Design Research Of Trust Repair Of Vehicle Agent Based On Social Cognition

With the development of advanced technology in intelligent automobile industry,the autonomy,natural interaction and anthropomorphism of vehicle agents(vehicle assistant or vehicle AI)have been improved,which makes them transform to the status of social participants in the human-computer relationship.At the same time,the selfdriving vehicle is heading towards fully autonomous,gaining the ability of selflearning,reasoning and problem solving which means that this technology has the potential to develop in a way that people cannot predict,causing a crisis of public confidence,especially in the case of intelligent automobile related technologies represented by autonomous driving,when mistakes or risky behaviors cannot be completely avoided,how to repair the damaged trust has become an urgent problem to be solved.The content of this thesis mainly includes four aspects: theoretical research,experimental research,construction of trust repair strategy design model and trust repair design practice.Firstly,this study analyzed the cognitive mechanism of humans’ perceptions and impressions forming of others from the perspective of social cognition,as well as the cognitive elements involved in the dynamic process of trust repair,and extracts two universal dimensions of social cognition and perceived trustworthiness impression and attribution theory in trust repair,compared and connected it with the existing humancomputer interaction framework,combined with the interaction design characteristics of vehicle agents,and constructed a trust repair strategy design and evaluation model for vehicle agents research Framework.Secondly,through the scoring evaluation and classification,the desktop research and analysis of the autonomous trust violation scenario and trust repair strategy are carried out.Two types of trust violation events(competence-based violation and intention-based violation)and four trust repair measures(warm-centered impression recovery,competence-centered impression recovery,internal attribution based information strategy,external attribution based information strategy).Through scenario simulation experiments,various types of scenarios and strategies were selected to generate trust repair plans and develop experimental materials.Based on online questionnaires,users’ perceived trustworthiness,intention to use,perceived warmth,and perceived competence were used to evaluate trust repair efficiency,user perceived impression evaluation.Through analyzing the data,so as to obtain the difference in the effect of different trust repair strategies in restoring user trust and improving user perceived impression under different trust violation situations.Then,based on the process of vehicle agent trust rest oration and the quantitative results obtained in the experiment,the design and evaluation model of vehicle agent trust repair and the design principles under the model are constructed.Finally,the experimental conclusions,design principles and design pr ocesses were applied to the trust repair design practice of vehicle agents,and the practicability and feasibility of this theoretical framework and trust repair design model developed in this paper were verified.In terms of theory,this research explored to apply knowledge of sociology and psychology into the research of human-computer interaction,and develop the strategy generation method and evaluation method of vehicle agent trust repair strategy.In terms of application,this paper analyzed the effectiveness of different trust repair strategies in two trust violation scenarios,and obtains the corresponding trust restoration design rules and principles to improve and consolidate users’ trust and impression on in-vehicle systems.