Alignment and calibration method for intelligentized aircraft missile movable base

Abstract

This invention relates to an intelligent airborne missile moving base alignment and calibration method. Based on the time variation property of Strapdown Inertial Navigation System model under the moving base condition, a neural network input and output sample construction method which is adequate for the alignment and calibration of moving base is provided. In order to settle the question that when hanging beneath the aircraft wing, the inertial navigation subsystem measuring information of the missile is not the same as when flying freely, the error angle between inertial navigation system and subsystem is firstly estimated and compensated before neural network training. Thus, the neural network training sample can simulate the freely flying state of missile accurately. This invention can increase the navigating accuracy.

Description

technical field [0001] The invention relates to an intelligent airborne missile dynamic base alignment and calibration method, which can be used to improve the navigation accuracy of airborne, ship-borne and vehicle-mounted missile weapons. Background technique [0002] The initial alignment error is one of the main error sources of the inertial navigation system (INS). The initial alignment accuracy and speed are directly related to the strike accuracy and rapid response capability of the weapon system. Usually, Kalman filtering is used to solve the initial alignment problem of INS, which can estimate the state value of the system from the noise-contaminated observations. However, since the operation time of the Kalman filter is proportional to the cube of the system order, it is difficult to guarantee the real-time performance of the filter when the system order is high. Although the real-time performance of filtering can be improved by using distributed Kalman filter, it...

AI Technical Summary

Problems solved by technology

At this time, if the measured values ​​and filtering results of the Kalman filter are directly used as input and output samples to train the neural network, since the samples used in the training of the missile network are inconsistent with the samples used in the actual application of the missile self-electric flight, the static basis The network constructed and trained under the base alignment and calibration conditions will not be applicable to the moving base case

Images