INS/VNS/CNS integrated navigation method for planetary rovers

Abstract

Planetary rovers play an important role in deep space exploration missions. Inertial navigation system (INS), visual navigation system (VNS) and celestial navigation system (CNS) are three commonly used autonomous navigation systems for rovers. Although INS and VNS have high precision, their navigation errors especially the yaw angle error accumulate over time. CNS can make up for this shortage and provide accurate yaw angle information. Since INS, VNS and CNS have complementary characteristics, an integrated navigation system composed of them helps to improve the navigation accuracy and reliability. In this paper, an INS/VNS/CNS integrated navigation method for planetary rovers is presented. The INS error equation expressed in the navigation frame is used as state model. The relative motion error obtained by INS and VNS and the starlight vector provided by CNS are used as measurements. The unscented Kalman filter (UKF) is utilized to perform state estimation. Hardware-in-the-loop simulations show that the proposed method can remarkably improve the estimation accuracy of position and attitude compared with the INS, VNS, INS/VNS and INS/CNS navigation methods.