Research And Realization On The Attitude Algorithm In Gyroscope-free Strap-down Inertial Navigation System

Inertial technology is one of the critical technologies in the orientation and navigation of weaponry system. Strap-down Inertial Navigation algorithm is the key part of SINS. For a certain extent the orientation and navigation of Precision Guided Munitions depend on the assistant analyze of attitude parameter. The attitude arithmetic and navigation is in the core part of SINS. And what's more, it affects the precision mainly. Especially to the missile in highly dynamic situation, precise attitude measurement is a crucial factor that can determine if the system runs normally. So it is necessary to study the high precision attitude algorithms for GFSINS. Under a fund project, this thesis focused on promoting attitude precision in a new flight parameter measurement system which operates in high spin condition and has a miniature size.For the particularity of high spin attitude measurement, the design faces two major difficulties: choosing the Accelerometer configuration mode and calculating precise angular velocity. The following works have been done: (1) According to the actual space installation requirements and reducing navigation error accumulated over time. A twelve-accelerometer structure was designed providing all the attitude information. On the other hand, a ten-accelerometer and magnetic structure was designed for reducing system space in high-speed rotary missile and improving practicability. (2) The mathematical model of angular speed is established with fix- errors. The method is using the damping Gauss-Newton method of nonlinear least squares for compensating and amending the fix-errors of accelerometers. (3) Nonlinear attitude is identified by BP neural networks based on LM algorithm. Selecting the test samples and setting up the network model between the information of accelerometers and the real-time angular velocity after studying the effect of accelerometer output factors for attitude. The simulation results show that the project is feasible. After designing calibration algorithm, missile simulation and calibration experiments for tested parameters and fix-errors were undertaken. To providing reliable information for optimized arithmetic and neural network model. Finally, attitude algorithm and error forecast validate that the arithmetic and model are reliable and applicable. It is valuable and necessary for the further study.