The Research Of The Temperature Compensation On FOG Applied In Satellite

Fiber optic gyroscope (FOG) is an angle rate sensor based on the Sagnac effect. It has many advantages, such as high sensitivity, immunity to electromagnetic interference, small size, high environmental stability, all of which make it possible to be applied in the new environment. The research of the FOG system is an active trend in inertial technology research. However, in the experiment or application, it is found that the output of FOG has obvious temperature drift, which is one of the bottlenecks in the engineering application of the FOG. This paper mainly studies the compensation method of FOG temperature bias.This paper introduces the basic principle of FOG, types, the basic structure of interferometric fiber optic gyroscope (IFOG) and the performance evaluation standard of FOG. It also analyzes the main factors influencing FOG temperature bias and briefly introduces how the FOG works in the gesture measurement system and in the navigation system.The method for the compensation of FOG temperature bias can be divided into the hardware method and the software method. The hardware method usually makes the system more complex but has high precision . The software method has lower requirement for system but is limited on high precision, especially for the FOG with high precision.This paper pays more attention to the software method of the compensation for FOG temperature bias with the help of the Markov model. Through the experiment, we get a great amount of experiment data to adjust the input parameters of Markov model. Then this model is simulated in computer. The result of simulation shows that this model can raise the output precision of the FOG to some extent and that the software method is indeed effective. But for the different sorts of FOG, software method maybe needs some adjustment.The embedded solution to the FOG temperature bias compensation is proposed in this paper. We use the TI DSP chip TMS320VC5402 as the CPU to gather and process the data. It is found that this method can greatly increase the real-time speciality and decrease the amount ofcalculation. It has passed the hardware and software debug and circuit test. Also this technology can be applied in the other embedded system for inertial products and multisensor systems.