Research on Temperature Error Compensation for Fiber Optic Gyroscope Based on LSH-PSO

To address the issue that the output of fiber optic gyroscopes (FOGs) is susceptible to ambient temperature, a temperature compensation method based on locality-sensitive hashing-enhanced particle swarm optimization (LSH-PSO) is proposed. This method introduces both temperature and its change rate into the model, which characterizes the error mechanism, and over which the LSH-PSO is utilized to identify the involved model parameters for optimization. Experiments were designed with three sets of temperature-changing scenarios: heating-insulation, unidirectional continuous heating, and full-range reciprocating temperature variation. The compensation effectiveness of the model fixed by particle swarm optimization (PSO), PSO-SVR, GWO-GRU, and LSH-PSO algorithms was compared and analyzed. Experimental results indicate that with the LSH-PSO-based compensation model, the FOG's bias stability and random walk coefficient reached 0.02°/h and 0.0091°/√h, respectively, representing improvements of 90% and 64% compared to those without error compensation. The proposed method effectively suppresses bias drift under complex temperature conditions, offering a solution with potential for engineering applications in FOG temperature compensation.