基于激光测距扫描的室内定位方法研究

摘要

In unknown indoor environment,the algorithm of simultaneous localization and mapping (SLAM)based on iterative closest point (ICP)matching is complex and has more calculating works.It’s difficult to assure the real-time performance if the algorithm is implemented on the ARM-based embedded control system for micro UAVs.An indoor flight localization method based on laser ranging scanning is proposed to reduce matching time and improve the real-time performance of SLAMalgorithm,which uses feature points of indoor environment to replace laser scanning datum for ICP matching.In the process of extracting feature points of indoor environment,a real feature point selection mech-anism is proposed to reduce matching errors caused by false feature points and to assure matching accuracy.Test plat-form has been set up and the test results indicate that the proposed method can meet the requirements of real-time per-formance and accuracy of the indoor flight localization for the micro UAVs.%在未知室内环境下，基于迭代最近点（ICP）匹配的同步建图与定位（SLAM）算法较为复杂，运算量较大，在微小型飞行器 ARM嵌入式控制系统上实现难以保证实时性。本文提出了一种基于激光测距扫描的室内飞行定位方法，用室内环境的特征点替代激光扫描数据进行ICP 匹配，减少了匹配时间，提高了 SLAM算法的快速性。在环境特征点提取过程中，提出了一种真实特征点筛选机制，减少因虚假特征点导致的匹配误差，保证了匹配精度。设计搭建了试验平台，试验结果表明，本文提出的方法能够满足微小型飞行器室内飞行定位快速性和准确性的要求。