为了提高多旋翼无人飞行器机载光电平台的扰动补偿能力，实现机载光电平台的稳定跟踪控制，提出一种基于改进扰动观测器和径向基函数(RBF)神经网络逼近的复合补偿控制方法。首先，对现有扰动观测器结构进行改进，构建基于速度信号的改进型扰动观测器，并分析了干扰补偿能力和稳健性；然后，利用RBF神经网络的函数逼近性质解决非线性未知扰动的补偿问题；最后，基于Lyapunov稳定性原理设计出复合补偿控制结构。实验结果表明，机载光电平台的扰动得到有效补偿。该补偿控制方法具有较高的稳定精度和跟踪控制性能，满足多旋翼无人飞行器机载光电平台的稳定控制要求。

In order to compensate disturbance and accomplish the stabilized tracking control for airborne plat-form mounted on multi-rotor unmanned aerial vehicle (MUAV), a self-adjusting tracking control method based on an improved disturbance observer (DOB) and radial basis function (RBF) neural network approximation is pro-posed. First, a compensated control is introduced into feedback loop in the structure of original disturbance ob-server, an improved disturbance observer is established based on velocity signals, and the ability of disturbance compensation and robustness are analyzed. Second, aiming at the compensation problem of nonlinear un-known disturbance, a method based on the RBF neural network (RBFNN) approximation properties is utilized. Finally, a composite compensation control structure is designed based on Lyapunov stability theory. The experi-mental results show that after applying the proposed method, the disturbance of airborne opto-electronic plat-form is compensated effectively. The proposed method has high precision and stable tracking control perfor-mance, and it can fully meet the requirement of airborne opto-electronic platform stability control.