Developing semi-autonomous humanoid robots that perform various composite tasks via a task sequencer and dynamics simulator
Developing semi-autonomous humanoid robots that perform various composite tasks via a task sequencer and dynamics simulator
复制标题
开发半自主人形机器人，通过任务排序器和动力学模拟器执行各种复合任务
DOI：
10.1109/arso.2017.8025188
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
F. Kanehiro
中科院分区：
文献类型：
--
作者：
Shin'ichiro Nakaoka;Rafael Cisneros Limón;M. Morisawa;T. Sakaguchi;K. Kaneko;S. Kajita;F. Kanehiro
本文提出了一种系统，使一个人形机器人执行各种任务半自主根据高层程序称为任务序列。任务序列是描述如何基于传感数据、运动规划器和机器人控制器执行特定任务的顺序结构化过程。这些任务序列由任务定序器处理以移动机器人。我们的系统的目标是支持全面的任务，如灾难响应，与一些任务序列定义的各种子任务。为了有效地开发所需的任务序列，任务定序器的设计，使任务序列可以交互式执行;此外，任务定序器集成了动态模拟器，可以用来测试序列，而无需实际移动机器人。我们通过使用HRP-2Kai机器人执行DARPA机器人挑战赛中定义的任务来验证我们的系统，DARPA机器人挑战赛是一项涉及灾难响应机器人的比赛。
This paper proposes a system to enable a humanoid robot to perform various tasks semi-autonomously according to higher-layer procedures called task sequences. A task sequence is a sequentially structured procedure that describes how to perform a particular task based on sensing data, motion planners, and a robot controller. These task sequences are processed by the task sequencer to move the robot. The goal of our system is to support comprehensive tasks, such as disaster response, with a number of task sequences defined for various subtasks. To efficiently develop required task sequences, the task sequencer is designed such that task sequences can be executed interactively; further, the task sequencer is integrated with a dynamics simulator that can be used to test the sequences without actually moving the robot. We verified our system by using it to make our HRP-2Kai robot perform the tasks defined in the DARPA Robotics Challenge, a competition involving disaster response robots.