Português

Abstract

This paper presents a new algorithm for the in-motion alignment of a Strapdown Inertial Navigation System (SNIS), based on a classic method in the literature. The novelty is the introduction of estimates of the repeatability of biases of inertial sensors and roll and pitch attitude angles, based on indirect observations and without the aid of external attitude sensors. The algorithm allows restarting the SNIS without stopping the vehicle or interrupting its operation during navigation, recovering the system's accuracy while it is in motion. In-motion alignment avoids mission delays or temporary exposure of the vehicle in hazardous environments caused by forced stops for alignment after the occurrence of momentary failures, power outages, or shutdowns for preventive maintenance. The algorithm is performed in two steps, coarse and fine alignment, both using Kalman filters to estimate the biases of inertial sensors and the errors of the navigational variables of the system. With the estimation of the repeatability of biases, the SNIS calculates the position, velocity, and attitude data with greater precision at each sampling period. The developed algorithm was validated through simulations of warship navigation.