快速倾斜镜是星间激光通信终端精瞄系统的核心部件，其驱动装置为压电陶瓷执行器，而压电陶瓷具有迟滞特性，其严重影响了快速倾斜镜的定位精度，进而对星间通信链路的稳定性造成不利影响。为解决这一问题，本文设计了一种改进Prandtl-Ishlinskii(P-I)模型对压电陶瓷执行器进行建模。在此基础上，提出了压电陶瓷执行器前馈线性化方法，以对迟滞特性进行前馈逆补偿。接着，提出了一种结合改进的 P-I模型与增量式PID算法的复合控制算法，并在DSP中实现了该复合控制算法。最后，在试验平台上对该算法进行了验证。结果显示: 当分别对系统输入10Hz和100Hz减幅正弦、等幅正弦曲线时，模型误差在0.59%以内，在输入同频100 Hz以下的减幅正弦曲线时，传统PID算法的最大误差为59.31 μrad，而该复合算法的最大误差为14.22 μrad。实验数据表明，本文复合控制方法的动态跟踪性能明显优于传统PID方法，改进Prandtl-Ishlinskii(P-I)模型可以精确描述压电陶瓷的迟滞特性。本文设计的复合控制方法满足实际应用对快速倾斜镜的要求。

Fast Steering Mirror (FSM) is the core part in terminal fine aiming system of inter-satellite laser communication. As the driving device for FSM, Piezoelectric actuator’s (PEA) hysteresis characteristics seriously affect FSM’s position accuracy, and therefore have adverse influence on stability of inter-satellite communication link. In order to solve this problem, a modified Prandtl-Ishlinskii (MPI) model was proposed to describe PEA. On this basis, a feed-forward linearization approach for PEA was presented to achieve feed-forward inverse compensation for hysteresis. Meanwhile, a composite control method was developed by combining the MPI model and the incremental PID control algorithm, then the composite control algorithm was implemented in DSP later. At last, the performance of proposed algorithm was verified on test platform. The experimental results indicate that when system was input with damped sinusoidal and constant amplitude sinusoidal curves of 10 Hz and 100 Hz respectively, model error was within 0.59%. Under damped sinusoidal input curve of same frequency under 100Hz, the greatest error of traditional PID algorithm was 59.31μrad, and that of proposed composite algorithm was 14.22 μrad. Experimental data shows that MPI model can accurately depict hysteresis characteristics of PEA and corresponding composite control method has obviously better dynamic tracking performance than traditional PID method, which can satisfy requirements of practical application to FSM.