One sensor to rule all

    The Neurio smart sensor project, which two weeks before the end of the Kickstarter campaign has already exceeded the plan by $ 40,000, is based on an unusual approach to managing a smart home. Neurio is based on a very accurate and detailed energy analysis. A smart meter is similar to a conventional electronic electricity meter, but it captures much more parameters, and does it more accurately and more often than usual. These data are sent to the cloud and there they are analyzed by artificial intelligence algorithms that confidently recognize individual consumers of electricity by their characteristic “handwriting”.

    This is possible due to the fact that the Neurio sensor measures not only active power, but also reactive power, power factor, voltage, current, and frequency in two independent two-way channels. Data is updated once per second and transmitted to the home network via WiFi. In case of loss of communication, there is an internal memory that lasts for a year of continuous recording.

    In the cloud, the flow of information about the electrical characteristics of the home network is converted into a stream of messages about recognized events, such as turning on or off a certain electrical appliance or even higher-level, combined events - for example, if the garage door drive turns on and then the light comes on in the house, the system can do the conclusion that the user returned home from work. The reaction to each event is set by the user - it can be an SMS notification, turning on or off any electrical appliance - Neurio is already integrated with IFTTT and SmartThings. It is also possible to connect any user devices and scripts - the Neurio API is open.

    Included with the Neurio sensor and cloud service is the Wattson application, which analyzes the history of energy consumption and allows you to identify the most voracious consumers and optimize consumption. For paranoid people there will be an opportunity to deploy a personal Neurio cloud on their own server, though most likely you will have to pay for this with lower recognition accuracy - because the general cloud is constantly self-learning based on user data.

    Also popular now: