We use the term smart so much nowadays that it is possible to forget what it actually means. We have smart vehicles, smartphones, smart watches, and smart homes. But what does smart really mean? What makes the smartphone smart? Is it the biometric access enabled through fingerprint scanning or facial recognition? Is it the automatic rotation of the screen based on the physical orientation of the device? Or is it the device’s connectivity to the internet?
Till Recently, Connectivity Was All It Took To Be Smart
When it comes to smart homes, a quick search on the internet will reveal hundreds of smart home products. One of the special qualities that a huge number of so-called “smart” appliances have is that they can be controlled via apps on smartphones.
There are smart bulbs which are marketed as having the ability to help homeowners fall asleep at night and wake up in the morning simply by adjusting the light. Such changes can be made through an app on the user’s smartphone, which supposedly makes this lighting system smart.
Whether the above lighting system should be referred to as smart may be a debatable question, depending on an individual’s point of view.
However, such products have been the hallmark of smart home technologies for some time now. If you can change the music on your stereo system at home simply by talking, then it’s smart. If you can change the temperature in a room by touching your smartphone screen, then it’s smart.
But thanks to advances in technology, such as the introduction of the smart home chip and the AI accelerator module, manufacturers are redefining smart home technologies. We are taking the smart home to the next level, hopefully one that better deserves the label “smart.”
Shouldn’t The Term Smart Be Related to Intelligence?
Smart televisions have been all the rage for quite some time. When they were introduced, being able to watch YouTube on a big screen, as opposed to watching it on your mobile device, might have been considered revolutionary. So was the ability to conveniently stream your favorite Netflix shows on your television screen. And with some smart televisions, you could browse the internet.
That was all it took for the television to be smart: access to the internet.
According to the online version of the Merriam-Webster dictionary, the word smart refers to an excellent ability to learn and think about things. It can also be used to refer to the ability to exercise good judgment.
The television that can access Netflix and Google doesn’t seem so smart now, does it? The Amazon ecommerce website can learn about a user’s preferences and suggest new products based on what it has learned. That is smart. But the current “smart” television can’t learn about its user. It can’t think. And it definitely doesn’t have the ability to exercise good judgment. If it did, it would probably be able to stop you from binge watching the latest season of your favorite Netflix series way into the morning.
For appliances and technologies used in the smart home to really be smart, they should have the ability to think, that is, they should be able to process data and derive meaningful insights that can inform decisions.
They should also have the ability to learn, meaning that automatic optimization should be on the table.
And lastly, they should be able to exercise good judgment. For example, in a smart home, all smart devices should conspire to minimize the wastage of energy.
Enter VPU technology such as the smart home chip and suddenly, a home with true smart technology is in the cards.
How VPU Technology Enables True Smart Devices
What Does a True Smart Device Look Like?
Picture a lighting system that you don’t have to control via your smartphone and that automatically adjusts to provide you with the most convenient lighting experience possible in your home.
As opposed to some current “smart” lighting systems, you don’t have to reduce the level of blue light through your smartphone when going to sleep. The new smart system would have learned the time when you normally go to sleep. It would track your movement into the bedroom and adjust the light accordingly after having switched off the lights in the other rooms. It would then track your movement into the bed and switch off the lights or dim them – according to your preference.
And if you were reading in bed, it would notice the book and provide appropriate lighting for reading: enough to read comfortably but without blue light so that your body can prepare for sleep.
In the morning, when you usually wake up, the smart system would adjust the light to help your body wake up.
Such adjustments in lighting are considered important because light affects the production of melatonin, a hormone that affects the body’s sleep-wake cycle. Darkness triggers the production of melatonin, which helps the body sleep. On the other hand, light reduces the levels of melatonin in the body.
A truly smart lighting system could help with better sleep. But how would such a system work?
VPU Technology in the Smart Home
To implement smart home systems that learn, think, and exercise good judgment, certain conditions must be met.
To start with, data must be reliably collected. In the smart lighting example used above, such data can be collected through cameras spread out through the home. The footage should be of enough quality to enable the next stage, that of processing.
After smart home systems have collected data, they should be able to analyze it to derive actionable insights. In a smart lighting system, object analysis can help track a person’s movement into and out of rooms.
And the final basic ability of true smart systems is that of making decisions. A smart home surveillance system should be able to perform facial recognition on a person approaching the house. If the person is a stranger and he or she attempts to access the house, the system should send out an alert to the homeowner and probably trigger the alarm system.
For a surveillance system to be capable of such functions, the feed should be linked to a device with processing capability, such as an AI accelerator module. Such a module is usually made up of specially made processors called vision processing units (VPUs).
What Makes VPUs Appropriate For Smart Home Technologies?
VPUs are designed to facilitate neural processing and machine vision. Neural processing helps machines and computers think and learn like humans. Modern VPUs have parallel processing capabilities. In addition, they implement minimal data transfer, which minimizes power consumption.
The ability of VPUs to deliver powerful processing while using minimal energy makes them suitable for processing at the edge.
Edge processing means that instead of sending data to the cloud to enable smart systems to make decisions, the data is processed within the smart home system. This makes real-time applications such as the use of gestures to switch off the lights possible. It also eliminates the privacy concerns that come with sending smart home data to the cloud.
Conclusion
Since “smart homes” became a popular buzzword in the media, most smart technologies have been considered smart because of connectivity. However, technology has advanced to the point where we can have truly smart devices, capable of thinking, learning, and making helpful decisions.
And thanks to technologies such as the AI accelerator module, processing can be brought to the edge, making smart home technology all the more efficient.
