The fiber-optic gyroscope (FOG), based on the Sagnac effect, is a fully solid-state gyroscope with high reliability, strong anti-interference ability, and rapid startup. It is therefore widely used in inertial navigation. Among them, the interferometric fiber-optic gyroscope (IFOG) has become relatively mature and is widely applied; however, its accuracy is directly limited by the length and diameter of the fiber ring. In contrast, the resonant fiber-optic gyroscope (RFOG) can achieve the same accuracy as an IFOG with a smaller size. Compared to the RFOG driven by a highly coherent light source, the RFOG driven by a broadband light source can effectively suppress noise, such as backscattering and polarization fluctuations, with a simpler optical path structure. This offers greater advantages in terms of cost and size. However, the cost of the LiNbO3 phase modulator in broadband source-driven RFOG remains high, which restricts the reduction of their overall cost. To further reduce the cost, we propose an RFOG scheme driven by a broadband light source, which uses a 3×3 fiber coupler instead of a phase modulator and a balanced photodetector (BPD) to extract the rotation signal.

We establish theoretical principles and conduct a noise analysis on the RFOG driven by a broadband light source based on a 3×3 fiber coupler. We then experimentally measure the gyroscope’s response to rotation, noise levels, and accuracy. First, a mathematical model of this scheme is developed based on the characteristics of the fiber ring resonator (FRR) and the 3×3 fiber coupler, from which the theoretical response characteristics of the gyroscope to rotation are derived. Next, an analysis is conducted on the potential sources of noise within the gyroscope, including electrical noise, shot noise, and relative intensity noise (RIN). Particular emphasis is placed on RIN due to its role as the primary contributor to noise in the RFOG driven by a broadband light source. Experimentally, the resonant curve of the FRR, the core sensitive element of the gyroscope system, is measured, and the fitness of the FRR is obtained. Additionally, the theoretical sensitivity of the gyroscope system is calculated. The RIN of port 1 and port 3 of the 3×3 fiber coupler, as well as the final BPD output signal, is measured and compared with the theoretical analysis to verify the accuracy of the theory. Finally, the gyroscope system is calibrated and kept stationary, and Allan variance analysis is performed on 1 h static test data to evaluate the accuracy and stability of the gyroscope.

In this study, we construct a broadband source-driven RFOG based on a 3×3 fiber coupler, where the FRR is made of polarization-maintaining fiber with a length of 104 m and a diameter of 7.6 cm. The fitness of the FRR is experimentally measured to be 48, from which the ARW constrained by shot noise is calculated to be 0.0016 (°)/h^1/2. Utilizing the established mathematical model, the response curve of the gyroscope to rotation is simulated, and the response characteristics are experimentally tested, with the results aligning with the simulation. The experiment demonstrates that when the rotational angular velocity is within ±30 (°)/s, the gyroscope exhibits a linear response. Additionally, the RIN in the system is simulated, which indicates that in this scheme, the larger the correlation coefficient between the optical signals at port 1 and port 3 of the 3×3 fiber coupler, the more effectively the RIN can be canceled, thereby reducing system noise. The experimental RIN measurement results also verify the simulation results. The correlation coefficient between the optical signals at port 1 and port 3 of the 3×3 fiber coupler is measured to be 0.7435. Theoretically, the peak of the RIN spectrum after subtracting the optical signals at the two ports is 4.7 dB lower than that before subtraction. However, the presence of two photodetection links results in the shot noise and electrical noise of the signal after subtraction being marginally higher than those before subtraction. In the static test, the gyroscope system achieves an angular random walk (ARW) of 0.047 (°)/h^1/2 and a bias instability (BI) of 0.114 (°)/h, which indicates that the gyroscope attains tactical-level accuracy. However, due to the absence of modulation in this scheme, it remains challenging to effectively suppress optical noise and drift in the system.

We have demonstrated, through theoretical analysis and experimental verification, the feasibility of using a 3×3 fiber coupler as an alternative to a LiNbO₃ phase modulator in the construction of a broadband source-driven RFOG, with the objective of extracting the rotation signal using a BPD. The suppression of RIN has been confirmed through both theoretical investigation and experimental verification. Ultimately, tactical-level accuracy is achieved with a 104 m long FRR. Compared to the traditional broadband source-driven RFOG, the broadband source-driven RFOG based on the 3×3 fiber coupler is more cost-effective, with a simpler optical path structure and signal processing method. Additionally, compared to the LiNbO₃ phase modulator, the passive 3×3 coupler is easier to integrate with the light source and photodetector (PD). Therefore, this scheme offers a novel approach to the cost-effective miniaturization and integration of RFOG.