Energy consumption is a current concern in many aspects of daily life. The concept of smart buildings, where researchers propose solutions to tracking energy consumption by installing sensors and conducting smart analysis to predict and/or prevent future energy usage, is important. In this way, both users and suppliers are aware of how energy consumption is distributed.

One of the mostly widely used techniques for capturing appliance usage is non-intrusive load monitoring (NILM). At one extreme, energy is tracked at the building level; for example, smart energy meters are installed outside the building and provide only the total sum of power consumption. It is clear that this solution can only predict the main load. Another extreme case is when each smart plug is inserted into individual electrical sockets. This solution increases the monitoring cost significantly.

In this work, Roy et al. propose an alternative solution that falls between the two extreme cases while still keeping the finer granularity. The idea is to install the circuit breaker at room level, and then a novel analysis and algorithm are employed to “help disaggregate such measurements to infer the use of even relatively low-power appliances.” The contribution of this work is the proposed algorithm, which “provide[s] very high accuracy in identifying device usage” at a fine-grained level.

The proposed algorithm will be very useful for researchers in the smart buildings and sensors field. The paper provides the general public with a good vision of how energy consumption is monitored and how analysis is conducted.