arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2010.08017cs.RO
收藏

自主个人专用跟随机器人

Autonomous Person-Specific Following Robot

Monash University(莫纳什大学);JDQ Systems(JDQ系统公司);University of British Columbia(不列颠哥伦比亚大学)

Wesley P. Chan, Sina Radmard, Zhao Quan Hew, Jon Morris, Elizabeth Croft, H. F. Machiel Van der Loos

摘要:

在许多人机协作应用中,跟随特定用户是服务机器人的期望甚至是必需的能力。然而,大多数现有的人跟随机器人跟随人,而不知道它在跟随谁。在本文中,我们提出了一个身份特定的人跟踪器,能够跟踪和识别附近的人,使人特定的以下。我们所提出的方法使用了一个顺序最近邻与持票选择算法,我们设计融合在一起的匿名人跟踪器和人脸识别器。实验结果表明,我们提出的方法与其他方法相比,我们的方法在跟踪和识别人,以及提高机器人的性能在以下的目标个人取得了更好的性能。

英文摘要:

Following a specific user is a desired or even required capability for service robots in many human-robot collaborative applications. However, most existing person-following robots follow people without knowledge of who it is following. In this paper, we proposed an identity-specific person tracker, capable of tracking and identifying nearby people, to enable person-specific following. Our proposed method uses a Sequential Nearest Neighbour with Thresholding Selection algorithm we devised to fuse together an anonymous person tracker and a face recogniser. Experiment results comparing our proposed method with alternative approaches showed that our method achieves better performance in tracking and identifying people, as well as improved robot performance in following a target individual.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑