The two-dimensional(2D)lidar is a ranging optical sensor that can measure the cross-section of the geometric structure of the environment.We propose a robust 2D lidar simultaneous localization and mapping(SLAM)algorit...The two-dimensional(2D)lidar is a ranging optical sensor that can measure the cross-section of the geometric structure of the environment.We propose a robust 2D lidar simultaneous localization and mapping(SLAM)algorithm working in ambiguous environments.To improve the front-end scan-matching module’s accuracy and robustness,we propose performing degeneration analysis,line landmark tracking,and environment coverage analysis.The max-clique selection and odometer verification are introduced to increase the stability of the SLAM algorithm in an ambiguous environment.Moreover,we propose a tightly coupled framework that integrates lidar,wheel odometer,and inertial measurement unit(IMU).The framework achieves the accurate mapping in large-scale environments using a factor graph to model the multi-sensor fusion SLAM problem.The experimental results demonstrate that the proposed method achieves a highly accurate front-end scan-matching module with an error of 3.8%of the existing method.And it can run stably in ambiguous environments where the existing method will be failed.Moreover,it ccan successfully construct a map with an area of more than 250000 square meters.展开更多
为推动大宗茶机械化采收,提升大宗茶鲜叶采收效率与质量,针对目前仿形采茶机感知传感器易受接触作用力、自然光照或茶蓬面叶片间隙影响,提出了融合2D激光雷达与航向姿态参考系统(Attitude and heading reference system,AHRS)的采茶机...为推动大宗茶机械化采收,提升大宗茶鲜叶采收效率与质量,针对目前仿形采茶机感知传感器易受接触作用力、自然光照或茶蓬面叶片间隙影响,提出了融合2D激光雷达与航向姿态参考系统(Attitude and heading reference system,AHRS)的采茶机割刀仿形距离估计方法,在此基础上,设计并研制精度验证试验台与自动仿形采茶样机,分别开展了室内与田间试验。采茶机采用2D激光雷达测量采茶机割刀仿形距离信息,为提升测距精度与实时性,结合AHRS感知的加速度信息,提出了融合2D激光雷达测距与加速度信息(Fusion of 2D-LiDAR ranging and acceleration,FLRA)的采茶机割刀仿形距离估计算法,并研制了算法精度验证装置和方法,验证了算法有效性。室内试验结果表明,算法处理前仿形距离测距误差均值为36.53 mm,标准差为23.21 mm,算法处理后仿形距离估计误差均值为8.56 mm,标准差为6.31 mm,算法处理后的输出数据延迟更小,提升了仿形距离测距精度与实时性。田间试验表明,鲜叶采收效率达180~210 kg·h^(-1),割刀覆盖蓬面上鲜叶的平均采收率为92.38%,平均芽叶完整率为85.34%,平均杂质率为4.93%,一芽三叶及以下嫩梢占90.72%,满足大宗茶机采技术标准和后续加工工艺要求,与传统超声波感知的自动仿形采茶机相比,提升了大宗茶鲜叶采收效果。展开更多
针对目前仿真果园环境研究不考虑地形起伏的问题,基于二维正态分布函数生成具有起伏特征的地面模型,并搭建果园仿真环境测试基于2D激光雷达的同步定位与建图(simultaneous localization and mapping, SLAM)算法的性能。通过控制土地模...针对目前仿真果园环境研究不考虑地形起伏的问题,基于二维正态分布函数生成具有起伏特征的地面模型,并搭建果园仿真环境测试基于2D激光雷达的同步定位与建图(simultaneous localization and mapping, SLAM)算法的性能。通过控制土地模型的起伏变化程度、最大高低落差以及突起分布密度等参数生成不同地形特征的土地;通过2D激光雷达、里程计等传感器获取移动机器人在果园仿真模型中的观测数据;通过可视化SLAM定位偏差及SLAM建图效果评价4种经典2D激光SLAM算法(Hector SLAM,GMapping, Karto SLAM,Cartographer)在具有不同地形的果园环境中的性能。实验结果表明:在果园环境中,随着土地起伏变化程度及最大高低落差的增加,2D激光SLAM算法的定位性能与建图性能将会降低;Hector SLAM能够提供更精准的定位结果,但是建图鲁棒性较差;GMapping能够获得更准确的环境地图,但是定位鲁棒性较差;Cartographer的定位及建图的鲁棒性均较为良好,但会出现少量偏差;Karto SLAM相较于其他算法,在果园环境中不具备优势。展开更多
基金This work was supported by National Key Research and Development Program of China(Grant No.2017YFB1301300).
文摘The two-dimensional(2D)lidar is a ranging optical sensor that can measure the cross-section of the geometric structure of the environment.We propose a robust 2D lidar simultaneous localization and mapping(SLAM)algorithm working in ambiguous environments.To improve the front-end scan-matching module’s accuracy and robustness,we propose performing degeneration analysis,line landmark tracking,and environment coverage analysis.The max-clique selection and odometer verification are introduced to increase the stability of the SLAM algorithm in an ambiguous environment.Moreover,we propose a tightly coupled framework that integrates lidar,wheel odometer,and inertial measurement unit(IMU).The framework achieves the accurate mapping in large-scale environments using a factor graph to model the multi-sensor fusion SLAM problem.The experimental results demonstrate that the proposed method achieves a highly accurate front-end scan-matching module with an error of 3.8%of the existing method.And it can run stably in ambiguous environments where the existing method will be failed.Moreover,it ccan successfully construct a map with an area of more than 250000 square meters.
文摘为推动大宗茶机械化采收,提升大宗茶鲜叶采收效率与质量,针对目前仿形采茶机感知传感器易受接触作用力、自然光照或茶蓬面叶片间隙影响,提出了融合2D激光雷达与航向姿态参考系统(Attitude and heading reference system,AHRS)的采茶机割刀仿形距离估计方法,在此基础上,设计并研制精度验证试验台与自动仿形采茶样机,分别开展了室内与田间试验。采茶机采用2D激光雷达测量采茶机割刀仿形距离信息,为提升测距精度与实时性,结合AHRS感知的加速度信息,提出了融合2D激光雷达测距与加速度信息(Fusion of 2D-LiDAR ranging and acceleration,FLRA)的采茶机割刀仿形距离估计算法,并研制了算法精度验证装置和方法,验证了算法有效性。室内试验结果表明,算法处理前仿形距离测距误差均值为36.53 mm,标准差为23.21 mm,算法处理后仿形距离估计误差均值为8.56 mm,标准差为6.31 mm,算法处理后的输出数据延迟更小,提升了仿形距离测距精度与实时性。田间试验表明,鲜叶采收效率达180~210 kg·h^(-1),割刀覆盖蓬面上鲜叶的平均采收率为92.38%,平均芽叶完整率为85.34%,平均杂质率为4.93%,一芽三叶及以下嫩梢占90.72%,满足大宗茶机采技术标准和后续加工工艺要求,与传统超声波感知的自动仿形采茶机相比,提升了大宗茶鲜叶采收效果。
文摘针对目前仿真果园环境研究不考虑地形起伏的问题,基于二维正态分布函数生成具有起伏特征的地面模型,并搭建果园仿真环境测试基于2D激光雷达的同步定位与建图(simultaneous localization and mapping, SLAM)算法的性能。通过控制土地模型的起伏变化程度、最大高低落差以及突起分布密度等参数生成不同地形特征的土地;通过2D激光雷达、里程计等传感器获取移动机器人在果园仿真模型中的观测数据;通过可视化SLAM定位偏差及SLAM建图效果评价4种经典2D激光SLAM算法(Hector SLAM,GMapping, Karto SLAM,Cartographer)在具有不同地形的果园环境中的性能。实验结果表明:在果园环境中,随着土地起伏变化程度及最大高低落差的增加,2D激光SLAM算法的定位性能与建图性能将会降低;Hector SLAM能够提供更精准的定位结果,但是建图鲁棒性较差;GMapping能够获得更准确的环境地图,但是定位鲁棒性较差;Cartographer的定位及建图的鲁棒性均较为良好,但会出现少量偏差;Karto SLAM相较于其他算法,在果园环境中不具备优势。
