仿人机器人交互行为设计评估方法

摘要

研究目的在于为仿人机器人的交互行为设计制定评估方法.将动作捕捉系统采集的真人动作数据与采用虚拟现实技术设计的机器人动作数据进行了对比,并将两者制作成动画视频作为评估试验对象,采用语义分析法确定进行评估的感性量,采用层次分析法确定感性量的权值,组织志愿者观看所制作的动画视频并对视频中的交互动作的感性量打分,对打分结果进行计算,通过对计算结果的分析完成评估.为仿人机器人的交互行为设计提供了一套定量评估方法,实验分析了评估方法的效果.%In this paper,the evaluation method of the humanoid robot interaction design is discussed.Firstly,the motion capture system is used to collect the real action data and the virtual reality technology is used to design the robot motion data,both of which are made into animation video as the evaluation test object.Secondly,the evaluation of the semantic analysis to determine the Kansei dimension is implemented and the analytic hierarchy process (ahp) to determine the amount of Kansei dimensions weight is calculated.Thirdly,the volunteers are asked to watch the video and to grade according to the Kansei dimension.Finally,the results are analyzed and the Kansei graph is drawn to complete the evaluation process.The evaluation method discussed in this paper provides a new approach to quantitatively evaluate the humanoid robot interaction design and the feasibility is also studied.