利用扩展卡尔曼滤波对无人机进行在线辨识

Research of online identification of UAV using extended Kalman filter

为了实施带自适应增益的PID控制,较好地实现轨迹跟踪任务,提出一种基于扩展卡尔曼滤波的在线辨识方法。首先,对六旋翼无人机(UAV)以离散时间模型进行建模,考虑了UAV控制系统的非线性特征。然后,针对UAV非线性离散时间模型,开发基于扩展卡尔曼滤波器的神经估计器,利用状态估计完成在线辨识,从而调节PID增益,完成跟踪任务。实验结果验证了所提方法的有效性,所提辨识模型可应用于非线性和未知动力学的系统,具有一定实用性。

In order to implement PID control with adaptive gain and better track tracking task,an online identification method based on extended Kalman filter is proposed.Firstly,a discrete-time model of six rotor UAV is established,and the nonlinear characteristics of UAV control system are considered.Then,according to the nonlinear discrete-time model of UAV,a neural estimator based on extended Kalman filter is developed.The state estimation is used to complete on-line identification,so that the PID gain can be adjusted to complete the tracking task.Experimental results verify the effectiveness of the proposed method.The proposed identification model can be applied to nonlinear and unknown dynamic systems,and has certain practicability.