indoor localisation for minute positional changes of moving objects

[dr dre]

Hi, I've been thinking about a project in which computer controlled toy cars races together. The thing is that, I need to track down the exact location of each cars to determine next action should be performed for each of them. except one hero car, others will be simulated by the computer itself. The part I'm very confused at is that of localisation. I would like to implement something similar to GPS but it must detect small changes in location. There are different kinds of devices that I'll be using for communication like wifi, or bluetooth. So, what you guys think about it? Is there any specific devices available for this purpose? can you share your ideas? This is very important project to me. please help.

 

[davenn]

this query comes up from time to time

wifi, BT, GPS are not going to work

look into multiple cameras and object recognition systems

 

[dr dre]

why? I don't know much, but still I believe there's a way. Here's my idea.. the sensors keeps communicating with a server and cars keeps sending some data to those sensors.. can we use signal delay or something? may be we could place an accelerometer or gyro for more perfection. lets break it down..

this query comes up from time to time

wifi, BT, GSP are not going to work

look into multiple cameras and object recognition systems

 

[Alec_t]

If you did have sensors placed as shown, how far apart would they be? Metres? Kilometres? Bearing in mind that radio signals travel at around 30cm per nS, any method involving time delays would need extremely accurate timing of very short delays.

 

[BobK]

I started thinking about this last time someone wanted to do it. (It is a very common request). Here is a possible solution I came up with. You would use both IR and untrasound. The IR pulses would provide a clock signal that is common to all the cars and the beacons. Because the speed of light is so fast, the propagation delays would be neglible.

There would be 3 ultrasound beacons. Each sends a pulse at a specific delay from the clock pulse. The delays here would be longer than the speed of sound delay across the entire area of interest, so that they would never overlap. Then the cars would measure the delays for each of the 3 beacons from the last IR pulse, subtract off the expected delay, and could then triangulate the position from the different delays. Basically it is the same as a GPS system, but using ultrasound and a synchronization pulse rather than RF and atomic clocks.

Not a simple system to implement, but I think it could be done, even by an amateur like myself.

Bob

 

[davenn]

If you did have sensors placed as shown, how far apart would they be? Metres? Kilometres? Bearing in mind that radio signals travel at around 30cm per nS, any method involving time delays would need extremely accurate timing of very short delays.

He said indoor, so distances would be measured in metres
and yes as you say the the measurements would be tiny, not likely to be able to be measured with equip built by a home constructor

why? I don't know much, but still I believe there's a way.

if you have to ask why, it demonstrates that you don't understand the difficulty of the problem and the reasons why radio signals
couldn't be used as has been commented on by Alec

.... may be we could place an accelerometer or gyro for more perfection.

they wont help with location of the cars


Dave

 

[AnalogKid]

Think about football. Televised professional American football. The system that superimposes the line of scrimmage and the down line relies on three rotary encoders per stationary camera. The encoders send back signals related to rotation, tilt, and zoom lens setting, and a gigantic buttload of high school geometry does the rest. In dre's case there is no elevation component and no zoom lens to change the field of view, so a triangle of rotating narrow-beam scanners could locate anything in the covered space with nothing on the target object other than a reflecting surface. Because the controlling system knows the distances among the scanners, all that's needed to calculate a location are the three angles.

ak