Modeling and optimizing of the random atomic spin gyroscope drift based on the atomic spin gyroscope

Area of Science:

• * Inertial Navigation Systems
• * Sensor Technology
• * Signal Processing

Background:

• * Atomic spin gyroscopes (ASGs) are crucial for high-precision navigation.
• * Random drift in ASGs, caused by nonlinearities and instability, limits operational accuracy.
• * Existing error modeling techniques struggle to fully compensate for complex ASG drift characteristics.

Purpose of the Study:

• * To develop a hybrid model for compensating random drift errors in ASGs.
• * To improve the operational accuracy and stability of ASGs.
• * To establish a robust error model for ASG time-series data.

Main Methods:

• * Established a hybrid random drift error model using autoregressive (AR) and genetic programming (GP) combined with a genetic algorithm (GA).
• * Analyzed and preprocessed measured ASG data to obtain time series of random drift.
• * Developed a linear section model using AR technique and a nonlinear section model using GP, with GA optimizing GP-derived mathematical expressions.

Main Results:

• * The hybrid model effectively captures the characteristics of ASG random drift.
• * Achieved a 92.40% reduction in the square error of ASG random drift.
• * Demonstrated superior performance compared to standalone AR (9.34% reduction) and GP+GA (5.06% reduction) techniques.

Conclusions:

• * The hybrid AR + GP + GA modeling method effectively compensates for ASG random drift.
• * This approach significantly enhances the stability and operational accuracy of atomic spin gyroscopes.
• * The developed model offers a promising solution for improving ASG performance in demanding applications.