Sense — device provides a newsfeed showing what your garage door openers, toasters, microwave ovens, washing machines, heaters, and refrigerators have been up to all day.
A startup in Cambridge, Massachusetts, is the first to offer a consumer product that reads incoming household power levels a million times per second—enough to tease out telltale clues to which specific appliances, even low-wattage ones, are operating in real time. “It’s at the cutting edge of what I have seen people attempting in this area,” says Michael Baker, a vice president at SBW, an energy efficiency consultancy in Seattle.
The company says it can accurately disaggregate 80 percent of home energy use; it can do things like detect a microwave oven through its very specific startup and operating power “signature,” or sense a washing machine thanks in part to subtly increasing demand on the motor as the drum fills with water. As it identifies garage door openers, toasters, microwave ovens, washing machines, heaters, and refrigerators, it displays them on an app as a newsfeed and a series of labeled bubbles.
The Sense home energy monitor:
Sense—founded by speech-recognition veterans whose technology ended up in Samsung’s S-Voice and Apple’s Siri—consists of a box about the size of an eyeglasses case installed inside or next to an electrical service panel. Two inductive current sensors sense current, and two cables power the box and sense voltage. The box does some onboard processing, and then uses Wi-Fi to send data to the cloud for further analysis and aggregation with data from other users to improve its accuracy.
While Sense’s initial business model is based on selling the hardware for $299, the long-term play is in the data: Sense will retain rights to the data and expects to eventually serve personalized recommendations. It also hopes to sell anonymized data and insights to companies like utilities or insurance companies.
The utility already has so-called “smart meters” that collect data at eight- to 12-second intervals and plot it for customers in an app, but “we found that data at this resolution isn’t all that interesting,” says Todd LaMothe, a software development manager at the utility.
A number of consumer systems for monitoring home energy exist—but are generally of low resolution and look only at entire-household energy use. Navetas allows you to track electricity consumption in real time, analyze trends, and set goals, but does not disaggregate what is causing the load.
Bidgely services utilities that have installed smart meters. By sampling meter data every few seconds, it can spot trends or major anomalous events, like a big overnight load that suggests an appliance such as an electric oven was accidentally left on. Another startup, Neurio, is developing a similar system but is only able to see high-wattage devices.
“Intelligence in the home starts with good data about what is going on, so that’s our focus right now—developing that data,” says Sense CEO and cofounder Michael Phillips, who a decade ago cofounded Vlingo, a voice-recognition startup that developed speech recognition for mobile phones and virtual assistants. “Until now nobody has been able to make this work, because the real world is more challenging than expected.”
So far it has detected my fridge, washing machine, and dryer. I was surprised to see a big spike in demand at certain times when my dishwasher was running.
via: [MIT Tech Review]
