Aerospace Instrument-Making Annotation << Back A MODEL OF INTEGRATED NAVIGATION SYSTEM FOR MANEUVERABLE OBJECT A.S. Devyatisilniy, K.A. Chislov The article deals with estimation of motion parameters of highly maneuverable ob-ject, i.e. the parameters of the trajectory and the rotation of the object as a rigid body, when integrated system that solves this problem is represented by 1) the theo-retical model of object’s trajectory (Newton's equations, written in mobile rotating frame of reference), is the base for the construction of algorithms for assimilation of 2) inertial information delivered from the unit of linear accelerometers, and 3) infor-mation from the satellite navigation system with on-board multi-positions reception. A mathematical model of the integrated system is interpreted as a model of the inverse problem of the «state-measurement» form. To solve the inverse problem of the neural network is proposed, which implements pseudoinversion (dynamic inversion) based on Kalman filtering algorithm in its multi-model interpretation with the tuning mechanism of synaptic coefficients which in article by analogy with biological prototype called the «non-nuclear». The results of computational experiments are convincing evidence that the proposed integrated system without gyroscopes can effectively solves the problem of estimation of the angular velocity of highly maneuverable object rotation. Key words: angular velocity, accelerometer, navigation satellite system, neural network, inverse problem, pseudoinversion. Contacts: E-mail: kirillche@rambler.ru Pp. 3-8.