Method and system for nonlinear optimal estimation of dynamic processes in real time

摘要

System for non linear optimal estimation of the state of a real time dynamic process (1) comprising measurement sensors (2) providing sampled data (3) related to the state of said process at a given time t, a calculating unit (4) associated with memories and loaded with calculating programmes, capable of recurrently providing the estimated components (5) of the state of the process at a given time t. The system is characterized in that calculating unit (4) comprises N elementary processors (401i) arranged in parallel and each receiving said measurement data (3), each of said N processors (401i) including at least one evolution random generator (404i) which, taking into account random disturbance in said evolution, provides possible values (412i) for the state of the dynamic process (1) at time t. A specialized operator (405i) calculates, based on said possible values (412i) for the state of the process and of said measurement data (3), a scalar quantity (413i) representing the probability that this state (412i) is that at time t of the process.