INS/GPS Navigation Accuracy Improvement With Nonholonomic Constraints

Abstract

This paper describes an INS/GPS integration algorithm with nonholonomic constraints for improving navigation accuracy during periods of GPS signal outage. The constraints that the velocity of the vehicle in the plane perpendicular to the forward direction is almost zero are used as a new measurement of the Kalman filter for the navigation. By using the relationships between velocity equations and sensor noises, theoretical analysis of the new measurement is performed to decide the most appropriate measurement noise. The proposed method is tested for a land vehicle using two GPS receivers and a tactical grade INS. The experimental results show that the nonholonomic constraints provide a reasonable solution to the problem of GPS signal outage.