This paper focuses on evaluating the feasibility of using a low cost Micro Electro Mechanical System (MEMS) - based Inertial Measurement Unit (IMU) integrated with Differential Global Positioning System (DGPS), for land vehicle navigation. The main focus is on position accuracy analysis using a closed loop decentralized GPS/INS integration scheme. A Kalman filter is proposed for the IMU which models sensor scale factors and turn on biases. A performance comparative analysis of the MEMS Crista IMU with a tactical grade HG1700 IMU is presented. GPS outages are simulated in a clean data set collected in open sky conditions, by artificially omitting satellites during post-processing in order to asses the stand-alone performance of each IMU. Two different methods are investigated to prevent the errors from accumulating in the Inertial Navigation System (INS) in the absence of GPS. One is to use the constrained motion attributes of a land vehicle to bound the drift in the INS solution. This method is suitable for use in real-time applications. Another method is specifically a post processing scheme which involves the use of a Rauch Tung Streibel (RTS) smoother. Simultaneous field testing using DGPS and MEMS, tactical, and navigation-grade IMUs was conducted to allow for a direct comparison of their relative performance. The results show that in open sky conditions with MEMS IMU measurements and a pseudorange/Doppler derived DGPS solution, a positional accuracy better than 38 cm can be achieved. The use of vehicle movement constraints during GPS outages can prevent errors from accumulating, and can provide the improvement of up-to 80% in the position domain, whereas the corresponding improvements using the RTS smoother are about 99%.
展开▼
