INS/DGPS/VMS Integration for In-Motion Alignment

In this paper, we propose a new method for a land-vehicle In-Motion Alignment by integrating Inertial Navigation Sys- tem, the transfer (code-based) Differential GPS and Vehicle Motion Sensor (INS/DGPS/VMS). The VMS can provide moderately accurate speed of vehicle from measured wheel revolution by an optical encoder. It is similar to the so-called odometer. Our aim in this work is to integrate the advantages of these systems and to develop the navigation system that does not require initial attitudes information of the vehicle° In case of our previous conventional navigation system, the initialization of INS navigation states is completed prior to vehicle motion. By stopping the vehicle at the start point, these initial states can be computed by integrating INS data with static navigation data such as velocity 0 (ft/s) and ini- tial position obtained from DGPS. However this initializa- tion method usually requires 5 - 10 minutes and thus the vehicle must be stopped. However it is common occurrence that there is not enough time to stop at the start point. Thus, for reducing the initial alignment time, developing the In- Motion Alignment algorithm is desired. In case of In-Motion Alignment, there exist some similar works such as INS/GPS integrated system. In these works, it is assumed that the external information, i.e. GPS navi- gation data can be obtained continuously. In other words, if the GPS signal is not available because of obstructions such as tall buildings in the city, they cannot be performed. Therefore we propose a new integrated INS/DGPS/VMS In- Motion Alignment by applying the Kalman filter. Substi- tuting the VMS for DGPS as DGPS is not utilized, the al- gorithm is able to continue In-Motion Alignment in all cir- cumstances. In this paper, we consider In-Motion Alignment by using DGPS and VMS properly. When the DGPS is available, accurate velocity and position data can be obtained and they are used as the measurement to estimate the INS er- rors by the Kalman filter (INS/DGPS mode). On the other hand, VMS sensor provides only speed data. Therefore if the DGPS signals are not available, VMS speed is used as the measurement corresponding to DGPS (INS/VMS mode). For INS error model of the algorithm, the large azimuth er- ror model which formulates INS error (position, velocity, attitude) equations for large initial heading error and sensor error [ 1] is adopted with some suitable modifications for our coordinate system• The experimental results show that, in case of In-motion alignment, although GPS navigation data are utilized dis- continuously, our INS/DGPS/VMS integration switching mode provides almost the same accurate navigation data (position, velocity, attitude) comparing with INS/DGPS mode under GPS navigation data obtained continuously.