Estimating and reshaping human intention via human robot interaction
Estimating and reshaping human intention via human robot interaction
复制标题
DOI：
10.3906/elk-1306-85
复制
发表时间：
2016-01-01
影响因子：
1.1
通讯作者：
Erkmen, Aydan Muserref
中科院分区：
文献类型：
--
作者：
Durdu, Akif;Erkmen, Ismet;Erkmen, Aydan Muserref
人机交互的研究主要集中在两个重要的研究领域：意图估计和意图重塑。尽管文献中有许多定义人类意图的研究，但新的研究考察了在HRI中使用机器人重塑人类意图。为了重塑当前的人类意图，本文在真实环境中对两种不同的机器人动作进行了测试。在我们的智能机器人系统中，隐马尔可夫模型(HMM)被用来估计人类的意图。该系统的算法设计包括两个部分：第一部分跟踪环境中的运动物体，第二部分利用智能机器人估计人类意图并重塑估计的当前人类意图。在第一部分中，利用视频处理技术建立了由人类姿态标题和人类和机器人位置组成的特征向量。第二部分是通过隐马尔可夫模型估计人类参与者的当前意图，并将当前意图重塑为另一个意图。该系统在包括人类和机器人在内的真实实验环境中进行了测试，并在文章的最后给出了录制的视频结果。
Human robot interaction (HRI) is studied in two important research areas, intention estimation and intention reshaping. Although there are many studies in the literature that define human intention, new research examines the reshaping of human intentions by using robots in HRI. In this paper, 2 different robot movements are tested in a real environment in order to reshape current human intention. The hidden Markov model (HMM) is used to estimate human intention in our intelligent robotic system. The algorithmic design of the system comprises 2 parts: the first part tracks the moving objects in the environment, and the second part estimates human intention and reshapes the estimated current human intention by using intelligent robots. In the first part, a feature vector consisting of the headings of the human posture and the locations of the humans and robots is created by using video processing techniques. The second part is related to estimating the current intention of a human participant via HMM models and to reshaping the current intention into another intention. The system is tested in a real experimental environment including humans and robots, and the results in the recorded videos are given at the end of the paper.