Aerospace Instrument-Making Annotation << Back A neural network identification of mobile geodesic platform kinematical parameters Devyatisil’niy A.S. K.A. Chislov A problem to determine the rotation parameters of geodesic platform with wide spectrum of basing is investigated. For identification of corresponding kinematic parameters (orientation matrix and angular velocities) the inertial-satellite integrated system is proposed. It consists of onboard network of GLONASS receivers and gyroscopic angular velocity sensors. An integration of navigation information is realized by artificial neural network based on multimodel construction of Kalman-type dynamical inversion algorithm. A mathematical model of the integrated system is represented as model of «state-measurement»-type inverse problem, where state equations are the Poisson’s kinematical equations. In article is given the development of representations about neuromorphism and is suggested the mechanism of adjustment of neural network synaptic coefficients. Some results of mathematical simulation, which illustrate the efficiency of the problem solution, are presented. Key words: navigation satellite system, gyroscopes, technological platform, artificial neural network, in-verse problem Kalman filter. Contacts: E-mail: kirillche@rambler.ru Pp. 16-21.