Adaptive neuro-controller based on IMAC scheme for InnoSAT control

Abstract

In this paper, an Adaptive Neuro-Controller (ANC) based on Internal Model Adaptive Control (IMAC) scheme is developed to adjust the control parameters for the InnoSAT attitude. Controller architecture, which is combination of IMAC with MLP network, has been outlined and its effectiveness is demonstrated on the InnoSAT system. The control signal error is used with Recursive Least Square (RLS) algorithm with forgetting factor to update the weight of the ANC. This controller has been tested using the InnoSAT system with several operating conditions. In conclusion, the ANC based on IMAC scheme is analyzed, and compared with the classical PID controller results. As can be seen from the simulation results, the ANC based on IMAC scheme is successfully implemented to control the InnoSAT attitude.