Application of a genetic algorithm Elman network in temperature drift modeling for a fiber-optic gyroscope

Area of Science:

• Instrumentation and Measurement
• Signal Processing
• Artificial Intelligence

Background:

• Fiber-optic gyroscopes (FOGs) are crucial sensors for navigation and positioning in satellite and automotive applications.
• FOG performance is significantly impacted by temperature variations, leading to drift and inaccuracies.
• Existing error-processing techniques require enhancement to address temperature-induced errors effectively.

Purpose of the Study:

• To develop a novel error-processing technique for fiber-optic gyroscopes (FOGs) to mitigate temperature-induced drift.
• To improve the accuracy and reliability of FOGs in dynamic environments.
• To present a robust method for modeling and compensating FOG errors.

Main Methods:

• A two-part error-processing technique involving signal denoising and subsequent modeling/compensation.
• Implementation of a dynamic modified Elman neural network (ENN) for error modeling.
• Optimization of ENN parameters using a genetic algorithm (GA) based on prediction accuracy.

Main Results:

• The proposed method effectively reduces and compensates for temperature-induced drift in FOGs.
• The genetic algorithm-Elman neural network (GA-ENN) model demonstrated a 20% improvement in prediction accuracy compared to the standard ENN.
• Experimental results validated the efficacy of the denoising and compensation strategy.

Conclusions:

• The novel GA-ENN based error-processing technique offers a significant improvement in FOG performance under varying temperatures.
• This approach enhances the reliability of FOGs for critical navigation and positioning tasks.
• The study highlights the potential of intelligent algorithms in sensor error compensation.