There is a whole bunch of smart meters on the market, meters that measure how much electricity your home is using and that track how your energy consumption is changing with time. I have one in my own home, and I’ll admit it’s changed my behavior—it’s making me want to optimize the bottom line—and see if I can get the ‘tick over’ for my house—the figure that it reports when nothing much is happening in my home—even lower. But it doesn’t tell me a whole lot about how the individual appliances in my home are working.
In the closing hours of their Kickstarter campaign, Neurio is a smart meter with a difference. It doesn’t just monitor how much energy your homes is using. It tries to figure out the cost of running individual appliances using smart algorithms that look for patterns—and potential problems.
The system will have a public API so you can extend it—connect it to web services, or write your own applications. It also comes with something they’re calling Trigger,
Trigger is an open-source application for Raspberry Pi and Android. It enables you to receive a continuous stream of power data directly from the Neurio Sensor. Trigger also comes equipped with pattern matching capability, so you can detect your appliances in real-time. With Trigger, you can detect when your garage door opens and turn up your Nest Thermostat!
Yesterday I talked to Ali Kashani—the software lead—and Jon Hallam—the hardware lead—of the Neurio project and asked them about their technology, and how it could be used by makers.
What is Neurio exactly? What’s home intelligence?
I like to use an analogy to explain this. The wiring in our homes is analogous to a nervous system that connects everything to a central place — the breaker panel. But there’s no brain there, no intelligence. That’s how Neurio fits in a home. It sits at the center of all devices, and by monitoring their signals and looking for patterns, it figures out what’s happening with each device. It can then relay that information to users by notifying them when the laundry is done, or control wifi-enabled appliances like adjusting the thermostat when you leave home.
What’s the difference between Neurio and other smart energy meters?
In essence Neurio and smart meters both monitor the same electrical signals but they’re built for very different purposes. Smart Meters are designed to facilitate billing customers, so the focus is on kWh measurements and low frequency reporting. But Neurio was designed to perform granular measurements of many attributes beside just kWh, and communicate that in real-time to the cloud. That detailed data is what drives the algorithms that allows us to do things like break your bill down by appliance or use. Neurio was also designed for consumers, so it uses WiFi rather than the less common radio formats typically found in smart meters.
Neurio learns about the appliances in your home, how does it do that?
Appliances exhibit unique patterns in the way they consume electricity. Neurio monitors the main electrical lines of a home, and looks for those patterns. It can figure out when a dishwasher is running, or when a dryer has just finished.
Is there a stage where the user is involved in training the system?
While we have designed the algorithms to work without training, there may be cases every now and then where a new device is introduced that we don’t recognize. So the users are also able to provide feedback and “train” the algorithms.
The system talks to the cloud, is that necessary, can it stand by itself?
The cloud is important because our learning algorithms perform best when they look at large data sets to find patterns and create their models. It’s similar to a search engine; Google provides every individual with better results by looking at everyone else’s search behavior. All your data remains private though. We only look at broad trends in anonymized data in order to build our models.
Neurio Trigger is what makers will be interested in, tell us more about it?
We think there is tons of potential in energy data but without the right tools, that potential will remain untapped. So we decided to start building those tools, and release them for the community to use. Trigger will be an open-source application that retrieves real-time data streams from our sensor, processes that data to infer the states of appliances in your home, and then performs some kind of action based on that state. For instance, Trigger could notice that your garage door just opened in the early evening, and welcome you home by turning up the heat. The idea is to leverage energy data in order to allow you to take ordinary appliances and get them talking with connected devices. We’re trying to hide all the complexity involved in such a process and build something that will be easy for anyone to integrate into their existing projects, and share the result with the community. We’re really excited to see what they will do with Neurio.
What can you do with it?
You could connect your dryer to your Hue lights so they flash red when your laundry is done. Or do what our hardware engineer Jon did, and make it so that your noisy robotic vacuum only runs when you’re not in the house. It can essentially take the state of an ordinary appliance, and make it trigger some kind of action.
Why just Raspberry Pi and Android? Can you talk to Neurio from something more lightweight like an Arduino?
Trigger requires more processing power than is available in a typical Arduino, but you can definitely integrate Arduino-based projects into Neurio other ways. In fact, we’re integrating Neurio with the Spark Core, which is kind of like an Arduino with superpowers. It’s Arduino-compatible and integrates a powerful microcontroller and WiFi into a small module. Our support for the Spark Core means that your Arduino projects could get notifications from Neurio when someone wakes up in the morning or arrives home at night.
You talk about an open API, and then you talk about Trigger—which is an application—are they the same thing?
Trigger uses our API to get data from the Neurio Sensor. Both Trigger and the API are open. In a sense, Trigger actually serves two purposes: it’s a really useful application that also demonstrates how to develop an application that uses our API. We’re also hoping that the fact that it’s open-source will allow the community integrate it with all their favourite projects.
In a way, it’s a starting kit with some basic capabilities already implemented in it to make it really easy to start. It will even be able to send commands to other home automation products that we’re partnering with, so it’s like a do-it-yourself kit ready to be applied to your favourite projects.
Is this what you set out to build? How has it changed over the course of moving from prototype to Kickstarter?
A lot has changed actually. For instance, we had to announce a redesign to our sensor board to accommodate the demand from all the 3-phase homes, mostly in Europe, that we weren’t expecting. We also added an expansion port as a stretch goal, in order to allow the community to build their boards to extend the Sensor’s capabilities, and do things like control appliances from within the breaker panel.
The project is now in its closing hours and over 200% funded and closing in on the last of its stretch goals.