Embedded Vehicle Dynamics and LASER Aiding Techniques for Inertial Navigation Systems

Abstract

This work evaluates the impact of two state-of-the-art aiding techniques to enhance the performance of inertial navigation systems (INS). A new embedded methodology to integrate the vehicle dynamics (VD) in the navigation system is proposed, by modeling it directly in the Extended Kalman Filter. The embedded VD and the INS algorithm propagate simultaneously the inertial states, allowing for the estimation of the INS errors by exploiting the dynamical information enclosed in the vehicle model. Results show that the attitude, velocity and inertial sensors bias estimates are enhanced by the comprehensive number of states predicted by the VD. The proposed technique introduces computational savings, with an accuracy equivalent to the classical external vehicle model implementations. A LASER range finder sensor is also introduced as an external aiding source and integrated into the navigation system to provide high precision altitude readings for the critical takeoff and landing maneuvers. The paper shows that the proposed technique represents a step towards the use of Uninhabited Air Vehicles in mission scenarios with limited GPS availability and/or high accuracy positioning requirements. The performance of the INS aiding architecture is assessed in simulation, and results obtained with the full nonlinear dynamics of a model-scale helicopter are presented and discussed.