论文部分内容阅读
提出一种基于对比示范的演示学习方法,旨在指导家庭服务机器人快速智能地学习指定任务决策.在模拟家庭环境下,首先利用视觉手段检测和跟踪任务对象,获取“教师”演示正反例关键信息,通过动作识别和状态检测将信息抽象成符号形式,然后经过认知推理得到演示任务的执行决策和约束条件,并通过动作共享指导机器人重现任务.给定演示实例,在真实移动机器人平台上进行了多次成功演示验证,结果表明本方法在家庭服务机器人快速准确的任务学习方面是有效的.
This paper proposes a demonstration learning method based on comparative demonstration to guide the family service robot to learn the task assignment quickly and intelligently.Under the simulated home environment, the visual object is firstly used to detect and track the task object to get the “teacher” demonstration For example, the key information is abstract information into symbolic form through action recognition and state detection, and then the execution decision and constraints of demonstration task are obtained through cognitive reasoning, and the robot is re-tasked through action sharing. Given a demonstration example, The robot platform has been successfully demonstrated for many times. The results show that this method is effective in fast and accurate task learning of home service robot.