked very well. It used 4 ultrasonic transducers in the corners on an RS485 bus and the moving object was also on the bus. We also considered lasers but the dusty environment made them usless and also the wavelength was a problem.
     
  19. tkane

    New Member

    Aug 26, 2010
    1
    0
    Why not just use GPS? GPS is time sync'd and you can transmit LAT/LONG once every second. No need to calculate time between transmission and your recvr. Sync your server to GPS 1pps and your're good to go.
     
Loading...