【摘要】 无人水下航行器(UUV)运行时所携带的多元测距声呐在水下环境获取到的数据置信度较低。因此,采用均值漂移聚类算法对所获取到的数据进行预处理,将异常数据与正常数据分为不同的数据类,得到边界点的集合;然后利用Alpha-Shapes算法将边界点连接成轮廓线;并且为了满足UUV跟踪的理想路径需平滑的要求,利用贝塞尔曲线对构建的轮廓线进行拟合;最后,利用控制方法来实现对所构建地图的跟踪,从而解决了UUV在未知海洋环境下的地图构建问题。更多还原

【Abstract】 The confidence level of data measured by multivariate ranging sonar in an unmanned undersea vehicle(UUV)in an underwater environment is generally low. Therefore, the mean shift clustering algorithm was used to preprocess the obtained data. First, the abnormal and normal data are divided into different data categories to obtain the set of boundary points. Second, the boundary points are connected with contour lines using the Alpha-Shapes algorithm. Furthermore,the Bézier curve is used to fit the constructed contour lines to obtain an ideal, smooth tracking path of the UUV. Finally,a control method was used to track the constructed map. Thus, the problem of UUV map construction in an unknown marine environment was effectively addressed.更多还原