A novel self-alignment method for SINS based on parameter recognition and dual-velocity vectors

Abstract

In this paper, self-alignment problem for strapdown inertial navigation system based on tracing gravity drift in inertial frame and dual-vector attitude determination is studied. Aiming to further ease, the interference from sensor random errors, especially accelerometer random noise, without adding alignment time, a novel alignment method based on integrating gravitational apparent motion to form apparent velocity is designed, using recursive least squares algorithm to recognize the parameters describing the theoretical apparent velocity from the calculated velocity containing random noise. Analysis and simulation indicate that with this method, random noise can be effectively removed compared with the only operation of integration. Meanwhile, a reconstruction algorithm with current recognized parameters for dual-velocity vectors is devised which can fully utilize all measurement information and completely avoid collinear problem. Simulation and turntable results show that compared with the existing integrating methods, the proposed method can acquire sound alignment results with lower standard variances and can improve alignment accuracy with the same alignment time or shorten alignment time with the same accuracy.