On - line Parametric Identification and Discrete Op timal Command of the Flying Objects' Move

This paper presents a new online parametric identification and discrete optimal command algorithm for mono or multivariable linear systems. The method may be applied with good results to the automatic command of the flying objects' move. The simulation results obtained with this real time algorithm, with parametric identification for the longitudinal and lateral move of an aircraft are also presented. This algorithm may be used, with good results, for identification and optimal command of an airair rocket's move in vertical plain regarding to target's line (1). ECAUSE of fast flying parameters' modify for the modern a ircrafts and rockets, performing real time identification and optimal or adaptive command algorithms have to be made. The authors of this paper have made such an algorithm. First of all, an offline parametric identification is made, without command, for obtaining the initial values of these parameters for the online identification process. Using the control system ) A ( and model's outputs, a d iscrete optimal command law is projected, using a quality quadratic criterion, which assures the convergence of the difference between control system and model's outputs. The model's parameters, obtained by the online identification, are used for calculus of the command law. For the algorithm validation one uses as examples the automatic command of an aircraft longitudinal and lateral move or the rocket's move in vertical plain; time characteristics, representing evolution of state variables of A and their estimate, are plotted. These variables' stabilization and the convergence of the errors i i i x x e ˆ − = happen in maximum 2 seconds. The proposed algorithm produces very good results in the case of lateral move's stabilization for transport and fights aircrafts or in the case of parametric identification of an airair rocket's move in vertical plain regarding to target's line.