Rapid Initialization of SINS for Land Vehicle- based Vertically-launched Missiles

This paper describes a new method for rapid initialization of a strapdown initial navigation system (SINS) which is hosted on a missile intended for vertical launch from a land vehicle. The SINS initialization aims to determine initial values of the system, including position, velocity, and attitude. For a fixed host platform, the attitude is the most critical aspect of the initialization. There are two alignment procedures. The first is the analytical alignment, in which the attitude matrix is computed directly utilizing the output of an inertial measurement unit (IMU) in combination with known earth rotation and gravity. The second procedure, usually referred to as the precise alignment, utilizes rough-wrought output from the analytical alignment to estimate precise initial attitude using a Kalman filter. Commonly used methods for precise alignment include the single-position alignment, the two-position alignment and the multi-position alignment. The latter two methods are widely used on fixed platforms because of good observability of the system. The two-position alignment introduces the second position alignment by rotating the yaw angle by180 . The multi-position alignment is implemented by further rotating yaw or pitch based on the rotation of the twoposition alignment. For a missile which is verticallylaunched from a land vehicle, either the multi-position or two-position alignment reduces the effectiveness and flexibility of operations as both alignments require a machine to implement the yaw rotation. This paper presents a new two-position self-alignment method, which introduces a rotation of pitch by 90 , to naturally reflect the raising to an upright position of the while on a fixed land vehicle. Distinguishing itself from the commonly-used two-position alignment or multiposition alignment, the new method requires the rotation to the second position after the missile remains at the first position for 30s until the horizontal misalignment angles become stable. In computer simulations, the effect of wind disturbance is considered. The results confirm that the proposed method can reduce the alignment period and achieve a higher precision of solution than those of commonly used methods.