SLAM(Simultaneous localization and mapping,同时定位与地图创建)指机器人在自身位置不确定和未知的环境中创建环境地图,并且利用地图进行自主导航与定位。移动机器人研究的核心课题是同步定位与地图创建,而移动机器人又是室内测绘仪...SLAM(Simultaneous localization and mapping,同时定位与地图创建)指机器人在自身位置不确定和未知的环境中创建环境地图,并且利用地图进行自主导航与定位。移动机器人研究的核心课题是同步定位与地图创建,而移动机器人又是室内测绘仪器的数据采集平台。文中对SLAM在室内测绘仪器的研发现状进行总结,然后结合SLAM测绘仪器室内移动测量系统重点探讨了被动视觉SLAM、主动视觉SLAM、无线局域网(Wi Fi)SLAM和RGB-D SLAM方法的原理、优缺点和研究现状,对两种比较流行的SLAM算法改进思想和研究现状进行了阐述,最后根据SLAM室内测绘仪器的研发探讨了SLAM测绘仪器室内移动测量系统的评价标准与发展趋势。展开更多
巡检机器控制是电厂巡检自动化和智能化技术的核心,但现行方法在实际应用中存在一些不足和缺陷,不仅控制路径平滑系数较低,而且存在碰撞问题,智能巡检机器避障性能较差,无法达到预期的控制效果,为此提出基于SLAM(Simultaneous Localizat...巡检机器控制是电厂巡检自动化和智能化技术的核心,但现行方法在实际应用中存在一些不足和缺陷,不仅控制路径平滑系数较低,而且存在碰撞问题,智能巡检机器避障性能较差,无法达到预期的控制效果,为此提出基于SLAM(Simultaneous Localization And Mapping)算法和动静态规划的电厂智能巡检机器控制方法。利用激光雷达和相机获取巡检环境信息,采用YOLOv3对图像增强,通过点云旋转去除激光点云中离散点,实现对点云数据增强,采用SLAM算法对巡检环境图像和激光点云融合,构建巡检地图和定位巡检机器,采用动静态规划根据环境信息动态调整巡检机器运动轨迹,从而实现对电厂智能巡检机器导航跟踪控制。经实验证明,应用设计方法后,巡检机器路径平滑系数在0.9以上,未发生碰撞,该方法在电厂智能巡检机器控制方面具有良好的应用前景。展开更多
融合物联网技术云计算技术提出智能传感激光SLAM(Simultaneous localization and mapping)算法,首先基于成分分析法相邻帧的点云计算矩阵进行粗配准,再使用改良后算法提供的改进点到线迭代的最近配准算法来弥补传统算法精度较低的问题...融合物联网技术云计算技术提出智能传感激光SLAM(Simultaneous localization and mapping)算法,首先基于成分分析法相邻帧的点云计算矩阵进行粗配准,再使用改良后算法提供的改进点到线迭代的最近配准算法来弥补传统算法精度较低的问题。采用了多重采样的算法在多次复制大权重例子集合的背景下利用小权重粒子集合来提升移动机器人路径定位精准度。最后将改良后的算法运用于AI移动机器人,实验结果表明,改进后的SLAM算法对移动机器人的路径设计的定位精准度有了较大提升,AI机器人可以具备优良的避障功能,对于已知环境或者非完全已知环境中存在的障碍物都具有良好的适应能力。展开更多
Technologies of underground mobile positioning were proposed based on LiDAR data and coded sequence pattern landmarks for mine shafts and tunnels environment to meet the needs of fast and accurate positioning and navi...Technologies of underground mobile positioning were proposed based on LiDAR data and coded sequence pattern landmarks for mine shafts and tunnels environment to meet the needs of fast and accurate positioning and navigation of equipments in the mine underground without satellite navigation signals. A coded sequence pattern was employed for automatic matching of 3D scans. The methods of SIFT feature, Otsu segmentation and fast hough transformation were described for the identification, positioning and interpretation of the coded sequence patterns, respectively. The POSIT model was presented for speeding up computation of the translation and rotation parameters of LiDAR point data, so as to achieve automatic 3D mapping of mine shafts and tunnels. The moving positioning experiment was applied to evaluating the accuracy of proposed pose estimation method from LiDAR scans and coded sequence pattern landmarks acquired in an indoor environment. The performance was evaluated using ground truth data of the indoor setting so as to measure derivations with six degrees of freedom.展开更多
文摘巡检机器控制是电厂巡检自动化和智能化技术的核心,但现行方法在实际应用中存在一些不足和缺陷,不仅控制路径平滑系数较低,而且存在碰撞问题,智能巡检机器避障性能较差,无法达到预期的控制效果,为此提出基于SLAM(Simultaneous Localization And Mapping)算法和动静态规划的电厂智能巡检机器控制方法。利用激光雷达和相机获取巡检环境信息,采用YOLOv3对图像增强,通过点云旋转去除激光点云中离散点,实现对点云数据增强,采用SLAM算法对巡检环境图像和激光点云融合,构建巡检地图和定位巡检机器,采用动静态规划根据环境信息动态调整巡检机器运动轨迹,从而实现对电厂智能巡检机器导航跟踪控制。经实验证明,应用设计方法后,巡检机器路径平滑系数在0.9以上,未发生碰撞,该方法在电厂智能巡检机器控制方面具有良好的应用前景。
文摘融合物联网技术云计算技术提出智能传感激光SLAM(Simultaneous localization and mapping)算法,首先基于成分分析法相邻帧的点云计算矩阵进行粗配准,再使用改良后算法提供的改进点到线迭代的最近配准算法来弥补传统算法精度较低的问题。采用了多重采样的算法在多次复制大权重例子集合的背景下利用小权重粒子集合来提升移动机器人路径定位精准度。最后将改良后的算法运用于AI移动机器人,实验结果表明,改进后的SLAM算法对移动机器人的路径设计的定位精准度有了较大提升,AI机器人可以具备优良的避障功能,对于已知环境或者非完全已知环境中存在的障碍物都具有良好的适应能力。
基金Project(2011CB707102)supported by the National Basic Research Program of ChinaProjects(40901220,41001302)supported by the National Natural Science Foundation of China+1 种基金Project(122025)supported by Fok Ying Tong Education Foundation,ChinaProject(N100401009)supported by Fundamental Research Funds for Central Universities,China
文摘Technologies of underground mobile positioning were proposed based on LiDAR data and coded sequence pattern landmarks for mine shafts and tunnels environment to meet the needs of fast and accurate positioning and navigation of equipments in the mine underground without satellite navigation signals. A coded sequence pattern was employed for automatic matching of 3D scans. The methods of SIFT feature, Otsu segmentation and fast hough transformation were described for the identification, positioning and interpretation of the coded sequence patterns, respectively. The POSIT model was presented for speeding up computation of the translation and rotation parameters of LiDAR point data, so as to achieve automatic 3D mapping of mine shafts and tunnels. The moving positioning experiment was applied to evaluating the accuracy of proposed pose estimation method from LiDAR scans and coded sequence pattern landmarks acquired in an indoor environment. The performance was evaluated using ground truth data of the indoor setting so as to measure derivations with six degrees of freedom.
