• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.1
• /
• pp.62-67
• /
• 1998

Gimballed Inertial Navigation Systems(GINS) are sophisticated autonomous electro-mechanical systems which supply the position, velocity and attitude of the vehicle on which they are mounted. In order to maintain accuracy of outputs, the GINS are required to regularly calibrate senior errors. However, existing calibration methods take up a long time due to multiposition alignments needed to increase accuracy. A particular system formulation for calibration of a GINS is proposed to enhance system observability and thus to expedite calibration procedure. Performance of the proposed calibration method is compared with existing methods such as Schuler test and muliposition alignment. Simulation studies show the proposed system formulation associated with a suggested suboptimal filter is accurate as well as efficient in error identification essential to GINS calibration.

• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.48 no.6
• /
• pp.735-746
• /
• 1999

We design the robust and inherently fault tolerant decetralized$$H^infty$$ filter for the multisensor state estimation problem when there are insufficient priori informations on the statistical properties of external disturbances. For developing the proposed algorithm, an alternative form of suboptimal$$H^infty$$ filter equations are formulated by applying an alternative form of Kalman filter equations to the indefinite inner product space state model of suboptimal$$H^infty$$ filtering problems. The decentralized$$H^infty$$ filter that consists of local and central fusion filters can be designed effciently using the proposed alternative$$H^infty$$ filiter gain equations. The proposed decentralized$$H^infty$$ filter is robust against un-known external disturbances since it bounds the maximum energy gain from the external disturbances to the estimation errors under the prescribed level$$r^2$$ in both local and central fusion filters and is also fault tolerant due to its inherent redundancy. In addition, the central fusion equations between the global and local data can reduce the unnecessary calculation burden effectively. Computer simulations are made to ceritfy the robustness and fault tolerance of the proposed algorithm.