神经网络直接逆控制在无人机飞控系统中的应用研究

Research on the Use of Neural Network Direct Inverse Control in UAV Flight Control System

针对无人机发生故障时系统的非线性耦合特性,提出了一种基于神经网络直接逆控制的飞行控制方法,用于在飞机发生故障时进行控制律重构以恢复对飞机的控制;根据无人机发生故障时非线性运动的特点,确定适当的神经网络结构建立无人机非线性系统的逆系统,并与被控对象串连可以对非线性耦合系统进行线性化解耦,然后引入PID反馈控制进一步提高神经网络逆控制系统的性能;仿真分析表明该方法在无人机发生故障采用传统PID控制失效时,在误差允许范围内可以用来对无人机进行控制。

A kind of neural network direct inverse control is proposed to control the damaged system of UAV （unmanned aerial vehicle） which is nonlinear and correlative. The inverse model of the system which can be established by neural network according to the UAV＇s movement is connected with the object controlled so that the nonlinear and correlative system will become linear and irrelated. PID reactive control is used to improve the performance. Simulation has shown that this kind of control can be used to control UAV when traditional control failed at the ease of damaged UAV system.