In recent years, anxiety among college students has become an urgent social issue that needs to be addressed. In the field of human-computer interaction research, serious game design is rapidly emerging as an effective means of alleviating anxiety. Its combination of entertainment and functionality aligns well with the needs of college students for anxiety relief.
This study focuses on the cognitive-behavioral therapy (CBT) game design. By analyzing existing theories and design examples, the study identified opportunities in the use of Generative Agents technology for CBT game design. Building upon established therapeutic gaming strategies, this research leveraged the capabilities of Generative Agents such as free interaction, dynamic behavior planning, and adaptive state iteration to innovate gameplay interactions. The study also completed the design of related game narratives, scenes, interaction interfaces, and other experiential elements. During the development process, based on design objectives, the research resolved relevant technical issues, developed the corresponding Generative Agents system, and deliver a fully functional game application.
To verify the effectiveness of the serious game and the rationality of its design strategy, cooperation was conducted with the Tongji University. Mental Health Center, 50 college students were recruited and screened for on-site user experiments. After conducting group comparisons with traditional interaction mode, statistical analyses confirmed that the game application could effectively alleviate anxiety among college students, and that the interactive module of the Generative Agent based game character played a significant role in the final anxiety relief effect of the game.
In summary, this study addresses the design task of CBT games, through user experimentation, the study validated the rationality of the design strategies discussed during the process and the practical value of the application produced. This contributes verified design methods to the fields of Human-AI Interaction and Serious Game Design within human-computer interaction research.
目 录 · Contents
第1章 引言
1.1 选题背景
1.1.1 大学生焦虑情绪
1.1.2 面向焦虑缓释的严肃游戏
1.2 选题依据
1.2.1 以NPC作为CBT游戏优化切入点的依据
1.2.2 生成式智能体对于NPC交互优化的设计机会
1.3 国内外研究现状
1.3.1 焦虑缓释严肃游戏
1.3.2 焦虑缓释疗法
1.3.3 游戏内NPC
1.3.4 生成式智能体
1.4 研究目的、研究问题和研究意义
1.4.1 研究目的
1.4.2 研究问题
1.4.3 研究意义
1.5 研究方法和创新
1.5.1 研究方法
1.5.2 创新点
1.6 研究内容和论文框架
第2章 设计策略
2.1 现有设计理论及实例分析
2.1.1 CBT游戏
2.1.2 生成式智能体
2.2 目标用户研究
2.2.1 问卷调查
2.2.2 深入访谈
2.3 设计策略阐述
2.3.1 基于生成式智能体的互助玩法策略
2.3.2 策略可行性分析
2.4 专家访谈
第3章 CBT游戏设计
3.1 游戏流程设计
3.1.1 NPC焦虑状态观察
3.1.2 信念输出及讨论
3.1.3 互动结果及行为反馈
3.2 游戏叙事设计
3.2.1.游戏场景设计
3.2.2.角色设计
3.3 游戏玩法设计
3.3.1 自由对话互动玩法
3.3.2 饮品匹配及明信片收集玩法
3.4 游戏美术设计
3.4.1 游戏美术风格
3.4.2 游戏交互界面设计
第4章 智能体系统实现
4.1 NPC交互模块构建
4.1.1 智能体响应时延问题
4.1.2 双模型推理架构
4.1.3 NPC交互模块评估
4.2 NPC行为模块构建
4.2.1 约定行为空间中的动态生成
4.2.2 行为展现形式
4.2.3 行动模块评估
4.3 NPC成长模块构建
4.3.1 持久化NPC状态迭代
4.3.2 成长模块评估
第5章 应用开发及迭代
5.1 应用开发
5.1.1 客户端开发
5.1.2 服务端工程
5.2 形成性实验
5.2.1 实验目的
5.2.2 实验设计
5.2.3 实验环境及参与者
5.2.4 实验结果
5.3 游戏应用迭代
5.3.1 游戏体验流程调整
5.3.2 小循环间的过渡玩法
5.3.3 游戏交互调整
第6章 用户实验
6.1 实验目的
6.2 实验设计
6.2.1 实验设计方法
6.2.2 实验测量方法
6.2.3 实验记录方法
6.2.4 实验流程
6.2.5 参与者
6.3 实施过程
6.3.1 实验环境
6.3.2 实验材料
6.4 实验结果
6.4.1 焦虑缓释效果
6.4.2 用户参与度
6.5 问卷及访谈结果
6.5.1 问卷结果
6.5.2 访谈结果
6.6 实验结果的分析与讨论
6.6.1 结果汇总
6.6.2 结果讨论
第7章 结论与展望
7.1 结论
7.2 局限性与展望
7.2.1 研究局限性
7.2.2 未来研究方向
参考文献
附录A 相关量表及问卷
附录B 核心代码
致谢
个人简历、在读期间发表的学术成果
部 分 插 图 · Figures
Dis_Lijiale_2025_01
Dis_Lijiale_2025_02
Dis_Lijiale_2025_03
Dis_Lijiale_2025_04
Dis_Lijiale_2025_05
Dis_Lijiale_2025_06
Dis_Lijiale_2025_07
Dis_Lijiale_2025_08
Dis_Lijiale_2025_09
Dis_Lijiale_2025_10
本网站内所有资料属于数字动画与数字娱乐实验室及相应创作团队或企业,未经许可请勿传播(沪ICP备2022027791号)All content is copyrighted by the DAELab or corresponding organizations. No distribution is allowed without permission.