Controller design for human-robot interaction
Controller design for human-robot interaction
复制标题
DOI：
10.1007/s10514-007-9054-7
复制
发表时间：
2008-02-01
影响因子：
3.5
通讯作者：
Trinkle, Jeff
中科院分区：
文献类型：
--
作者：
Meisner, Eric;Isler, Volkan;Trinkle, Jeff
许多机器人任务需要机器人与人类共享相同的工作空间。在这样的环境中，重要的是机器人以不会对工作空间中的人类造成困扰的方式运行。在这篇文章中，我们讨论了设计机器人控制器的问题，使机器人在执行给定任务时产生的压力最小。我们提出了一种新颖的、数据驱动的算法，它计算人类友好的轨迹。该算法利用生物反馈测量并结合一组几何控制器来实现人类友好性。我们使用电流皮肤反应(GSR)传感器来评估人体的舒适度。我们给出了一个人类跟踪任务的结果，在这个任务中，机器人被要求保持在指定的距离内，而不会造成高应力值。
Many robotics tasks require a robot to share the same workspace with humans. In such settings, it is important that the robot performs in such a way that does not cause distress to humans in the workspace. In this paper, we address the problem of designing robot controllers which minimize the stress caused by the robot while performing a given task. We present a novel, data-driven algorithm which computes human-friendly trajectories. The algorithm utilizes biofeedback measurements and combines a set of geometric controllers to achieve human friendliness. We evaluate the comfort level of the human using a Galvanic Skin Response (GSR) sensor. We present results from a human tracking task, in which the robot is required to stay within a specified distance without causing high stress values.