Pressure Pad Lecture - How to make your own 

Velostat is an amazing material, but can be difficult to house in a way that consistant readings can be obtained. I have not tested the pressure limits for these arrangements, but these designs are intended to sense low pressure levels (with a person stepping on the sensor serving as a maximum). This approach is great for detecting if and when an event occurs but is much harder to configure when the amount of pressure is needed. That being said, better results can be obtained by refining the sensor's physical enclosure and by applying filtering to the raw sensor data (usually a simple low-pass filter can garner much better data).

Hmm, good question. One possibility is to decrease the amount of overall pressure which is applied to the velostat. If it is stretched or compressed too much it can "creep" as you describe. Is there a way to reduce the amount of pressure coming into contact with the material? Perhaps you could build a brace mechanism next to the sensor to help support the load.

Hmm, you should have no trouble getting a sensor of that size to work. It might be mechanically more difficult to build but the principles are the same just scaled down. The size of the sensor does effect how the device will behave, but you should be able to make it work with some code or mechanical tweaking.

The problem is you are unable to solder directly to the velostat, it has a plastic-like consistency and soldering irons just melt right through it.

I think it would really come down to the construction of the pad. I would try to make sure that all the weight of the animal is centered on the sensor. If that is not possible try to have four sensors on each corner and perhaps on in the center. Any method should work, just analyze the output data in the serial monitor and determine where your threshold is. If you are not getting enough "action" from the sensor still try to reduce the size of the sensors and the size of the contact area, this way more pressure is applied per square inch and you can get a better reading with less weight. I hope this helps. Share your creation if you can and are willing =)

Velostat itself can get wet without any issues, the primary issues would be unintentional bridges between wires and corrosion of the metal components of the sensor. As long as you spend some extra energy to ensure that any exposed metal is securely sealed either inside an IP enclosure or by using a product such as "liquid tape" or "plasti-dip" to seal the exposed components you should be good to go.

If you are using for Pet Alert, just get a regular motion detector and a small light or alarm. Once the Light or Alarm on, you will know.

I agree, pressure pads have very limited use-cases where they are the "best" technology. Usually a motion detector (ultrasonic or IR) are an easier to use and more effective choice. Especially for detecting things which dont weight too much.

I would probably go with Velostat. I imagine it would be a big plus to have an analog value to work with instead of a simple digital signal. The analog signal you can get from Velostat would allow you to map the pressure on your sensor to the amplitude of the samples. If you use the simple digital methods you will only be able to tell is the pad has been hit but will be unable to know how hard it was hit.

It depends on what you want the sensor to do. If you just want a digital output of on or off the button method is probably good enough, but if you want to change a function based on how much weight or pressure is on the sensor you should probably go with the velostat.

I am getting mixed information on what the top layer should be above the velostat. One guy on flashgamer say use a conductive/knit material. Do the top layer need to be conductive or will carpet work?

So the video highlights 2 ways you can set up the velostat as a pressure sensor. Check out 4:38 and 7:19. But in either way you could put carpet over the entire set up.

Hmm, that would be a very complicated project and something that would likely be very hard to make in a "hobby" environment on your own. However, that being said, some sort of a capacitive system would likely work best. None of the methods discussed in this video would provide you with an X/Y position.

@ Even im also thinking about the same. Can you please make a video on getting x/y co ordinates ? Thanks for the info.

use a wacom tablet or cintiq