​Schneider Electric is currently developing an affordable, non-invasive assisted-living solution to detect unusual situations in elderly people's homes. And the solution leverages two major innovations. First of all, it uses a low-cost network of sensors selected because they are both non-invasive and flexible. The sensors are not worn by the patient; they are installed in the living environment and provide information about the occurrence of events like opening and closing doors, the patient's presence in strategic rooms, and more.

Plus, the fact that the sensor network is personalized makes it even more effective. "We developed an algorithm that factors in the patient's habits and time of day to pinpoint any abnormal situations." The algorithm just needs a simple data log to automatically determine "reasonable" alert thresholds, ensuring an excellent compromise between reducing false alarms and achieving a high incident-detection rate. The statistical method used works independently of the type of sensor and the algorithm improves its performance over time. The solution has been evaluated in test homes and is currently being assessed by the manufacturer.