Using of Adaptive Artificial Neural Net in a Future Launch Vehicle Autopilot

Paper number

IAC-07-D2.5.07

Author

Dr. Valery Dineev, Central Research Institute of Machine Building (TSNIIMASH), Russia

Coauthor

Mr. Andrei Efimov, Central Research Institute of Machine Building (TSNIIMASH), Russia

Coauthor

Dr. Anton Kolozezny, Central Research Institute of Machine Building (TSNIIMASH), Russia

Coauthor

Mr. Alexander Muhin, Central Research Institute of Machine Building (TSNIIMASH), Russia

Year

2007

Abstract

The aspect of use adaptive artificial neural net in autopilot is considered with reference to a task of the external disturbances compensation on a launch vehicle. Under the external disturbances the random process of wind action is considered.

One of the relevant issues for future space transportation systems is the controlling forces minimization during compensating perturbations. 

During the flight it can be reached by an operative estimation and prediction of the perturbation. For this purpose the using of adaptive neural net is offered. The selected algorithm of adaptation uses the algorithm of inverse error propagation.  

The example of an autopilot with using of a neuron network at the representation of wind perturbation as random process is considered.

Abstract document

IAC-07-D2.5.07.pdf

Manuscript document

IAC-07-D2.5.07.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.