Conference Paper

• Trust, Receptivity and Behavior: Pedestrian Interactions with Autonomous Vehicles in Real-World Experiment. (ASPIRE 2025 Under Review)
  Xiang Chang, Zhijie Yi, Hongling Sheng, Yichang Liu, Dengbo He.

• From Driver to Passenger: Understanding Evaluation Gaps in “Fantastic” Driving Behaviour Delivery. (Submitted to CSCW 2025)
  Zhijie Yi, Yueteng Yu, Xiang Chang, Xinyu Yang, Mengdi Chu, Junrong Lu, Yiyao Liu, Jingli Qin, Ye Jin, Jialin Song, Guyue Zhou, Jiangtao Gong*
  Abstract: The accurate evaluation of driving behaviors is crucial for optimizing and implementing autonomous driving technology in practice. However, there is no comprehensive understanding of good driving behaviors currently. In this paper, we sought to understand driving behaviors from the perspectives of both drivers and passengers. We invited 10 expert drivers and 14 novice drivers to complete a 5.7-kilometer urban road driving task. After the experiments, we conducted semi-structured interviews with 24 drivers and 48 of their passengers (two passengers per driver). Through the analysis of interview data, we found drivers’ considerations and efforts to achieve good driving, passengers’ assessing logic of driving behaviors, and gaps between these perspectives. Our research provided a systematic understanding of driving behavior design and evaluation for autonomous driving and valuable insights for future autonomous vehicle design.

• SurrealDriver: Designing LLM-powered Generative Driver Agent Framework based on Human Drivers’ Driving-thinking Data. (IROS 2024)
  Ye Jin, Ruoxuan Yang, Zhijie Yi, Xiaoxi SHEN, Peng Huiling, Xiaoan Liu, Jingli Qin, Li Jiayang, Peizhong Gao, Guyue Zhou and Jiangtao Gong*
  Abstract: Leveraging advanced reasoning capabilities and extensive world knowledge of large language models (LLMs) to construct generative agents for solving complex real-world problems is a major trend. However, LLMs inherently lack embodiment as humans, resulting in suboptimal performance in many embodied decision-making tasks. In this paper, we introduce a framework for building human-like generative driving agents using post-driving self-report driving-thinking data from human drivers as both demonstration and feedback. To capture high-quality, natural language data from drivers, we conducted urban driving experiments, recording drivers’ verbalized thoughts under various conditions to serve as chain-of-thought prompts and demonstration examples for the LLM-Agent. The framework’s effectiveness was evaluated through simulations and human assessments. Results indicate that incorporating expert demonstration data significantly reduced collision rates by 81.04% and increased human likeness by 50% compared to a baseline LLM-based agent. Our study provides insights into using natural language-based human demonstration data for embodied tasks. The driving-thinking dataset is available at https://github.com/AIR-DISCOVER/Driving-Thinking-Dataset.

• Improving Knowledge Asymmetry in Group Discussions with Smart Assistants. (HCII 2024)
  Hongfei Wu, Chiju Chao, Zhijie Yi, Zhiyong Fu*
  Abstract: This study aimed to explore the role of smart assistants in alleviating knowledge asymmetry in interdisciplinary group discussions. Due to diverse disciplinary backgrounds and foundational knowledge differences, group members struggle to engage effectively in group discussions. To address this issue, we introduced a smart assistant named CaseAssistant, designed to provide relevant information for discussions, thereby promoting cognitive consensus among participants. Through qualitative research, including user interviews, semi-structured group discussions, and observer feedback, we deeply analyzed the impact of CaseAssistant on enhancing discussion efficiency, reducing participant workload, and expanding the depth and breadth of discussions. The findings indicate that the introduction of smart assistants can lower the workload of group members, promotes cognitive consensus, and effectively improves the quality and efficiency of group discussions.

  
  

Journal Paper

• An Emotional Design Model for Future Smart Product Based on Grounded Theory. Systems. 2023; 11(7):377. https://doi.org/10.3390/systems11070377
  Chiju Chao, Yu Chen, Hongfei Wu, Wenxuan Wu, Zhijie Yi, Liang Xu, Zhiyong Fu*
  Abstract: Recently, smart products have not only demonstrated more functionality and technical capabilities but have also shown a trend towards emotional expression. Emotional design plays a crucial role in smart products as it not only influences users’ perception and evaluation of the product but also promotes collaborative communication between users and the product. In the future, emotional design of smart products needs to be regarded as an important comprehensive design issue, rather than simply targeting a specific element. It should consider factors such as design systems, values, business strategies, technical capabilities, design ethics, and cultural responsibilities. However, currently, there is a lack of a design model that combines these elements. Currently, there are numerous practices in emotional design for smart products from different perspectives. They provide us an opportunity to build a comprehensive design model based on a large number of design case studies. Therefore, this study employed a standardized grounded theory approach to investigate 80 smart products and conducted interviews with 12 designers to progressively code and generate a design model. Through the coding process, this research extracted 547 nodes and gradually formed 10 categories, ultimately resulting in a design model comprising 5 sequential steps. This model includes user requirements, concept definition, design ideation, design implementation, and evaluation, making it applicable to most current and future emotional design issues in smart products.