The alarm systems in the field of physical security have two specific problems of dependency.
A direct dependency on central alarm monitoring system, which involves a monthly cost, that can make the service unaffordable to middle or low income households, and other dependency for the deactivation of the system by presence, caused by the need to identify when the users are in specific controlled areas or not, requiring, in this cases, the deactivation of the system, fully or partially.
This project began with the need to provide greater value to the alarm systems related industry, treating in a "smart" way the different situations that cause alarm triggers to discern, based on patterns of behaviour, if that event has its origin in an abnormal or dangerous situation, or, on the contrary, it is a normal situation in the area controlled by the alarm.
This project allows alarm systems to always be connected (armed) obviating the current need for connection / disconnection of these systems according to different everyday situations: presence or not at home / office / factory, end of working hours, etc.
"Ideally, an alarm system available to every consumer should be created, with a very low fixed cost of maintenance, and active at all times, regardless of the presence of users in the controlled areas."
Consequently, we have developed a tool capable of detecting when an alarm trigger corresponds to an anomalous or potentially dangerous situation throwing at this time the relevant warning to the central alarm and avoiding false alarms or triggers by normal and usual situations.
In itself, the system represents an innovative paradigm in the context of so called “smart alarms” due to the qualitative leap that means the partial elimination of the customer verification systems by multicontextual inference derived from information extracted from all available devices (cameras, sensors, etc.).
In addition, the whole system developed on detection tools is formed by an inference unit consisting of artificial intelligence algorithms that process the information received to provide an intelligent real-time response. This integration represents one of the first applications of artificial intelligence in the field of physical security at European level.
In summary, our Algorithmic Processing Units receive all alerts interactions with the system and processes them to make the inferences with the data already collected. The answer to a particular alert is sent to the client system, so the two systems are fully related but disjoint.