Abstract: The integration of kinematic equations in the Rodrigues-Hamilton parameters is characterized by asymptotic instability. The result is a significant accumulation of errors in their calculations with the relative long-term movement of an autonomous moving object (AMO). The use of neural network principles of the organization of the computational process in free-form inertial navigation systems (SINS) will significantly reduce the error in calculating their kinematic parameters and, consequently, increase the efficiency of the AMO control system. The article investigates the error of neural network calculations of the Rodrigues-Hamilton kinematic parameters for the required motion conditions of the APS. All calculations are carried out on the neuroprocessor of the AMO control system.
Index terms: kinematic parameters, artificial neural networks.

Contacts

Russia, 659305, Altai region, Biysk,
Trofimova Street, 27, room 404B
Tel. + 7-923-162-93-27
(executive secretary -
Golykh Roman Nikolayevich)
e-mail: info@s-sibsb.ru

The certificate