Application of Extended Kalman Filtering to a Dynamic Laboratory Calibration of an Inertial Navigation System

摘要

The report describes a data reduction technique that obtains estimates of inertial sensor error model coefficients from a dynamic laboratory calibration of a typical Inertial Navigation System. The error model coefficients are those associated with gyros, accelerometers, and their misalignment errors that have been found by test and analysis to be the predominant sources of error affecting system accuracy. All the error terms considered are categorized as either fixed (independent of applied acceleration), first-order (proportional to the first power of acceleration), or higher-order terms, which are proportional to the square or cube of acceleration. In the case of the higher-order terms, the error model coefficients of inertial grade sensors are from one to four orders of magnitude smaller than the fixed and first-order terms. (Modified author abstract)