ECOA-MAML-Enhanced Framework for Accurate Temperature Drift Compensation in Fiber Optic Gyroscopes

A temperature drift compensation method is proposed, integrating the Enhanced Coati Optimization Algorithm (ECOA) with a meta-learning framework to improve the measurement accuracy and adaptability of Fiber Optic Gyroscopes (FOGs). Experimental data are preprocessed using FFT and low-pass filtering, effectively removing high-frequency noise while preserving key low-frequency drift information. Experimental results confirm that ECOA significantly outperforms COA in both compensation performance and optimization efficiency. MSE is reduced by 90.63% with a 10.67-fold improvement, while RMSE decreases by 68.26% with a 3.15-fold improvement. ECOA achieves rapid convergence within 50 iterations, compared to COA, which remains unstable beyond 100 iterations. Initial fitness values for ECOA (0.96009–0.99632) are much closer to the optimal solution than those of COA (2.9637–4.2647), and final fitness values for ECOA (0.059477–0.068461) are 3 to 5 times smaller than COA’s (0.18606–0.29207). Moreover, ECOA exhibits smaller fitness fluctuations, indicating greater stability and robustness. These results demonstrate ECOA’s ability to minimize ineffective search paths, enhance accuracy, and achieve efficient optimization, making it a reliable solution for high-precision temperature drift compensation in FOGs.