基于动作捕捉的无人机运动状态识别

Motion State Identification of Unmanned Aerial Vehicle Based on Motion Capture

无人机运动状态识别是无人机运行状态分析的基础,是实现无人机航迹预测的必要条件。对于非合作目标来说,动作捕捉系统可以有效采集其航迹数据。提出一种基于动作捕捉的无人机运动状态识别方法。首先,通过插值、重采样、滤波等方法对包含噪声的无人机航迹数据进行预处理;然后,通过特征提取与特征选择方法,针对速度、加速度、曲率、转角这4个运动参数,提取无人机运动特征;并分割无人机航迹。最后运用支持向量机的方法进行无人机运动状态识别,对速度、加速度、曲率的分类精度分别达到了95%、90%和100%。证明了本方法的可行性。

The unmanned aerial vehicle(UAV)motion states recognition is the basis of UAV operation status analysis and it is a necessary condition for UAV track prediction.For non-cooperative targets,the motion capture system can effectively collect its track data.A UAV motion state recognition method was proposed based on motion capture.Firstly,the noise-containing drone track data is preprocessed by interpolation,resampling,filtering and other methods.Then,feature extraction and feature selection methods are used to extract the four motion parameters:velocity,acceleration,curvature,and rotation angle,drone movement characteristics and segmented drone tracks.Finally,using the method of support vector machine to identify the motion state of UAV,the classification accuracy of speed,acceleration and curvature reached 95%,90%and 100%respectively.The feasibility of this method is proved.