Computational aspects of continuous-discrete extended Kalman-filtering (Q626232)

The author discusses computational efficient implementation of nonlinear extended continuous-discrete Kalman-Schmidt filtering procedures. A new Taylor-Heun-scheme is proposed for numerical integration of the corresponding differential equations. Its A-stability is demonstrated. A Gauss-Legendre approximation is used for the state error covariance. A technique of adaptive step size control is developed to minimize the computational effort. The algortitm performance is analyzed numerically for the Ornstein-Uhlenbeck process, Van der Poll-oscillator and the vertically falling body problem.