Hi Raj,

I once had a customer (about 1982 or 1983) who had a large dog with his own pet door. The dog was allowed to come and go through his pet door (or doggie flap) even when the alarm was active. His "rest area" was in the back yard. But since the dog was so large, his pet door was large enough for a human to fit through. So we had to secure the flap.

I did some experiments and found that there was a marked difference between the elapsed time it took the dog to go through and that of a person. I put a magnetic contact on the pet door, that triggered open, and put a parallel capacitor across the switch. This meant the capacitor had to charge before the panel would recognize the alarm condition. I then put a resistor in series, setting the time constant long enough that the dog would not trip the alarm but a person would.

If you can make the light curtain you're describing work, discerning a human versus a pet could be as simple as looking at how many receivers are blocked. A dog walking through is going to block no more than a foot or two off the floor. A person on the other hand will block five or six feet from the floor up.

If you're concerned about someone laying down, where they would now only block the bottom foot or so of sensors, you would need to look at how long the sensors stay blocked. A person lying flat is going to take a lot longer to slide through than someone walking vertically.

Then just think about other ways a person could move through the light curtain. If they go through on all fours, the last foot and a half or so is only going to block the bottom 5 or 6 inches of sensors (their legs from the knees down). Comparitively, a dog or cat will also break an inch or so of sensors in the last foot or so of motion, but it will be off the floor by a half foot to a foot or so (their tail).

These are a few examples of how to discriminate pets from people. How well your final algorithm works will depend on how many such scenarios you can reliably define.

Good luck.

Joe