摘要
针对移动机器人人机交互时对环境感知的需求,该文设计一种基于多传感器融合的环境感知方法,以实现在复杂环境下对目标人员的识别与对障碍物的检测。该文通过Aruco码检测,云台相机控制及目标位置预测策略实现移动机器人对目标行人的定位,同时结合点云聚类与拟合实现移动机器人在复杂环境下对障碍物的检测。所提出的目标识别策略在仿真平台和物理的移动机器人平台上进行测试,结果表明该策略能够实时且鲁棒的定位目标行人与检测障碍物。该策略成功实现对移动机器人对复杂环境的检测,并可应用于移动机器人的其他任务中。
To address the environmental perception requirements for human-robot interaction in mobile robots,this paper designs an environmental perception method based on multi-sensor fusion to achieve the recognition of target personnel and the detection of obstacles in complex environments.This paper realizes the localization of target pedestrians through Arucomarker detection,gimbal camera control,and target position prediction strategies.At the same time,point cloud clustering and fitting are combined to realize the mobile robot's detection of obstacles in complex environments.The proposed target recognition strategy was tested on a simulation platform and a physical mobile robot platform.The results show that the strategy can locate target pedestrians and detect obstacles in real time and robustness.This strategy successfully realizes the detection of complex environments by mobile robots and can be applied to other tasks of mobile robots.
作者
林骏杰
吴炜
LIN Junjie;WU Wei
出处
《科技创新与应用》
2025年第28期1-8,共8页
Technology Innovation and Application
基金
国家自然科学基金项目(61807016)。
关键词
移动机器人
人机交互
目标定位
多传感器融合
障碍物检测
mobile robot
human-computer interaction
target positioning
multi-sensor fusion
obstacle detection
