Modeling and analysis of gyroscope's random drift based on GP+GA method

Gyroscope is the key component in an inertial navigation system (INS). It depends on the precision of the INS. In order to improve the gyro's operational accuracy and compensate the random error caused by the nonlinear and weak-stability characteristic of gyro's random drift, the nonlinear random drift error model based on genetic programming (GP)+genetic algorithm (GA) was established taking the time series of gyro's random drift as study object. Firstly, the time series of gyro's random drift were got through analyzing and preprocessing the measured data of gyro. Then the model from the data based on the GP was established and the nonlinear mathematic relationship between the current time and the former times was obtained. Finally, GA was used to optimize the parameters of the mathematic relationship in order to get the more accurate model. The experiment result indicates that compared with classical auto regressive (AR) model, this model can effectively reflect the characteristics of gyro's random drift. The square error of the gyro's random drift has decreased by 73.72% and the effects have increased by 4.72% compared with classical AR model. The establishing model method effectively compensates the gyro's random drift and improved the stability of the system.