A constraint selection technique for recursive set membership identification

Recursive approximation of the set of feasible parameters is a key problem in set membership identification. In this paper a new technique is presented, aimed at recursively computing an outer orthotopic approximation of polytopic feasible parameter sets. The main idea is to exploit the concept of binding constraints in linear programs, in order to select a limited number of constraints providing a good approximation of the exact feasible set. Numerical tests demonstrate that the proposed technique outperforms existing recursive approximation algorithms, with a limited increase of the required computational burden.