Improving MEMS IMU/GPS Systems for Accurate Land-Based Navigation Applications

To overcome the problems encountered with GPS stand-alone navigation, the concept of integrating other navigation sensors with GPS has been implemented. For the last two decades, the integration of an Inertial Navigation System (INS) with GPS has become a standard tool for accurate navigation applications. In INS/GPS integrated systems, GPS is used for positioning and calibrating the INS time-dependent errors while the INS is used for determining attitude and providing navigation information when GPS signals are blocked. For accurate INS/GPS navigation, navigation-grade and tactical-grade Inertial Measurements Units (IMUs) are commonly used. However, the high cost of these IMUs ($15,000-100,000), their considerable size and their restricted access regulations have made several limitations on using them extensively. Thus, the navigation community has been directed to investigate other inertial sensor possibilities. Recently, this has been accomplished by the evolving development of low-cost inertial sensors based on Micro-Electro-Mechanical Systems (MEMS) sensors. However, due to the fabrication process, MEMS inertial sensors have large bias instability and sensor noise. Thus, special considerations have to be taken into account for MEMS IMU/GPS integration. In this paper, the possibility of using MEMS IMU/GPS integrated systems for accurate land-based navigation applications is investigated through improving the integrated system accuracy using two successive approaches. The first approach is to improve the IMU/GPS Kalman filtered solution by applying additional velocity updates for the MEMS IMU that can be used during GPS full availability and during GPS signal blockage periods. Then, in the second approach, the improved solution is further enhanced through backward smoothing. To perform the analysis, a land-vehicle MEMS IMU/GPS data is used with several induced GPS signal blockage periods. Compared to the original obtained solution during GPS signal outage intervals, the results showed a remarkable accuracy improvement after applying the proposed approaches where the final obtained navigation accuracy met the requirements of several accurate land-based navigation applications.