论文部分内容阅读
对家用的私人机器人来说,个性化服务和预先设计的任务同样重要,因为机器人需要根据操作者的习惯调整住宅状况。为了学习由诱因、行为和回报构成的操作者习惯,本文介绍了行为足迹,以描述操作者在家中的行为,并运用逆向增强学习技巧提取用回报函数代表的操作者习惯。本文用一个移动机器人调节室内温度,来实施这个方法,并把该方法和记录操作者所有诱因和行为的基准办法相比较。结果显示,提出的方法可以使机器人准确揭示操作者习惯,并相应地调节环境状况。
For home private robots, personalized service and pre-designed tasks are equally important because the robot needs to adjust the home’s condition according to the operator’s habits. In order to learn operator habits that are made up of incentives, behaviors, and rewards, this paper presents behavioral footprints to describe the behavior of operators at home and to extract operator habits represented by the reward function using the inverse-enhanced learning technique. In this paper, a mobile robot is used to adjust the room temperature to implement this method and to compare this method with a baseline method of recording all operator inducement and behavior. The results show that the proposed method enables the robot to accurately reveal the operator’s habits and adjust the environmental conditions accordingly.