Design And Implementation Of Intelligent Education Companion Robot Based On Deep Reinforcement Learning

With the development of Internet technology,the education mode has ushered in a new opportunity for reform.In order to meet the growing demand of students for autonomous learning,the online education mode has emerged and become an indispensable education mode.Various online education platforms are emerging,amount them,the Intelligent Education Platform developed by the School of Computer Science and Technology of Xidian University aims to provide an excellent online learning environment for students,which has aroused strong repercussions among teachers and students.Through the previous multi-dimensional data analysis,it was found that there was a strong correlation between students’ scores and the performance on the platform.Therefore it was necessary to improve the enthusiasm of students in using the Intelligent Education Platform.Deep reinforcement learning is an important part of artificial intelligence.It has achieved great success in many fields,especially in games.Its super game ability has attracted attention.In this context,relying on the Intelligent Education Platform,the thesis builds the game mode of Intelligent Education Companion Robot System based on deep reinforcement learning.The purpose is to strengthen students’ interest and enthusiasm of autonomous learning on the Intelligent Education Platform from the perspective of enhancing interest through the game mode,so as to achieve the purpose of optimizing teaching.The specific of the thesis is as follows:(1)Through the investigation of the development status of online education and deep intensive learning at home and abroad,the thesis explores the lack of attractiveness of online education platforms to students,analyzes the effectiveness of introducing game mode into online education The thesis proposes the overall design of the Intelligent Education Companion Robot System’s game mode,builds the deep reinforcement learning Agent to provide game ability for each student’s exclusive Intelligent Education Companion Robot.The Agent is automatically trained according to the training steps converted from students’ learning points,so that its game ability becomes the intuitive manifestation of students’ learning amount on the Intelligent Education Platform.Then the students’ sense of their exclusive Intelligent Education Companion Robot can be enhanced through the competition mode,so as to encourage them to carry out more learning activities in order to win the competition.(2)The game mode of Intelligent Education Companion Robot System can be divided into two main function modules: training module and competition module.The core function of the training module is to drive training of the Agent based on DQN and A3 C by students’ learning points,and gradually improve their game ability.The game score represents the students’ learning performance on the Intelligent Education Companion Robot System.In order to increase the interest of the game and stimulate students’ interest,the trained Agents can participate in the multiplayer game of the web game and defeat others,so as to simulate the effect of the intelligent learning companion robot in the same game,which is the core function of the game module.The actions of different individuals in the web game are completely controlled by the action instructions.The Agents send action instructions to different individuals in the game according to the dynamic game environment for real-time control.That is to explore a new application method of deep reinforcement learning in the real-time competition of web game.The Intelligent Education Companion Robot System has solved the problems of real-time communication and multi-player competition in web games.It meets the practical application requirements and provides a reliable theoretical and practical basis for more similar application scenarios.