Abstract: In order to solve the difficulties to establish the mathematical model of tow target and the influence of height measurement error, active disturbances rejection control method based on neural network is proposed. The influence of mathematical model inaccuracy to the flight is reduced by design of active disturbances rejection control. With height and vertical acceleration as the input signals to neural network, the compensation for disturbance rejection control factor coefficient is got, which decreases the influence of height measurement error, and makes the compensation of controllor be corrected in real time. Simulation indicats that this method makes the flight performance of the tow target more accurate, effectively restrains the external interference ,and accomplishes goal of a constant high flying.

Key words: Tow target, Neural network, Active disturbances rejection control, Simulation