导弹自适应自动驾驶仪的设计

本文对基于辨识和波波夫超稳定性理论的自适应自动驾驶仪设计进行了研究。这里研究的对象是导弹横滚回路,可用二阶时变微分方程近似描述。由于弹体参数在大范围内激烈时变,我们选择了具有遗忘因子的递推最小二乘法,实现实时地辨识对象参数,并且采用了比例加积分型模型参考自适应规律,使对象输出较好地跟踪参考模型输出。计算机仿真的研究给出了两种不同输入信号时的参数辨识的结果,仿真还表明了采用估计的参数进行自适应调参与采用真值实现自适应调参,其响应是一致的。

Design of Adaptive Autopilot for Tactical Missiles

Adaptive control technology has been developed significantly over the lastten years for tactical weapon applications, A successful analog adaptive autopilotapplied to air-to-air missile design by American scientists was reported in 1977[1]. They selected a simple method, so called external sinusoidal signal exci-tation method, to implement the on-line parameter identification of the plant.The disadvantage is that the adaptive range is limited. Since then, many re-searchers have studied the digital autopilot, which was made possible throughminiaturization, availability, flexibility, and low cost of digital hardware[2][3], but, to the author's best knowledge, no papers have been reported inthe open literature on the design of the digital adaptive autopilot of tacticalmissiles. The aim of this paper is to study the possibility of designing digital adap-tive autopilot of tactical missiles, using model reference adaptive control(MRAC) with explicit identification. The plant considered is the roll loop ofthe air-to-air missile whose dynamic parameters vary within wide limits. Therecursive least square (RLS) method, with a forgetting factor for the secondorder plant, is selected to implement the on-line parameter identification, anda P+I adaptive law is used to make the plant output follow the referencemodel output very closely. All the differential equations are transformed into difference equationsusing backward difference method in order to avoid the trouble caused by theinherent one-step sampling lag. Some practical problems encounted in the designare discussed. The parameter identifications for two kinds of inputs are given.It is shown that the results of using MRAC with real-time identification are inagreement with those of using MRAC with real values of the plant parameters.