Due to the advances in polarization-maintaining fiber technology, coil-winding process technology, and other optical fiber device technologies, high-precision fiber optic gyroscopes have been made possible for high-volume applications in navigation. However, in many unmanned vehicle platforms, the size, weight, power consumption, and cost control of the navigation system have high requirements, so it is not operable to suppress the temperature drift problem of fiber optic gyroscopes by adding a temperature control system. The fiber optic coil is the core component of the fiber optic gyroscope, but its preparation process requires the intervention of manual operation, resulting in a difference and poor consistency in product design and finished product, and its performance will deteriorate due to the external temperature perturbation. Although the development of low-temperature-sensitive optical fiber can improve the winding process to improve symmetry, optimize the cavity structure design, and slow down the rate of temperature perturbation to strengthen the gyroscope's self-suppressing ability of temperature drift, the gyroscope's temperature performance degradation caused by some human factors, device defects, and other factors is still unable to be effectively solved. Based on this, by relying on the fiber optic gyroscope system platform, we find the relationship between gyroscope output and temperature and other related factors and use algorithmic compensation to weaken thermally induced error effects of fiber optic gyroscopes. In this thesis, we start from the mechanism level and discuss and deduce in detail the deep-seated reasons for the deterioration of gyroscope performance due to the phase error caused by the temperature influence of the fiber optic coil, which is the core component of the fiber optic gyroscope, and we carry out the process correlation analysis of the influence of the temperature factors and put forward a new type of zero-drift polynomial temperature compensation model that can be realized in an engineered way. The proposed compensation scheme based on this model is verified to be effective and can significantly suppress the gyroscope temperature drift error.

Through the fiber optic gyroscope temperature drift profiling derivation, the deep-seated causes of gyroscope drift error caused by temperature perturbation are analyzed, and the correlation of each temperature term influence factor with the actual output of fiber optic gyroscope is verified by combining with the process correlation theory. It is found that a temperature sensor can only characterize two temperature factors, temperature and temperature variation rate, but not the temperature gradient factor. Simulation analysis based on the process correlation theory shows that the compensation effect can be improved by introducing the temperature gradient factor under variable temperature conditions. In view of the fiber material properties, when the temperature changes, it will cause the material properties to change. Through simulation analysis, it is found that the output of the fiber optic gyroscope is correlated with the coupling factors of the product of temperature, temperature variation rate, and temperature gradient under variable temperatures. Finally, based on the relevant theoretical analysis, a temperature compensation algorithm model is established by simultaneously considering the temperature, temperature variation rate, temperature gradient, and the product of the three factors, and the validity of the model is verified through experiments.

Through theoretical analysis, it is found that the fiber optic gyroscope cannot characterize the temperature gradient factor with the help of only one temperature sensor. Therefore, an implementation method of adding two temperature sensors inside the fiber optic gyroscope is proposed (Fig. 4). Simulation analysis with the help of process correlation theory (Fig. 6) reveals that there is indeed a correlation between the output of the fiber optic gyroscope and the temperature gradient factor during the temperature change process, which further supports the accuracy of the aforementioned theoretical analysis. Therefore, the temperature gradient factor is introduced when the compensation model is established. Through the offline comparison simulation test, it is concluded that the compensation model considering the temperature gradient factor can further improve the accuracy of the compensation model and enhance the gyroscope temperature performance (Fig. 7). In addition, by further analyzing the mechanism of thermally induced error in fiber optic gyroscope and verifying it with the help of process correlation theory simulation, a more comprehensive compensation model (Eq. 12) is proposed by simultaneously considering the temperature factor, the temperature variation rate factor, the temperature gradient factor, and the coupling term of the product of the three factors. Finally, the new temperature compensation model is verified to be more accurate and better compensated by means of multi-sample experiments (Fig. 11).

In this paper, a new multinomial algorithm compensation model is proposed, which simultaneously considers the temperature factor, the temperature variation rate factor, the temperature gradient factor, and the coupling term of the product of the three factors. We combine the analysis of thermally induced error mechanism derivation and process correlation theory as the designation idea and analyze the temperature compensation algorithm model of fiber optic gyroscopes. The feasibility of the algorithm is verified through offline compensation, and the zero-bias stability accuracy of the gyroscope after compensation is significantly improved compared with the compensation algorithm that only considers three factors, namely, temperature factor, temperature variation rate, and temperature gradient. In addition, the compensation parameters are burned into the gyroscope by means of multi-sample experiments, and the full variable temperature experiments are carried out under variable temperature conditions (-40-65 ℃, 1 ℃/min) for verification. The experimental results show that under the variable temperature conditions, the zerobias stability of the three fiber optic gyroscope samples is better than 0.005 (°)/h (100 s smoothing), and the compensation effect reaches the expected effect. Due to the variability of the fiber optic gyroscope, the next step is to verify the effectiveness of the algorithm through large-volume experiments.