A Quick Forecasting Evaluation Method for the Comprehensive Performance of PINS Based on the Long-Term Reproducibility

The long-term reproducibility of the platform inertial navigation system (PINS) is key parameter to evaluate whether the comprehensive performance of PINS has met the navigation accuracy. Under the conditions of given standard trajectory parameters, PINS navigation error model can accurately describe the propagation relationship between each individual performance parameter of the PINS and comprehensive performance, but the model calculation consumes a long time, which can´t meet the demand of PINS application and decision, for which a comprehensive PINS prediction method is proposed combined PINS navigation error model with least squares support vector machine regression, and the PINS overall performance evaluation function is given. Using initial alignment errors and long-term repeatability parameters which characterize the overall performance of the PINS to test the combination forecasting method, the results show that this method has high prediction accuracy and pretty predicting aging, the overall performance of the PINS can be quickly and accurately judged by using the evaluation function, results a good practical application value.