Corresponding author: Devyatisil'nyi Aleksandr S., Chief Researcher, Institute of Automation and Control Processes (IACP), Far Eastern Branch of the Russian Academy of Sciences (FEB RAS), Vladivostok, 690041, Russian Federation, e-mail: devyatis@iacp.dvo.ru

Received on December 23, 2014
Accepted on February 05, 2015

The authors of this paper demonstrate how, with the on-board multi-position information received from a satellite navigation system and the inertial information delivered from three dimension accelerometers with orthogonal axes of sensitivity, forming a movable coordinate trihedron, for the latter one the orientation and angular velocity of the rotation in projections on its own axis can be determined. The mathematical model of the inverse problem in the "state-measurement" form is presented as 1) a dynamic group of equations of functioning of the inertial navigation system (without gyroscopes), with the state vector, which includes the coordinates, specific impulses and angular velocities of the instrumental trihedron rotation, and 2) equations for measurement of the object's location coordinates, identified with the coordinates of trihedron tops in the projections on its axis. A dynamic pseudo inversion (solution) of the problem is realized by the neural network, the model of which is based on Kalman-type algorithm in its multi-model presentation, which allows a judgment about a fair competition between the models in the progress of the state vector evaluation. The concept of "nuclear" and "nuclear-free" setting mechanisms of the neural network synaptic coefficients is introduced. Hypothesizes about a possibility of broadcasting of the characteristics, used in the description of an artificial neural network, are introduced in presentations concerning organization and functioning of the populations of the biological ("live") neurons.
The results of the computational experiments are presented.

Keywords: angular velocity, accelerometer, satellite navigation system, neuromorphism, inverse problem, algorithm of solution

For citation:
Devyatisil'nyi A. S., Chislov K. A. A Model of Competitive Neuromorphic Aggregation of Information in the Integrated Satellite-Inertial Navigation System, Mekhatronika, Avtomatizatsiya, Upravlenie, 2015, vol. 16, no. 10, pp. 716—720.
DOI: 10.17587/mau.16.716-720

To the contents