机械与电子Issue(7)：61-65,5.
基于人工情感与Q学习的机器人行为决策
Robot's Behavioral Decision-making Strategy Based on Artificial Emotion and Q-learning
祝宇虹 1毛俊鑫1
作者信息
- 1. 哈尔滨工业大学机器人技术与系统国家重点实验室,黑龙江,哈尔滨,150080
- 折叠
摘要
Abstract
This paper presents a novel emotion - weighed Q learning strategy and apply it in the behavior selection of a virtual robot. By simulating natural emotion generation mechanism, an artificial emotion model of 'fear'is proposed. Moreover,his paper combines this model with reinforcement learning based on the survival function of natural emotion. The comparative simulation experimentswith standard Q learning without emotion and Option HRL prove that the strategy has the advantage of high success rate. This method has great theoretical significance and high application value for enhancingrobot'ssurvivability and autonomous decision - making capacity in harsh and complex environment.关键词
自主移动机器人/人工情感/强化学习Key words
autonomous mobile robot/ artificialemotiom/ reinforcement learning分类
信息技术与安全科学引用本文复制引用
祝宇虹,毛俊鑫..基于人工情感与Q学习的机器人行为决策[J].机械与电子,2011,(7):61-65,5.
