Abstract:As a complex system combined with optics, electronics and mechanics, the RLG (ring laser gyroscope) can measure the angle rate with high accuracy. In order to meet the requirement of longer-time and higher precision in inertial navigation system, the characteristics between the RLG drift error and different signals from various sensors were studied. Based on the traditional RLG drift error compensation model, the compound signals from thermometers, photodiodes and piezoelectric ceramics attached on the dithered mechanism were introduced into the model. Moreover, by utilizing the support vector machine algorithm which has good nonlinear fitting capacity, the model was optimized by the correlation features between the mechanically dithered RLG drift error and the compound signals. The experimental results show that the precision of the mechanically dithered RLG drift error compensation model is higher than the traditional one.