A user experience map design method based on emotional quantification of in-vehicle HMI

User experience mapping (UEM) is widely used to capture user requirements across domains, but existing maps often lack precise expression and quantification of users’ emotional needs. This study targets the in-vehicle human–machine interface (HMI) and proposes a quantitative user experience map that integrates UEM with Kansei Engineering to quantify drivers’ emotions at behavioral touchpoints. Using the semantic differential (SD) method, the authors measure users’ affective perceptions of the automobile interface, visualize emotional change via user emotion curves, identify pain points, and perform iterative redesign with subsequent validation. The research addresses two questions: (1) Is a quantitative user experience map effective for automotive HMI? (2) Does continuous emotion quantification help improve visual perception and functional layout effects? The goal is to provide a reliable, iterative design method to improve user emotional experience in smart car HMIs.

HMI structure: In-vehicle HMIs are central to intelligent automobiles and influence drivers’ experience through functional layout, visual effects, and interactive behavior. Prior work explored multimodal interaction (gestures, voice) and the impact of icons, voice assistance, and visual perception on user emotion and usability. Usability and emotional experience evaluations for automotive HMIs have been developed, including models linking perceivable characteristics to satisfaction and methods emphasizing the management of complexity for positive emotional experience. Kansei engineering: Originating with Nagamachi, Kansei Engineering links users’ emotions (Kansei) to product design attributes, making difficult-to-quantify perceptions actionable for design. Its application has expanded from physical aesthetics to virtual interfaces, with methods to quantify aesthetic emotions, map interface elements to perceived needs, and apply iterative, task-structure strategies. The SD method is a core tool, employing bipolar adjective pairs rated on Likert scales to quantify affective content. User experience map: UEM visualizes emotional experiences across stages, using elements such as user behavior, touchpoints, user emotions, pain points, and opportunity points. It helps locate pain points by plotting emotional curves over time. Prior studies combined automotive products with experience processes and analyzed in-vehicle music UI usability based on emotional needs. This paper improves UEM by filtering emotional vocabulary via Kansei Engineering and quantitatively scoring touchpoints, increasing accuracy in expressing emotional change.

Study design comprised two parts: Design 1 (baseline evaluation and redesign) and Design 2 (validation). Design 1: Experience staging and evaluation. The user action trajectory was divided into three stages: (S1) observation experience (open driving interface; observe style, color, layout of home interface); (S2) interface-level experience (navigate to secondary interfaces; assess visual effect and functional layout); (S3) overall experience evaluation (complete interface operations and evaluate overall experience). Participants: A six-member expert design team prepared materials and screened Kansei words. Questionnaire participants were licensed drivers aged 20–45 with HMI experience; informed consent obtained and compensation provided. Materials: Interface samples were collected with a focus on Geely Star HMI (home page and secondary interfaces) covering entertainment/multimedia, navigation, communication, car adjustment, and general settings. Procedure: (1) Sample collection: 18 in-vehicle HMI images collected; after screening for clarity/similarity, 10 representative samples retained. (2) Kansei word collection: 196 Chinese Kansei imagery words relevant to in-vehicle HMI gathered from literature, websites, interviews, and prior Kansei studies. (3) Screening: Cluster analysis reduced words to 61; then aligned to design characteristics and the brand theme (“futuristic, intelligent, technological, new standard”) to 20. (4) Formation of final pairs: Using the KJ method with a focus group of six experts, words were categorized and mapped to behavioral touchpoints (color, style, and layouts of home and secondary interfaces), yielding 10 bipolar Kansei word pairs matched to touchpoints. (5) Questionnaire: SD-based 7-point Likert scale (−3 to +3) was used. For S1–S2, pairs included balanced–unbalanced, simple–complex, technological–outdated, compact–decentralized, clear–vague, colorful–uniform in color. For S3 (overall), pairs included decentralized–compact, balanced–unbalanced, clear–vague, convenient–difficult, simple–complex, calming–restless, technological–outdated, cold–warm, individual–ordinary. Data collection and analysis: 163 responses; 161 valid after excluding patterned responses. Reliability was acceptable (Cronbach’s α=0.741). Validity supported (KMO=0.919; Bartlett’s χ²≈5121.41, df=861, p<0.001). Mean scores per touchpoint were computed to quantify emotional satisfaction, and quantitative user experience curves were plotted to identify pain points. Iterative redesign: Based on 9 low-satisfaction touchpoints (of 41), designers conducted interviews (designers and design professionals), aggregated feedback, and applied visual principles (consistency, hierarchy, personalization). Visual redesign focused on aligning with a technological theme, adjusting color strategy toward lower-saturation blue tones, coordinated color balance, and enhancing personalization via functional color blocks (notably for touchpoints P2, P3, P9). Functional layout redesign followed principles of logic, hierarchy, and suggestibility, simplifying the main interface (P1, P4), using larger color-block modules (prioritizing navigation), clarifying functional boundaries, and reducing cognitive load by simplifying layouts and spacing (e.g., air-conditioning interface). Design 2: Validation of improved HMI. Participants: New cohort (non-overlapping with Design 1), including professional interface designers and experienced vehicle users aged 20–45; consent and compensation provided. Materials and procedure: The improved interfaces were evaluated using the same 7-point SD scale on the corresponding Kansei pairs. Data: 109 questionnaires received; 108 valid after excluding inadequate response times. Analysis: Paired t-tests compared before vs after mean scores at α=0.05 and 0.01 levels.

Baseline (before redesign): Quantitative UEM showed substantial variability in touchpoint scores. The multimedia interface exhibited the lowest emotional satisfaction among visual touchpoints, with mismatches between color scheme/product positioning and users’ emotional needs; overall findings indicated insufficient technological feel, color inconsistency, and lack of personalization. For functional layout, lower satisfaction stemmed from complexity and irrational layout; the main page icon layout was not conspicuous and had the lowest layout evaluation; however, simplicity of the main interface, air-conditioning adjustment interface, and multimedia layout received relatively higher scores; increased icon spacing in the air-conditioning interface contributed positively. Overall, 9 of 41 touchpoints were below average (pain points), concentrated in: visual effects (P2, P3, P5, P7, P9) and functional layout (P1, P4, P6, P8). After redesign (validation): Most mapped Kansei touchpoints improved. Significant improvements (paired t-tests) included: technological–outdated for certain settings or related interfaces (e.g., No.8: mean increased from 0.16 to 1.34; t=−6.846; p<0.001), vague–clear and vague–conspicuous (e.g., No.11: t=−3.066; p=0.003; No.29: t=−4.051; p<0.001), and complex–simple across multiple touchpoints (e.g., No.18: t=−4.733; p<0.001; No.30: t=−4.200; p<0.001; No.12: t=−1.987; p=0.049). Balance improved at several points (e.g., No.7: t=−4.298; p<0.001; No.13: t=−2.317; p=0.022; No.25: t=−2.418; p=0.017). Color uniformity improved in some cases (e.g., No.4: t=−2.274; p=0.025), though not all color-related measures changed significantly. A few items showed negligible or mixed changes (e.g., No.20 outdated–technological decreased; p<0.001), but the aggregate pattern indicated higher mean emotional experience values than before. The nine previously identified pain points received higher post-redesign ratings, achieving the intended design improvements.

The findings support the effectiveness of integrating Kansei Engineering with UEM to quantify and visualize emotional experience in automotive HMIs. The method accurately identified pain points—especially deficiencies in technological feel, color consistency, and functional simplicity—and guided targeted redesigns that significantly improved users’ affective ratings at key touchpoints. Visualization via quantitative emotional curves made it straightforward to locate low points and compare pre/post changes, aiding design decisions. The results highlight the influence of functional layout on affective experience during task performance: clearer hierarchy, conspicuous elements, and simplified layouts reduced cognitive load and improved satisfaction. Overall, the approach addresses the research questions by demonstrating that a quantitative UEM is effective for automotive HMI and that continuous emotion quantification helps improve both visual perception and functional layout.

Combining UEM with Kansei Engineering yields a quantitative method that effectively quantifies and visualizes user emotional experience and supports iterative HMI design. The approach: (1) accurately maps emotional expressions to behavioral touchpoints, (2) quantifies and visualizes emotional changes using SD-based metrics and curves, and (3) guides targeted iterations that enhance technological feel, clarity, balance, and simplicity. This toolset helps designers understand users’ emotional needs and streamline product iterations, improving overall user experience. Future research will deepen evaluation in dynamic, real-use contexts through behavioral observation and interviews, extend beyond visual aspects to interaction-level factors, and assess deeper-level interfaces beyond the main and first-level screens.

- The automotive context is complex; stationary test environments differ from real driving conditions, potentially affecting emotional evaluations. - The study primarily quantified visual-level experiences; interaction-level aspects need further exploration. - Analysis focused on the main and first-level interfaces; deeper interface layers require future investigation.