'smart homes' could track your electrical noise - household electrical appliances

Through Kurt kleiner, one day, when you turn on and off your home appliance, your computer can track your movements around the house by monitoring the noise from your home appliance.
This system may be cheaper and simpler than the sensor kit researchers currently envision for the "smart home.
For a long time, scientists and visionaries have always believed that smart homes can satisfy every whim of us.
But their ideas all depend on the cameras, microphones, or other sensors in each room to track the location of the occupants.
This new approach relies on a device that plugs into a single standard wall socket that monitors power noise due to electrical switches.
The computer that monitors the device can then infer that a person must be in that location.
"The problem I 've seen in a lot of pervasive computing research is that it needs to create new infrastructure and technologies.
Gregory D. abward said: "A lot of the things that we 've been focusing on are how you can achieve these things without the need for Joe Thorpe to buy something new, "computer scientist at Georgia Institute of Technology in Atlanta.
In a paper to be published next week at the International Conference on Pervasive Computing in Innsbrook, Austria, abowd and colleagues said that when their lights are turned on and off, they can train computers to distinguish between activities in different rooms, or tell the features of the microwave from the ceiling fan.
To do this, they use a laptop to analyze the frequency of noise generated in the home power cord when the appliance is turned on and off.
Researchers train computers to identify specific devices by turning them on and off.
All in all, they tested 19 different electrical devices in six different homes with an accuracy of between 85 and 90%.
This system may be useful for a family
An automated system, for example, turns on and off the heating or sound system when people enter and leave the room.
This may also be useful for monitoring the level of activity of older persons living alone.
The main drawback of this approach is the amount of time it takes to train a computer-about four hours for a typical house.
Although Abowd pointed out, it is also time consuming to install cameras and sensor networks.
Researchers still need to prove that the system can differentiate individual devices even if other devices are running.
For this study, they only turn on and off one device at a time.
And the system has obvious restrictions & colon;
If you don't turn anything on or off, it won't know where you are.
Still, Gede kotham, a computer scientist at Lancaster University, says it still has potential.
"It is wise to take advantage of the existing infrastructure. ” .
But it is not clear whether the new technology can completely replace the sensor, he said-more likely, the system will be used in conjunction with at least some sensors.