RECURSIVE ALGORITHMS FOR IDENTIFICATION IN CLOSED-LOOP - A UNIFIED APPROACH AND EVALUATION

A unified presentation of recursive algorithms for plant model identif ication in closed loop is given. From the basic formulation of the pro blem of finding the plant model that gives the best predictor for the closed-loop system, two families of algorithms using either a reparame terized predictor for the closed loop or a plant predictor operating o n filtered data are presented, and their asymptotic properties are exa mined. Validation tests for the models identified in closed loop are p roposed. A comparative evaluation of the various algorithms in simulat ion and on real-time experiments concludes the paper. (C) 1997 Elsevie r Science Ltd.