Font Size: a A A

Research On The Robot's Behavior Decision-Making Based On Artificial Emotion

Posted on:2012-11-02Degree:MasterType:Thesis
Country:ChinaCandidate:J X MaoFull Text:PDF
GTID:2218330362950713Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
With the develop of robot technology,people demand robot to work in the environment which are more complex and can not be modeled.Robots are not only implement simnple mechanical actions,but also need greater adaptability and more harmonious machine-user interaction. Traditional artificial intelligence is unable to meet these requirements while as an important component of human intelligence, emotional intelligence begin to receive attention from researchs.The studies of artificial emotion provides the potential to improve robot's intelligence and personification.To explore the role of human emotions can play in the robot,the paper combine the researchs of traditional artificial intelligence and the latest developing artificial emotion, then put forward the methods of robot behavior decision which are all based on artificial. Unlike most existing methods having the disadvantage of complex algorithm,slow learning, real-time differential, the method proposed in this paper is simple, fast learning and real good.Firstly,the paper analyze the research about artificial emotion and robot behavior decision making,points out the therocial siginificance and implementation background of the behavior decision making motheds based on artificial emotion.Secondly,on the base of analyzation of natural emotion features,a fuzzy artificial emotion model is established and a behavior decision making method of robot based on artificial emotion and computational put forward.A virtual experiment is designed to prove the effectiveness.Thirdly,on the base of analyzation of fear emotion features,an artificial emotion model is established for fear and a behavior decision making method of robot based on this artificial emotion model and reinforcement learning put forward.A virtual experiment about goal navigation in the enviroment which have many obstacles is designed to prove the effectiveness.Finally,because of the importance of harmonious machine-user interation,an architecture of virtual agent based on artificial emtion and many machine learning algorithm is proposed and implement it in an emotional agent.A interface is established in Matlab/GUI to verify the effectiveness and rationality of the architecture. In short, this paper presents a variety of behavior decision-making methods based on artificial emotion and they were all verified by simulation.the paper do some therocial attempt on the combination of the traditional robot behavior decision making and new developing artificial emotion. It has great theoretical significance and high application value for enhancing robot's survivability and autonomous decision-making capacity in harsh and complex environment.
Keywords/Search Tags:artificial emotion, behavior decision making, computational intelligence, reinforcement learning, virtual agent
PDF Full Text Request
Related items