State space identification of flight dynamics models

This paper is devoted to the state space identification of the flight dynamics models in the presence of sensor noise and biases. The goal of the identification procedure is not only the estimation aircraft stability and control derivatives, but also the biases of sensors. It is achieved by using the procedure of the likelihood function minimization, based on the Kalman filter and the stochastic approximation procedure. The application technique of the least-squares method to a state space model in order to determine initial values of unknown parameters which are necessary to identify the state space model by maximum likelihood method is created. This procedure was used for state space identification of the model of lateral-directional dynamics of small 6-seat aircraft and results of this identification are presented.