Practical combination forecasting evaluation method for PINS comprehensive performance

The long-term reproducibility of platform inertial navigation system (PINS) is a key parameter for evaluating whether the comprehensive performance of PINS has met the navigation accuracy. For the PINS dynamic flight, the PINS navigation error model can accurately calculate the comprehensive performance of PINS according to its individual performance parameters, but this method takes too much time and can't meet the timeliness requirements of the rapid evaluation during decision-making stage in PINS application. Therefore, we proposed a PINS performance prediction method based on the PINS navigation error model and the least squares support vector regression (LS-SVR). By combining the navigation error model with LSSVR combination method, we can build the PINS performance prediction model in the storage preparation stage of PINS, and real-time predict the individual performance parameters in the decision-making stage of PINS. One can quickly get the comprehensive performance evaluation results by sufficient comprehensive performance prediction of each PINS using the PINS performance prediction model and the comprehensive performance evaluation of each PINS using the evaluation indicators of the PINS comprehensive performance. The calculation results verify that the combination forecasting method not only has high prediction accuracy, but also has effectively improved the prediction efficiency and evaluation quality of the PINS comprehensive performance. It has a good practical application and potential in evaluation and decision-making.