Tuning of the Estimation Procedure for the B-BAC Methodology

This paper deals with the problem of the modeling accuracy of the balance-based model, which is a basis for the B-BAC (Balance-Based Adaptive Control) methodology. This model has the nonstationary form because all the modeling inaccuracies and uncertainties are represented by the only one time-varying parameter that must be estimated on-line. Thus, the estimation procedure is a very important part of the B-BAC methodology and it influences the modeling accuracy and, consequently, the control quality. Tuning of this estimation procedure requires the choice of the values of two parameters and in the practice, this choice must be made without the knowledge about the true process. In this paper, we suggest the optimal choice of the tuning parameters by simulation on the basis of the comparison with the true system. The experiments have been carried out for both the noiseless and the noisy cases and the results allow formulating the suggestions for tuning of the estimation procedure in the practice.