Responsiveness is a challenge for space systems to sustain competitive advantage over al-ternate non-spaceborne technologies.For a satellite in its operational orbit,in-orbit responsiveness is defined as the capabilit...Responsiveness is a challenge for space systems to sustain competitive advantage over al-ternate non-spaceborne technologies.For a satellite in its operational orbit,in-orbit responsiveness is defined as the capability of the satellite to respond to a given demand in a timely manner.In this paper,it is shown that Average Wait Time(AWT) to pick up user demand from ground segment is the ap-propriate metric to evaluate the effect of ground segment location on in-orbit responsiveness of Low Earth Orbit(LEO) sunsynchronous satellites.This metric depends on pattern of ground segment access to satellite and distribution of user demands in time domain.A mathematical model is presented to determine pattern of ground segment access to satellite and concept of cumulative distribution function is used to simulate distribution of user demands for markets with different total demand scenarios.Monte Carlo simulations are employed to take account of uncertainty in distribution and total volume of user demands.Sampling error and standard deviation are used to ensure validity of AWT metric obtained from Monte Carlo simulations.Incorporation of the proposed metric in the ground segment site location process results in more responsive satellite systems which,in turn,lead to greater customer satisfaction levels and attractiveness of spaceborne systems for different applications.Finally,simula-tion results for a case study are presented.展开更多
同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环...同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环境中,广泛存在的动态点使激光点云前后帧的配准精度不高,降低了动态场景下定位和建图的准确性.针对激光点云中存在动态点的问题,本文对SLAM的前端特征提取及后端回环检测模块分别进行改进,以去除动态点,提升SLAM在动态环境下的性能.针对SLAM前端,提出了一种分步的地面分割方法,依据点云高度信息完成地面点粗提取以矫正点云,再使用随机采样一致性方法对矫正后的点云进行精细的地面分割,最后根据高度阈值采用种子生长聚类方法提取非地面动态点,并进行特征提取与配准;针对SLAM后端,使用点云描述子替代传统方法中基于空间位置关系的回环检测方法,以减小累计误差、提高回环检测灵敏度.实验结果显示,本方法在M2DGR street_08序列数据集上较现有方法均方根误差最大降低29.8%,在KITTI04序列数据集上均方根误差最大降幅达42.7%,说明本方法能有效增强动态环境下SLAM系统的全局一致性与定位精度.展开更多
激光雷达同时定位与建图(LiDAR SLAM)技术通常适用于静态环境下,而在动态场景下,定位与建图效果会受到影响;同时,地面分割模块通常用作点云分类处理,然而地面欠分割问题会影响特征点的选择;并且,通常的框架只使用一种回环检测方法,这可...激光雷达同时定位与建图(LiDAR SLAM)技术通常适用于静态环境下,而在动态场景下,定位与建图效果会受到影响;同时,地面分割模块通常用作点云分类处理,然而地面欠分割问题会影响特征点的选择;并且,通常的框架只使用一种回环检测方法,这可能会导致漏检现象。针对上述问题,提出一种动态场景下基于地面分割与回环优化的LiDAR SLAM系统(GSLC-SLAM)。首先,利用lmnet对点云进行动态剔除,该算法将生成的距离图像与残差图像作为网络的输入,并通过SalsaNext网络预测出动态物体;其次,利用高效的gridestiamte算法进行地面分割,该算法利用不均匀网格划分的方法来减少网格的数量,从而保证分割的效率,并利用正交性、高度和平坦度这3个指标进一步筛选地面点;最后,使用由LinK3D(Linear Keypoints for Three Dimensions point cloud)描述子与BoW3D(Bag of Words for Three Dimensions point cloud)词袋构成的新回环检测方法检测回环,该方法利用边缘特征点生成描述子,使用类似于汉明距离的方式进行描述子匹配,并采用类似于词袋的方法构建BoW3D作为LinK3D描述子的数据库,从而对关键帧提取的描述子进行存储以及回环检测。在数据集KITTI(Karlsruhe Institute of Technology and Toyota Technological Institute at Chicago)上的实验结果表明,在KITTI00、02与05序列中与Lego-Loam(Lightweight and ground-optimized LiDAR odometry and mapping)相比,GSLC-SLAM的均方根误差(RMSE)分别降低了5.8%,78.2%,12.5%;相较于F-LOAM(Fast LiDAR Odometry And Mapping),在KITTI00与05序列中GSLC-SLAM的RMSE分别降低了76.7%和53.8%,而在KITTI02序列中GSLC-SLAM表现不佳。经过验证可知,GSLC-SLAM可以有效减少动态物体的干扰、精确分割地面点并减少回环检测的漏检,进而使系统定位精度更高且更鲁棒。展开更多
基金Supported by the Research Council of Shahid Beheshti University,G. C.
文摘Responsiveness is a challenge for space systems to sustain competitive advantage over al-ternate non-spaceborne technologies.For a satellite in its operational orbit,in-orbit responsiveness is defined as the capability of the satellite to respond to a given demand in a timely manner.In this paper,it is shown that Average Wait Time(AWT) to pick up user demand from ground segment is the ap-propriate metric to evaluate the effect of ground segment location on in-orbit responsiveness of Low Earth Orbit(LEO) sunsynchronous satellites.This metric depends on pattern of ground segment access to satellite and distribution of user demands in time domain.A mathematical model is presented to determine pattern of ground segment access to satellite and concept of cumulative distribution function is used to simulate distribution of user demands for markets with different total demand scenarios.Monte Carlo simulations are employed to take account of uncertainty in distribution and total volume of user demands.Sampling error and standard deviation are used to ensure validity of AWT metric obtained from Monte Carlo simulations.Incorporation of the proposed metric in the ground segment site location process results in more responsive satellite systems which,in turn,lead to greater customer satisfaction levels and attractiveness of spaceborne systems for different applications.Finally,simula-tion results for a case study are presented.
文摘同时定位与建图(Simultaneous localization and mapping,SLAM)能够在未知环境中构建地图并为机器人提供定位信息,是移动机器人领域重要研究方向之一.当前,大多数SLAM算法在静态环境中有较好的表现,但是在车辆和行人等运动物体较多的环境中,广泛存在的动态点使激光点云前后帧的配准精度不高,降低了动态场景下定位和建图的准确性.针对激光点云中存在动态点的问题,本文对SLAM的前端特征提取及后端回环检测模块分别进行改进,以去除动态点,提升SLAM在动态环境下的性能.针对SLAM前端,提出了一种分步的地面分割方法,依据点云高度信息完成地面点粗提取以矫正点云,再使用随机采样一致性方法对矫正后的点云进行精细的地面分割,最后根据高度阈值采用种子生长聚类方法提取非地面动态点,并进行特征提取与配准;针对SLAM后端,使用点云描述子替代传统方法中基于空间位置关系的回环检测方法,以减小累计误差、提高回环检测灵敏度.实验结果显示,本方法在M2DGR street_08序列数据集上较现有方法均方根误差最大降低29.8%,在KITTI04序列数据集上均方根误差最大降幅达42.7%,说明本方法能有效增强动态环境下SLAM系统的全局一致性与定位精度.
文摘激光雷达同时定位与建图(LiDAR SLAM)技术通常适用于静态环境下,而在动态场景下,定位与建图效果会受到影响;同时,地面分割模块通常用作点云分类处理,然而地面欠分割问题会影响特征点的选择;并且,通常的框架只使用一种回环检测方法,这可能会导致漏检现象。针对上述问题,提出一种动态场景下基于地面分割与回环优化的LiDAR SLAM系统(GSLC-SLAM)。首先,利用lmnet对点云进行动态剔除,该算法将生成的距离图像与残差图像作为网络的输入,并通过SalsaNext网络预测出动态物体;其次,利用高效的gridestiamte算法进行地面分割,该算法利用不均匀网格划分的方法来减少网格的数量,从而保证分割的效率,并利用正交性、高度和平坦度这3个指标进一步筛选地面点;最后,使用由LinK3D(Linear Keypoints for Three Dimensions point cloud)描述子与BoW3D(Bag of Words for Three Dimensions point cloud)词袋构成的新回环检测方法检测回环,该方法利用边缘特征点生成描述子,使用类似于汉明距离的方式进行描述子匹配,并采用类似于词袋的方法构建BoW3D作为LinK3D描述子的数据库,从而对关键帧提取的描述子进行存储以及回环检测。在数据集KITTI(Karlsruhe Institute of Technology and Toyota Technological Institute at Chicago)上的实验结果表明,在KITTI00、02与05序列中与Lego-Loam(Lightweight and ground-optimized LiDAR odometry and mapping)相比,GSLC-SLAM的均方根误差(RMSE)分别降低了5.8%,78.2%,12.5%;相较于F-LOAM(Fast LiDAR Odometry And Mapping),在KITTI00与05序列中GSLC-SLAM的RMSE分别降低了76.7%和53.8%,而在KITTI02序列中GSLC-SLAM表现不佳。经过验证可知,GSLC-SLAM可以有效减少动态物体的干扰、精确分割地面点并减少回环检测的漏检,进而使系统定位精度更高且更鲁棒。
