Method for pattern recognition in a plurality of signals

A method for pattern recognition in a plurality of time signals of different types for medical or industrial applications, a hierarchical model of standard contexts being defined, each standard context being associated with a respective level of the hierarchical model, the method comprising the steps of:

• a) receiving the plurality of time signals;
• b) for each received signal of the plurality of signals, creating an asynchronous time signal comprising events, said events being of a same form as the asynchronous signal representative of a characteristic of the received signal;
• b′) using the events of each of the asynchronous signals as current events and using standard contexts of a first level of the hierarchical model as current standard contexts;
• c) for each of the created asynchronous signals, creating an activity profile of said asynchronous signal, the activity profile comprising at least one activity value that decreases as a function of the time elapsed since an event that is a most recent event among successive ones of the events of said asynchronous signal;
• d) for each given time among a plurality of given times:
• d1) determining a context being defined as a set of the values of the activity profiles, at the given time, of the created asynchronous signals,
• d2) determining a standard context among the current standard contexts, the determined standard context having, among the set of current standard contexts, a minimum distance to the determined context, and
• d3) when a level of the hierarchical model has not been used, generating an event for one of the asynchronous signals as a function of the determined current standard context;
• e) when the level of the hierarchical model has not been used:
• using the generated events generated in step d3) as current events, and
• using the standard contexts of a level of the hierarchical model not yet used as current standard contexts;
• re-executing steps c-e;
• and
• f) when all of the levels of the hierarchical model have been used:
• determining a pattern in the plurality of received signals via a comparison of the standard contexts determined at a last occurrence of the step d with a base of standard signatures.