An inertial navigation system (INS) exhibits relatively low noise but tends to drift ov time. In contrast, Global Positioning System (GPS) errors are relatively noisy, b exhibit no long-term drift. Integrated INS/GPS navigation systems provide the best both worlds: the low short term noise characteristics of INS and the long term stabi of GPS are combined to provide a navigation solution with accuracy, reliability a robustness far beyond the sum of the constituent parts. However, in order to fu evaluate the performance of an integrated INS/GPS system, it is necessary stimulate both the GPS and inertial components of the system simultaneously. In the paper, the architectures for common filter algorithms in GPS/INS loose integrati are presented. The error dynamics for attitude calculation are derived. Algorithm based on quaternions and direction cosines are used. Simulation results a analyzed to evaluate the different systems. This system evaluation are required help to create specification data; to aid integration algorithm design and tuning; determine if the receiver meets a given specification; to create conditions beyo those which can be created during live trials; and to recreate a known anomaly which occurred in the real world.