This paper presents the Meta-Collaborative Workstation concept and a gesture-based robot program builder software named MEGURU. The software is ROS-based, and it is publicly available on GitHub. A hand-gestures language has been developed to create a fast and easy to use communication method, where single-hand gestures are combined to create composed Commands, allowing the user to create a customized, powerful, and flexible Gestures Dictionary. Gestures are recognized using an R-FCN Object Detector model fine-tuned on a custom dataset developed for this work.

The system has been tested in two experiments. The first one was aimed at evaluating the user experience of people of different age, sex, and professional background concerning the proposed communication method. The second one was aimed at comparing the traditional teach pendant programming method and MEGURU in the task of assembling a small Moka coffee maker. The results of both experiments highlight that MEGURU is a promising robot programming method, especially for non-expert users.

本文介绍了Meta-Collaborative Workstation的概念以及一个基于手势的机器人程序生成器软件MEGURU。该软件基于ROS，可在GitHub上公开获得。已经开发了手势语言来创建一种快速且易于使用的通信方法，该方法将单手手势组合在一起以创建组合的命令，从而允许用户创建自定义，功能强大且灵活的手势字典。使用针对此工作而开发的自定义数据集进行了微调的R-FCN对象检测器模型可以识别手势。

该系统已在两个实验中进行了测试。第一个旨在评估与所建议的通信方法有关的不同年龄，性别和专业背景的人的用户体验。第二个旨在比较传统的示教器编程方法和MEGURU在组装小型Moka咖啡机的任务中。这两个实验的结果都表明MEGURU是一种很有前途的机器人编程方法，特别是对于非专业用户。