CAJ | 학술논문

为了完成野外环境下无人自主导航车的障碍检测任务，应用AK—means聚类算法对四线激光雷达数据进行挖掘，聚类结果输出为候选障碍，为了克服车辆颠簸引起的错误聚类，引入加权欧氏距离对聚类评价指标DBI进行改进。试验结果表明，障碍检测算法在车辆低速行驶时稳定可靠。
위료완성야외배경하무인자주도항차적장애검측임무，응용AK—means취류산법대사선격광뢰체수거진행알굴，취류결과수출위후선장애，위료극복차량전파인기적착오취류，인입가권구씨거리대취류평개지표DBI진행개진。시험결과표명，장애검측산법재차량저속행사시은정가고。
In order to obtain obstacle information in field environment for an unmanned ground vehicle（ UGV）, AK -means clustering algorithm was applied in four - layer laser radar data mining. The result of clustering output candidate obstacles. To overcome the false clustering due to vibration of UGV, using euclidean distance with weight to improve the evaluation - DBI. The experimental results show that the obstacle detection algorithm is reliable in low speed driving.