Design and analysis of inertial navigation systems with skew redundant inertial sensors

The performance of current MEMS based inertial units is usually insufficient to be used in applications for which high accuracy is required. For such applications, if the size and power constraints do not allow the use of traditional high-end inertial sensors, an alternative solution is strongly required. In this study such an alternative INS design method using redundant low cost inertial sensors is described.;In the second part, the optimum navigation solution for the SRIMU based INSs is defined. In general, such optimum configurations are hard to implement due to the structure of the navigation filters. Therefore, in this study, alternative suboptimal SRIMU based INS configurations are introduced and the conditions for which these alternative configurations become optimal are described. These configurations can be conveniently implemented in the existing navigation software/libraries without requiring any significant modification.;Finally, the comparison of SRIMU based single-INS configurations with the multi-INS configurations is presented. In the existing literature, the multi inertial sensor navigation problems are usually handled within multi-INS approaches. However, such multi-INS approaches i) are numerically very problematic, ii) require extensive modification in the existing navigation libraries and iii) cannot be used when sensors are arbitrarily oriented. In this study it is shown that the proposed single-INS configurations for SRIMU systems are theoretically equivalent to the multi-INS configurations and hence can be implemented instead of any multi-INS configuration without dealing with any of the aforementioned problems.;In the first part of this study the solution of the optimum inertial sensor fusion problem is derived. In the skew redundant inertial measurement units (SRIMU), the redundancy of the sensors generates additional observations which can be used to estimate the sensor errors without requiring any external aid. In this study, it is proved that the projection of raw sensor outputs to the left null space of the sensor configuration matrix can be used to define these redundancy observations. Based on these observations, the computation of best acceleration and rotation rate which can be used to execute the standard INS equations in an SRIMU based navigation system is explained.