Effects of Human Personal Space on the Robot Obstacle Avoidance Be havior: A Human-in-the-loop Assessment
Effects of Human Personal Space on the Robot Obstacle Avoidance Be havior: A Human-in-the-loop Assessment
复制标题
人类个人空间对机器人避障行为的影响：人在环评估
DOI：
--
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Boyi Hu
中科院分区：
文献类型：
--
作者：
Yuhao Chen;Trevor Smith;Nathan Hewitt;Yu Gu;Boyi Hu
为了确保人-机器人交互过程中人类的身体和精神安全，人们积累了大量的文献，并产生了社会可接受的机器人行为的概念。具体地说，它要求机器人的运动不仅要物理上没有碰撞，而且要考虑和尊重在人类社会背景下发展和实施的社会习俗。在这些社会习俗中，个人空间或代理是机器人行为设计中最常见的考虑之一。然而，之前的大多数研究都假设机器人可以通过简单地模仿人类来产生类似人类的运动。机器人的行为算法很少被人类参与者评估和验证。因此，为了填补这一研究空白，进行了一个类似图灵的模拟测试，其中包含两个智能体(每个智能体可以是人类或机器人)在共享空间中的交互。参与者(共33人)被要求识别并标记这些代理人的类别，然后进行问卷调查。结果表明，对适当的机器人行为有不同态度和先前期望的人对该算法的反应不同，他们的识别精度也有显著差异。总的来说，通过在机器人避障算法中考虑个人空间，机器人可以表现出更接近人类的运动行为，这一点得到了人类实验的证实。
To ensure both the physical and mental safety of humans during human-robot interaction (HRI), a rich body of literature has been accumulated, and the notion of socially acceptable robot behaviors has arisen. To be specific, it requires the motion of robots not only to be physically collision-free but also to consider and respect the social conventions developed and enforced in the human social contexts. Among these social conventions, personal space, or proxemics, is one of the most commonly considered in the robot behavioral design. Nevertheless, most previous research efforts assumed that robots could generate human-like motions by merely mimicking a human. Rarely are the robot’s behavioral algorithms assessed and verified by human participants. Therefore, to fill the research gap, a Turing-like simulation test, which contains the interaction of two agents (each agent could be a human or a robot) in a shared space was conducted. Participants (33 in total) were asked to identify and label the category of those agents followed by questionnaires. Results revealed that people who had different attitudes and prior expectations of appropriate robot behaviors responded to the algorithm differently, and their identification accuracy varied significantly. In general, by considering personal space in the robot obstacle avoidance algorithm, robots could demonstrate more humanlike motion behaviors which are confirmed by human experiments.
影响因子：
4.7
作者：
Smith T;Chen Y;Hewitt N;Hu B;Gu Y
通讯作者：
Gu Y