Abstract: A new efficient real- time flight path reconstruction method of the longitudinal aircraft motion is presented in this paper. The reconstruction of flight path is formulated as a discrete, linear, time-variant state reconstruction problem which can be solved by Kalman filtering techniques. The U-D covariance factorization filter, which has excellent numerical characteristics and needs less computation, is used. Furthermore, by exploiting some special features of the system model, the new partitioned U-D factorization filter and the algorithm to make the bias terms controllable are proposed. This method makes the algorithm implementation more efficient and improves the convergence of filter scheme. The simulating and real flight test data computations show that the proposed method in this paper is robust and efficient.

Key words: Kalman filter, U-D covariance factorization filter, flight path reconstruction