Gyroscope Drift Estimation in Tightly-coupled INS/GPS Navigation System

The objective of this paper is to analyze the estimation performance of gyroscope drifts in INS/GPS integrated navigation systems, while system is in motion. Strapdown inertial navigation system (SINS) error model is augmented by inertial sensor bias and drift and GPS clock bias and drift to improve the integration performance. To estimate the augmented state error, instead of GPS position, GPS pseudoranges and pseudorange rates are used. In tightly-coupled integration GPS and INS under different motion conditions, it is observed that the vertical gyro drift estimation is very weak in all cases of stationary, linear motion, and three-axis sway motion via Kalman filter covariance analysis. In order to solve the vertical gyroscope drift estimation problem, zero update velocity (ZUPT) algorithm is proposed to improve the observability of vertical gyro drift. Computer simulation is carried out by this algorithm to validate the feasibility of the concept. The simulation results demonstrate that vertical gyro drift can be better estimated for the sake of compensation afterward.