Sustainable Homecare Monitoring System by Sensing Electricity Data

The ageing of the population has raised the interest in homecare monitoring systems and assisted living in recent years. Most studies pursue non-intrusiveness and scalability to meet the requirements of this ever-growing community. Nonetheless, they are still limited by the technique: wearables need to be carried and sensor networks imply installation. This work presents new solutions to monitor elderlies by using one single sensor: the smart meter or another similar device. This enables a new level of scalability and non-intrusiveness. Likewise, a non-intrusive load monitoring algorithm has been devised to infer the pattern usage of daily used appliances from the total power consumption. These usage patterns have been modeled using Gaussian mixture models and the Dempster-Shafer Theory to create an activity monitoring system that triggers alarms, whenever the subject has deviated from their normal routine.