Research On Techniques Of Modeling And Compensating For Temperature Drift Error In Fiber Optic Gyroscope

Fiber optic gyroscope(FOG) is a new type of angle-velocity transducer, with the merits of wide dynamical range, small volume, light in weight, and so on. But when the ambient temperature changes, the bias has reduced the gyro's precision, which has been seriously affected its application in engineering.In recent years, foreign and domestic experts study a lot on temperature drift of fiber-optic gyroscopes, which is mainly divided into three dimensions, e.g., improvement of the structure of FOG, compensating for temperature-induced drift through hardware measures and software model construction. Because of the limitations of the technology conditions and elements, it is difficult to resolve the problem of temperature-induced drift in ways of changes of structure and mechanism. As a result, the third approach has received a lot of attention.This dissertation studies the technique of modeling and making up for temperature-induced drift in FOG from the perspective of software compensating.At first, the fundamental principles of FOG are given. The effects of temperature on the performance of FOG are discussed later, and then the main ways of curbing the temperature-induced drift of the Fiber optic gyroscope (FOG) are shown and compared. And all this paves the foundation of method of software compensating.Secondly, through the measurement of signals of Medium accuracy gyroscope, we study the factors leading to the changes of temperature. What is more, we analyze the temperature, the rate of change of temperature, gradient of temperature qualitatively and quantitatively. Based on the method of polynomial fitting, we apply the approach of split modeling and compensating. The model proves to be effective considering the results of simulation.Moreover, we use the property of nonlinear mapping of neural network and apply BP neural network to model and compensate for the drifts of temperature. Through comparison, we find that this method is better than the method of polynomial fitting.Finally, by combining the above two methods, we design the unified approach. For the cases of constant temperature and varying temperature, take the rate of the change of temperature, we set conditions of justification, and show the unified framework.Through simulation, the drifts of temperature can be reduced by 40%, that is to say, within certain ranges of the rate of changes in temperature, this unified method is effective.