环境特征提取是移动机器人导航中的研究难点,以2D激光雷达为环境感知传感器,针对室内环境中的应用提出一种环境特征提取方法;通过激光雷达的测距精度评估,基于近邻方法对环境中障碍进行空间分类特征提取;对该方法不足进行分析的基础之上,为了能够提取更为精确的特征信息,将SOM方法应用于激光雷达的环境特征提取中;结合装备2D激光雷达LMS291的自行研制的移动机器人采集的数据进行实验,实验结果验证了所提方法的有效性。
It is a crucial problem for mobile robot navigation to extract environmental features. Using a 2D laser scanner as environmental perception sensor, it presents a method to detect the features of indoor environment. By the estimation of ranging accuracy of laser scanner, the neighboring approach is adopted to realize the spatial classification and features extraction of obstacles according to the ranging data of laser scanner. Then, in order to obtain more accurate features information than the neighboring approach, SOM is utilized into features extraction of ranging data of laser scanner. Combined with the sample data set from mobile robot designed by us which is equipped with ZD laser scanner LMS291, the experiment is implemented and its results verifies the validity of this approach.