Method and system for merging data or information

A method for merging information of high semantic level representing complex situations comprising several items of information or data originating from several sensors, said information or data possibly being heterogeneous, said method being executed on a processor and comprising:

• acquiring the various items of information or data arising from at least two sensors in raw form and transforming them using said processor so as to cast them into the form of conceptual graphs, a conceptual graph representing several concepts and the relations which exist between them, a conceptual graph being composed of several entity nodes and of relation nodes, with E a set of concept nodes defined on a support S, G1 and G2 two conceptual graphs defined on S, the graphs thus obtained being stored in a database;
• defining a knowledge base containing information specific to a sector of application of data merging and rules applied in said sector of application, and transforming said information or data using the processor so as to present it in a form of conceptual graphs;
• determining a merging strategy denoted strategymerge as follows:
• strategymerge=fmerge◯fcomp:E×E→E∪{E×E}, where fmerge:{true, false}×E×E→E∪{E×E} is a merge function for merging concept nodes of graphs, and fcomp:
• E×E→{true, false}×E×E is a function for testing compatibility between two concept nodes of the graphs;
• applying the merging strategy to the various items of information arising from the knowledge base and from an observation base taking the form of graphs to produce merged data;
• and
• sending the merged data to a decision making system, wherein
• the two conceptual graphs G1 and G2 have a common generalization H, and
• projections P1:H→G1 and P2:H→G2, P1 and P2 are compatible according to the function fcomp if, for each concept c of the graph H, the following conditions are complied with:
• P1(c) and P2(c) have a common sub-type different from an absurd type,
• referents or components of P1(c) and P2(c) conform to their most general common sub-type,
• the referents of P1(c) and P2(c) are either equal, or one of the two is undefined, or fcomp(P1(c), P2(c))=(true, P1(c), P2(c)).