Search

The well-known conventional Kalman filter requires an accurate system model and exact stochastic information. But in a number of situations, the system model has an unknown bias, which may degrade the performance of the Kalman filter or may cause the filter to diverge. The effect of the unknown bias may be more pronounced on the extended Kalman filter (EKF), which is a nonlinear filter. The two-stage extended Kalman filter (TEKF) with respect to this problem has been receiving considerable attention for a long time. Recently, the optimal two-stage Kalman filter (TKF) for linear stochastic systems with a constant bias or a random bias has been proposed by several researchers. A TEKF can also be similarly derived as the optimal TKF. In the case of a random bias, the TEKF assumes that the information of a random bias is known. But the information of a random bias is unknown or partially known in general. To solve this problem, this paper proposes an adaptive two-stage extended Kalman filter (ATEKF) using an adaptive fading EKF. To verify the performance of the proposed ATEKF, the ATEKF is applied to the INS-GPS (inertial navigation system-Global Positioning System) loosely coupled system with an unknown fault bias. The proposed ATEKF tracked/estimated the unknown bias effectively although the information about the random bias was unknown.

A nonlinear observer is proposed for a strapdown inertial navigation system (SDINS) in-flight alignment problem using an [H.sub.[infinity]] filter Riccati equation and a freedom parameter. The proposed observer improves the filtering stability, convergence, and performance. The characteristics of the observer are analyzed using a Lyapunov function. Simulation results demonstrate a significant reduction in alignment errors by employing the proposed nonlinear observer. The observer is developed in general such that it can be applied to estimating nonlinear systems other than the SDINS in-flight alignment.

We propose new equivalent tilt error models which are applicable to the analysis of the terrestrial strapdown inertial navigation systems (SDINS), based on the quaternions. The currently available equivalent tilt error models, like the conventional [Phi] model of the gimbaled inertial navigation systems (GINS), are derived only by the true frame approach. The true frame approach has a computational disadvantage that it produces an error model where the attitude error equation is coupled with its position and velocity error equations. The motivation of this work is to solve this problem. As a result, two kinds of error models are derived. Among them, one is derived by the computer frame approach for practical onboard implementations. Thus, like the conventional GINS [Psi] model, its attitude error equation is decoupled from the position and velocity error equations. The other is derived in order to show the relationship between the true frame approach and the computer frame approach which are applied to the quaternion-based SDINS. Thus, like the GINS [Delta][Theta] model, it can be used to transform the error variables into each other which are calculated by the two different approaches.

The relationship between quaternion errors and tilt angles from true navigation frame to analytic platform frame is newly derived for strapdown inertial navigation system (SDINS). Using the relationship it is shown that the quaternion error model for attitude and velocity is equivalent to the conventional perturbation error model. Based upon the equivalency, the quaternion errors during a multiposition alignment are determined and analyzed. Furthermore, it is shown that the heading rotation of 180 deg achieves the minimal quaternion error for 2-position ground alignment.

Presented in this paper is a complete error covariance analysis for strapdown inertial navigation system (SDINS). We have found that in SDINS the cross-coupling terms in gyrocompass alignment errors can significantly influence the SDINS error propagation. Initial heading error has a close correlation with the east component of gyro bias error, while initial level tilt errors are closely related to accelerometer bias errors. In addition, pseudostate variables are introduced in covariance analysis for SDINS utilizing the characteristics of gyrocompass alignment errors. This approach simplifies the covariance analysis because it makes the initial error covariance matrix to a diagonal form. Thus a real implementation becomes easier. The approach is conformed by comparing the results for a simplified case with the covariance analysis obtained from the conventional SDINS error model.

This paper presents a new approach to noise covariances estimation for a linear, time-invariant, stochastic system with constant but unknown bias states. The system is supposed to satisfy controllable/observable conditions without bias states. Based on a restructured data representation, the covariance of a new variable that consists of measurement vectors is expressed as a linear combination of unknown parameters. Noise covariances are then estimated by employing a recursive least-squares algorithm. The proposed method requires no a priori estimates of noise covariances, provides consistent estimates, and can also be applied when the relationship between bias states and other states is unknown. The method has been applied to strapdown inertial navigation system (SDINS) initial alignment. Simulation results indicate a satisfactory performance of the proposed method.