Abstract:

Aiming at the remarkable influence of wing flexure noise on transfer alignment (TA) filtering, an online compensation algorithm is proposed on the assumption that the wing flexure noise is a time varying multi-order colored measurement noise from its character analysis. On the basis of typical Kalman filtering equations, new filtering equations capable of compensating multi-order colored noise without dimension extension are derived. According to the series measuring data and aided by a model estimating error F testing method, the algorithm is used to iteratively estimate auto-regressive orders, auto-regressive coefficients and white noise of wing flexure noise online which are thus taken advantage of to construct the measurement data, measurement matrix and measurement noise to perform compensating filtering. Finally, the algorithm is applied to a velocity plus attitude TA matching plan and compared with the method substituting wing flexure noise by pouring white measurement noise and extending system states. Simulation results show that the algorithm is of strong adaptability to new measurement data under time varying wing flexure noise as well as steady convergence and high accuracy estimating performance, which is feasible to achieve rapid and high accuracy airborne missile TA under hard air environment.