| 英文摘要 | With the development of robotics and its technology, robots are increasingly used in manufacture, service and military and so on. At the same time, intelligent robots with more perceptual knowledge are increasingly demanded by people. The interaction between the robots with emotions and human should include not only the information, but also their emotions, which makes the communication between the human and the robots more natural, and the human will feel closer to the robots. Furthermore, having emotions is also helpful to the behavior decision making of robots, such as visual attention. Therefore, robotic emotion model is becoming the research hotspot in robot research field. It is of great importance theoretically and practically. Firstly, it is a burgeoning and cutting-edge cross discipline, which involves several fields, such as neuroscience, psychology and information science. Secondly, the research on the robotic emotion promotes the development in the field of the behavior decision making of robots. Finally, robotic emotions have positive effects on developing service robots. The introduction of emotion models can help the robots to understand the emotions and intentions of human, which makes the interaction more flexible and efficient. Though the research about robotic emotion has achieved some results, most of them focus on the emotion recognition of human and humanoid head portrait robot, there are still many challengeable problems facing in the research on the robots with emotions. In this thesis, we try to make some further research on some of the critical problems and have made a series of results, as introduced in the following: Firstly, for the problem that the robot’s emotional state is transferred by the external stimulus, we propose to utilize the Hidden Markov Model to research the emotional transfer under the outside stimulation. Emotions are assumed as hidden states and stimuli are assumed as observable states, since stimuli can be observed while the emotions cannot. Moreover, we propose self-transfer matrix according to the emotion’s gradual recovering. Also, we propose stimuli-transfer matrix according to emotion memory effects to stimuli. Based on the self-transfer matrix and stimuli-transfer matrix we proposed, emotions can be evaluated from stimuli more accurately Secondly, for the problem about unconscious emotions, we proposed an algorithm of emotion eliciting with biologic mechanisms. In this algorithm, the ... |
修改评论