智能网联汽车交互体验的关键维度：安全感知度、交互自然性 与可理解性的评估方法
文献类型：学位论文
| 作者 | 曹剑琴 |
| 答辩日期 | 2022-06 |
| 文献子类 | 硕士 |
| 授予单位 | 中国科学院大学 |
| 授予地点 | 中国科学院心理研究所 |
| 其他责任者 | 张警吁 |
| 关键词 | 安全感知度 交互自然性 可理解性 智能网联汽车 评估方法 |
| 学位名称 | 理学硕士 |
| 学位专业 | 应用心理学 |
| 其他题名 | Key dimensions of interaction experience in intelligent connected vehicle: assessment methods of perceived security, interactive naturalness and comprehensibility |
| 中文摘要 | The purpose of this paper is to construct an assessment tool to measure perceived security, interactive naturalness and comprehensibility of intelligent connected vehicles through a systematic approach like qualitative research and large-scale questionnaire, and to explore the relationship between these three variables and critical influencing factors in three studies, including eight sub-studies. The method used in this article consists of three main steps: 1) Establish a comprehensive and representative questionnaire set by qualitative research method, dictionary retrieval, literature summary and expert interview; 2) Based on the set of representative items, quantitative research was adopted to investigate the factor structure formed by these items, and the validity was verified by criterion-related variables. 3) We further verified the above findings by using new samples or creating differentiated indicators and traditional usability to further verify the perceived security scale, interactive naturalness scale and the comprehensibility scale of intelligent connected vehicles to ensure the reliability, validity and uniqueness of the structure of the measurement tools developed. In Study 1, the perceived security scale of intelligent connected vehicles was developed and verified. In Study 1a, users in Beijing (N = 373) were selected as the research objects, and an 8-item perceived security scale was established. Through exploratory factor analysis, a two-dimensional model of cognitive safety and emotional safety was found. Pearson partial correlation analysis showed that these two components may have unique effects on different cause (e.g., social support, familiarity) and effect variables (e.g., purchase intention, driving intention). In Study 1b, using a new sample from Shenzhen (N = 352), confirmatory factor analysis consolidated the validity of the two-dimensional model. Further correlation analysis with the addition variables of perceived government support and recommendation intention suggested that the scale had good criterion-related validity. Based on the results of Study 1a and 1b, the main factors affecting cognitive safety are perceived controllability, tendency to seek out new technologies, intelligent connected vehicles driving experience, perceived government support, drive intention and intention to be other road users. The main factors affecting emotional security are social support, familiarity, perceived benefit, purchase and recommendation intention. Study 2 explored the structure and function of interactive naturalness through three sub-studies. In Study 2a, dictionary retrieval, literature review and expert interview were carried out to obtain a 9-item interactive naturalness scale of intelligent connected vehicles. In Study 2b, 353 intelligent connected vehicle users were surveyed by questionnaire. Exploratory factor analysis found two factor structures (joyful fluency and universal awareness). Further regression analysis showed that these two factors had significant and unique predictive effects on key indicators such as satisfaction. In Study 2c, a new sample (N = 349) was used to verify the stability of the two-factor model. The regression results also showed that the two interactive natural experience dimensions also significantly predicted the recommendation intention, loyalty and other important variables. We also found that the joyful fluency significantly predicted the recommendation intention, loyalty and other important variables. Universal awareness was more influenced by interaction and intelligence-related functions. Implications for how the scale can be applied to future human-computer interaction research were discussed in Study 2. In Study 3, the comprehensibility scale of intelligent connected vehicles was developed and verified. In view of the lack of measurement of scene and interpretation requirements in previous studies on the comprehensibility scale, Study 3a obtained the 19-item comprehensibility scale of intelligent connected vehicles Study 3b constructed the comprehensibility scale of intelligent connected vehicles (random grouping 1, N = 1036), and obtains a four-factor scale with good reliability and validity. In Study 3c (random grouping 2, N = 1035), according to existing research conclusion and expert assessment, differentiated explanations were created (state information only, state information and reasons, state information / reasons and forecast information before takeover, state information / reasons and forecast information after takeover). According to literature analysis and industry reports, 8 typical takeover scenarios with time urgency (high urgency, low urgency) were interpreted, and it was found that there was significant difference in comprehensibility scale scores of intelligent connected vehicles under different interpretation levels (p < 0.01). The results showed that it was more accurate in the interpretation quality of distinguishing different temporal urgency (p < 0.001). Meanwhile, the measurement results were more consistent with the objective performance results (p < 0.01), and were significantly correlated with the scores of usage intention, satisfaction and curiosity (p < 0.01). The comprehensibility evaluation system with good reliability and validity has been verified in a specific system. This study developed a set of psychological measurement tools for measuring safety perception, interaction naturalness and comprehensibility in intelligent connected vehicles, which made up for the lack of qualitative research basis, quantitative research verification and single objective in previous studies. The specific mechanism of psychological factors involved in this tool deserves further attention. The influencing factors identified in this study can be used to conceive new ways to improve people's perception of safety, natural interactive experience of product interaction, and comprehensibility. |
| 英文摘要 | 本研究旨在通过质性研究和大规模量表的定量研究等系统方法，对智能网联汽车交互体验的关键维度进行测量工具的开发。研究内容包括三部分构建测量智能网联汽车的安全感知度（研究一）、交互自然性（研究二）和可理解性（研究三）的量表，并探究这三个构念与关键影响因素关系。主要包括三个步骤：1）采用质性研究方法，结合字典检索，文献总结和专家访谈等方式建立全面而具有代表性的条目集；2）以代表条目集为基础，采用定量研究方法，考察这些条目所形成的因素结构的稳定性，并结合关键效标进行效度验证。3）使用新的样本和创设有区分度的和传统可用性指标对智能网联汽车的安全感知度量表、交互自然性量表和可理解性量表进行进一步验证，以保证所开发的测量工具的结构有可靠性、有效性和独特性。 研究一开发智能网联汽车安全感知度测评量表并验证其有效性。研究 1a 以北京地区用户为研究对象(N = 373)，建立了 8 个条目的安全感知度量表。通过探索性因素分析，发现“认知安全和情感安全”二维模型。皮尔逊偏相关分析表明，这两个成分可能对不同的原因（如社会支持、熟悉度）和后果变量（如购买意愿、驾驶意愿）有独特的影响。研究 1b 以深圳地区的新样本为研究对象(N = 352)，通过验证性因素分析再次证明了二维模型的有效性。增加了感知政府支持和推荐意愿两个变量的进一步相关分析表明，量表具有良好的标准相关效度。基于研究 1a 和 1b 的结果，影响认知安全的主要因素包括感知可控性、寻求新技术的倾向、智能驾驶汽车驾驶经验、感知政府支持、驾驶意愿和成为共同道路使用者的意图共 6 个。影响情感安全的主要因素包括社会支持、熟悉度、利益感知、购买和推荐意愿共 4 个。 研究二通过 3 个子研究探究了交互自然性的结构与作用。研究 2a 通过开展字典检索，文献回顾和专家访谈，得到了包含有 9 个条目的智能网联汽车交互自然性量表。研究 2b 问卷调查了 353 名智能网联汽车用户，探索性因素分析发现了通达舒畅和随境应人的两因素结构。后续回归分析表明这两个因素对满意度等关键效标有显著独特的预测作用。研究 2c 使用新样本 (N = 349) 验证了通达舒畅和随境应人的双因素模型的稳定性，回归结果还显示这两个交互自然性体验维度对推荐意愿、忠诚感等重要变量有显著预测作用，通达舒畅更多与基本驾驶辅助系统等功能有关，而随境引人更多与交互和智能相关功能有关。研究二还讨论了该量表如何用于未来的人机交互研究。 研究三开发了智能网联汽车的理解性量表并验证其有效性，研究 3a 针对以往有关可理解性量表的研究缺乏对场景和解释需求进行测量的问题，根据行业手册规范与专家访谈结果，得到 19 个条目的智能网联汽车可理解性量表。研究3b 对智能网联汽车的可理解性量表进行结构构建 (随机分组 1，N = 1036)，得到信效度良好的四因子量表。研究 3c (随机分组 2，N = 1035)，根据已有研究的一致结论和专家评估，创设有区分度的解释（只呈现状态动作信息，状态+原因信息，状态+原因+接管前预测信息，状态+原因+接管后预测信息）。根据文献分析和行业报告得到的 8 个典型接管场景中，对有时间紧迫性（高紧迫性、低紧迫性）区分的场景进行解释，发现不同解释水平下，智能网联汽车的可理解性量表得分有显著差异 (p < 0.01)，说明它在区分不同时间紧迫性的解释质量上更准确 (p < 0.001)；同时，其测量结果与客观绩效结果的一致性更高 (p < 0.01)，且与使用意愿、满意度和好奇心的得分有显著相关 (p < 0.01)。研究三得到了主客观结合，信效度良好的智能网联汽车的可理解性量表。 本研究开发了一套用于测量智能网联汽车中安全感知度、交互自然性和可理解性的心理测量工具，弥补了以往研究缺乏质性研究基础，缺乏定量研究验证和单条目的不足。该工具涉及的心理因素中的具体机制值得进一步关注。本研究发现的影响因素可以用来构思新的方法来提高人们的安全感知、产品交互自然性体验和可理解性。 |
| 语种 | 中文 |
| 源URL | [http://ir.psych.ac.cn/handle/311026/43117]  |
| 专题 | 心理研究所_社会与工程心理学研究室 |
| 推荐引用方式 GB/T 7714 | 曹剑琴. 智能网联汽车交互体验的关键维度：安全感知度、交互自然性 与可理解性的评估方法[D]. 中国科学院心理研究所. 中国科学院大学. 2022. |
入库方式： OAI收割
来源：心理研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。