Abstract:

In order to solve the problem that temperature drift errors (TDE) of interferometric fiber optical gyroscopes (IFOG) can not be compensated exactly, a modified compensation system for TDE of IFOG is designed with radial basis function artificial neural network (RBF-ANN). According to the detailed analysis of the thermal induced nonreciprocity phase delay, a TDE model is built based on the temperature, temperature variation and temperature product term; the RBF-ANN is trained to identify the parameters of the TDE model precisely; the compensation system is realized based on a microcontroller TMS320F28335, a high-precision temperature measurement unit, a decode-encode unit and an auxiliary unit. The methods to design the compensation system and the realization steps are shown in this article. The comparable results of temperature experiments show that the compensation system is high-accurate and wellreliable, and the actual output precision of IFOG can be raised to a high order of magnitude. The modified compensation system for TDE of IFOG can be used in compensation for TDE of IFOG.