Non-camera static object tracking

Hi all, first post.
I’m curious if anyone can think of a way to keep track of all reasonably sized human/life/robotic movement in a room. I thought of cameras, motion detectors, and infrared cameras already.

  1. Cameras: too much coding and I would assume out of the scope of an arduino project to track movement of a large object with a camera
  2. motion detectors: as far as i know, motion detectors only detect motion, they don’t know exactly which part of the viewing area it came from.
  3. infrared cameras: too expensive and again, most likely too much coding and out of the scope of an arduino project.

WITH THAT SAID, any ideas on how to basically make a grid system of where any foreign object is? if you had allot of time and money you could make a bunch of vertical laser beams from the ceiling to the floor and be able to make a grid of which lasers connected to a light receptor on the ground and which didn’t, hence knowing where any object in the room is at all times, from a top down viewpoint (all i need is top down), except thats not actually a practical idea. basically i need radar lol. and TERCOM.

Life like would suggest you need to scan for 98.6 degree temperatures using infrared sensors
Life size suggests you need to have motion detectors that detect from the waist up.

Unless you're willing to lower your goals a bit I don't think that infrared detection will be possible, because the things you list don't have the same thermal profile (I have no idea what temperature a robot would be at, and some animals are practically invisible thermally).

Video motion detection would be the most obvious approach. That would require a computer not a microcontroller, but you can get compact computers that are not much bigger or more expensive than an Arduino so that shouldn't be a problem.

Tracking objects inside a room is a complicated process, and usually requires the compliance of the objects being tracked for best results. Otherwise a combination of thermal and image are best, but not fool-proof (what if it is dark AND they aren't very thermally visible?). Beyond something like the grid you mentioned, which is relatively un-feasible due to cost and easily created blind spots, I'm not sure its even possible.

I have always thought something like this is beyond Arduino but an Atmel Simply Design contest entry may prove otherwise...

Hi lulzteddy,
This is actually quite easy to do - we make similar stuff for high end security applications based on scanning lasers.
You can make a low cost scanning laser using an SF02/F with a servo. Keith1024 has been experimenting with a fast version of the SF02/F for scanning purposes here Laser Measurement (SF02) Setup & Library - Sensors - Arduino Forum. You can map the room to establish the position of the fixed, background objects. The moving target then shows up as an anomaly against the background data.
If there are large shadows in the scanning data then it's possible to have two scanning lasers at different locations and then merge the data into a single image. If the coding is a bit daunting, I think that keith1024 is going to provide some Arduino code soon.