Linearized analysis of inertial navigation employing common frame error representations

This paper expands upon previous work that explores a new paradigm for inertial navigation systems. Errors in filter applications using inertial navigation system equations have been previously defined from an abstract vector point-of-view. For example, the error in velocity has always been expressed using a straight difference of the truth minus the estimate without regard to each of the vector’s frame representations. In previous work an alternative vector state-error is defined using common coordinates over all vector error realizations, thereby providing a true-to-life representation of the actual errors. A modified extended Kalman filter was derived that employs the alternative vector state error representation. Here a linearized analysis of the new error equations is conducted to determine effects on INS performance. Schuler and Foucault frequencies are calculated using a stationary analysis.