Performance Comparison Analysis of FOGS and MEMS IMUs Under An Enhanced GPS/Reduced INS Land Vehicles Navigation System

摘要

Most of the land vehicles today are equipped with various motion sensors and, to keep the vehicles affordable, the inertial sensors are usually based on low-cost Micro-Electro-Mechanical System (MEMS) technology. Although MEMS technology is improving, MEMS sensors still have large drift errors which accumulate with time and degrade the performance very quickly. Since Global Positioning System (GPS) has good long term accuracy, it is usually integrated with inertial sensors to limit their error growth. However, in scenarios where GPS satellite signal is partially or fully blocked, there is a limited or no help from GPS and inertial sensors have to coast through those GPS outages. Use of higher end IMUs has the advantage of smaller drift errors but they come at a significantly higher cost which limits their use in lowcost commercial applications. One approach to enhance the performance of MEMs-based integrated navigation systems is to incorporate measurements from other sensors such as speed data. To further decrease the cost and complexity, recent researches have been conducted to use measurements from fewer inertial sensors and this system is called reduced inertial sensor system (RISS). The RISS measurements are integrated with GPS measurements to form an affordable RISS/GPS integrated navigation system for land vehicles. To assess the extent of improvement in navigational solution, there is a need to quantify the performance of MEMS-based IMUs and compare them with higher grades IMUs under different systems configurations. The objective of this work is to demonstrate and compare the performance of high-end FOGs-based IMU and a very low-cost MEMS-based IMU under the configuration of a RISS/GPS integrated navigation system for land vehicles. The comparison is performed under an enhanced EKF design with a new error model that was developed by the authors for RISS/GPS integrated navigation system under the availability of vehicle speed measurements. The enhanced EKF design benefits from the more advanced error model which leads to better estimation of sensors errors and consequently enhances the overall navigation accuracy especially during multipath and long GPS outages. The comparison of the two types of IMUs was conducted on two real road trajectories where GPS signal outages were introduced in post processing and detailed quantitative analysis was performed. To further investigate the error specifications of the two IMUs, Allan Variance analysis was performed which indicated that the dominant error component in both IMUs is the angular random walk. The results showed that in low dynamics, and with de-noising and down-sampling, MEMS and FOGs navigation performance is comparable. Results also demonstrate that, in complete GPS outages, a full 3D IMU/GPS system configuration where speed is utilized as EKF updates, is better than the case of the RISS configuration. Given the huge cost difference and marginal improvement, it is obvious that using the proposed system, many application can benefit from MEMS sensors and avoid excessive cost of higher grade IMUs where speed data is available.