Inertial measurement units (IMU)s have been in use for decades. When first developed the uses were usually limited to space travel and military applications. With the development of new technologies and new sensors IMUs are being used in many more applications like street mapping, airborne surveying, down-hole logging and directional drilling, unmanned vehicles, and even in smart phones. The variety of applications has lent itself to a wide variety of sensors being developed with costs ranging from a few dollars to few thousand dollars. While the development of these sensors and IMUs has led to more widespread uses and a general improvement in the performance to cost ratio, there has remained a gap in understanding of the effects of vibration on the accuracy of the system. The result of this gap is poor performance under vibration. Both the designer and user must be made aware of the vibration related errors to avoid them.
This thesis explores the effects of nonlinearity, system noise, computational error, filtering, and system dynamics with the goal of providing an understanding of the errors and their interactions when subjected to vibration or high dynamics. By providing methods of analyzing and in many cases compensating for errors, it is possible to design a more robust system.