How to Make Your Sensors Smarter

Building the Internet of Things from Smarter Sensors

Alasdair Allan
3 min readMay 11, 2017

To fulfill the real promise of the Internet of Things there needs to be computing and sensing in each and every object in your home. In your office. In your neighborhood.

But not only would that be impractical, and expensive, it would be intrusive, and awkward. So what if, instead of deploying thousands of sensors in a room, you could install just a single sensor that could indirectly monitor the whole room?

Prototype general-purpose sensor tag. (📷: Carnegie Mellon University)

Gierad Laput, Yang Zhang, and Chris Harrison — researchers at the Future Interfaces Group of the Human-Computer Interaction Institute at Carnegie Mellon University — have been working to do just that. Their sensor tags allow everyday locations to become “smart environments” without invasive instrumentation.

“The idea is you can plug this in and immediately turn a room into a smart environment. You don’t have to go out and buy expensive smart appliances, which probably can’t talk to each other anyway, or attach sensors to everything you want to monitor, which can be both hard to maintain and ugly. You just plug it in to an outlet,” said Laput.

A single sensor board with multiple sensors monitors a number of sensor data streams. But instead of all this low-level sensor data being presented to the user from the individual sensors—such as accelerometers, pressure, temperature, humidity, electro-magnetic, and acoustic sensors—data from all these different sensors are joined together into ‘synthetic sensors.’

Stack spectrograms of accelerometer, microphone, and EMI sensors. (📷: Carnegie Mellon University)

The raw sensor events are combined, using machine learning to automatically recognise distinct patterns of sensor activation and expose these as higher level events. These are events that have meaning in the real world, like you turning on a faucet in the kitchen, or opening and closing the door of the microwave.

Remotely detecting a faucet being turned on. (📷: Carnegie Mellon University)

By then constructing rules engines around these real-world events you can build up a picture not only the current state of the world around you, but how the world works.

“It can not only tell you if a towel dispenser is working, but it can also keep track of how many towels have been dispensed and even order a replacement roll when necessary,” Laput added.

Then from these more advanced events you can infer human activity around the sensor; such as when someone is sleeping, showering, or watching television.

The group has already started to build prototype hardware to prove their ideas, and at the heart of their synthetic sensor tag design? A board most makers are probably going to be familiar with, a Particle Photon.

Want to delve deeper into the project? You can download the team’s paper, which was recently presented at the Conference on Human Factors in Computing Systems, in Denver.