SparkFun Forums

[lwclasses]

Hi all,

Imagine a 30 cm box, with 20 bugs in it, all running around. I need to get the position (distance from a static tracker is fine) of each bug at about a rate of 60 times every second. The bugs could have something attached to them (~1-2 cm?).

Requirements:
60 updates per second - although could drop down to 30
Accuracy is pretty important here - sub 1cm.
small transmitter/tracker
must be able to track multiple 'objects' at once
sometimes bugs are on top of each other so cannot be line-of-sight
cheap - the cheaper the better (as usual )
Oh and preferable not needing a battery, or possibly a small one (or could us wireless power??? since such a small area?)

Would probably work with triangulation if needed

Have looked at:
BT - too slow and too inacurate
BLE - bluetooth low energy has same shortfalls of BT
WIFI - too inacurate?
RFID - too inacurate also?
NFC - Near Field Communiction - too slow
UWB - Ultra Wide Band - ?

So any suggestions?

Not too much to ask is it?

Jono

[Mee_n_Mac]

Good luck !

The only thing I can think of that has any chance of coming close to your req's is optical. Use a camera array. If a bug is on top of another one, too bad. Use a tracking algorithm on each detection. If one "hides" for a while, you'll know where it last was. If it "unhides", it'll be picked up again.

[jremington]

Agreed, video is your only option. You will need a fairly high quality camera connected by a fast connection to a PC so that video frames can be stored on disk. People often use Matlab programs for the image processing and frame to frame object tracking. For example, bacteria swimming on a microscope slide can be tracked, see http://www.rowland.harvard.edu/labs/bac ... /index.php

[stevech]

yes, video is best option but complex if automated rather than eyeballs on screens.
There is a technique using a pulse-coded bright infrared emitting diode on the tracked object. Sensed by 3+ cameras. Then complex IR code (ID#) detection feeding into hairy math with photogrammetry to estimate location using precisely located cameras with well known lens focal length, etc.

No RF technique can get what you want.
Other than UWB with large equipment on the mobile side, heavy infrastructure, maybe 5cm. Oh, and $20K.
Time Domain or Ubisense.

Maybe electronic mileposts.. very short range (under table/floor) IR receiver or transmitter to know "I just passed milepost x and I know x's location". That requires a cooperating infrastructure. But not 2cm, not 60Hz.

This is hard.

[lwclasses]

Oh stink! Pretty tall order I know. Hey thanks so much for your insights guys.

I just a thought now and would love your far more experienced thoughts...

What about a small device (ultrasonic speaker) that gives different frequencies and could be triangulated?

would that be hard to implement, parts don't exist, be really expensive, triangulation isn't that precise or doesn't update that quickly?

Any of your insights appreciated?

[jremington]

Using your method of choice, first determine how to locate one object in a similar setup. If google can't help, that is a strong hint.

[lwclasses]

Thanks to everyone that replied. I really appreciate your thoughts. I think that will be enough to keep me busy for a while!

Thanks again,

Merry Christmas

[Valen]

Why do people suggest (edit: consider) methods of datacommunication as a way of distance /position sensing? I really cannot fathom that.

Is it perhaps because the world got smaller as the telephone came into existance and internet blossomed? (Intended rethorical, but feel free to reply

[jremington]

Good question! Many beginners seem to think you can do just about everything with radio waves.

[waltr]

Yea, its all FM (Freaking Magic) that can do anything like in all the really BAD SF movies.

[Mee_n_Mac]

Radar works w/RF.
GPS works w/RF.
Hogan's Heros could DF via RF.

It's *logical* all RF can yield position ... just as your microwave does.

[stevech]

Radar works w/RF.
GPS works w/RF.
Hogan's Heros could DF via RF.

Classic radar gives only range (not position) by RF means. Azimuth is by mechanical aiming of the narrowed beam RF.

GPS works because there are many transmitters received in quick succession, and the receiver can get the precise 3D location of each transmitter, then use range to each to do the trig (a bit simplified).

And, GPS receivers are cheap because the transmitters aren't.

[macrob]

Possibly a powered rfid or non, powered stronger signal longer range quicker maybe. with a analog output sensing device to pic up on the rfid transmission or non powered rfid. Analog might be capable to cover more tracking area then digital that has to be pointed in the general direction to detect. I don't know might be looking into I was told for tracking... :shifty: