Research on the improved design of lightweight outdoor apparel for urban travel based on the integration of AHP-KE and artificial intelligence techniques

Abstract

To address issues including generally low user satisfaction in the outdoor apparel market, the lack of research on lightweight outdoor apparel for urban travel, and narrowly focused studies, this paper proposes a user-centered design approach. The approach integrates the Analytic Hierarchy Process (AHP), Kansei Engineering (KE), and Artificial Intelligence-Generated Content (AIGC) for improved apparel design. First, the Analytic Hierarchy Process (AHP) was used to construct a hierarchical index of user requirements, highlighting key elements such as garment structure and craftsmanship (weight: 0.4103) and material performance (weight: 0.3184). Second, a Kansei word set was developed, and semantic differential scales combined with Principal Component Analysis (PCA) were employed to analyze data collected from 100 user evaluations, identifying specific design features under the dimensions including (1) structure and craftsmanship and (2) materials. Subsequently, high-priority user needs and recognized design features were converted into actionable design instructions via an AIGC platform, facilitating the intelligent development and optimization of design ideas based on representative samples. Finally, the effectiveness of the optimized design proposals was validated through feedback from target users. The results demonstrate that the proposed AHP–KE–AIGC integrated model effectively converts users’ weighted requirements and ambiguous Kansei preferences for urban travel light outdoor apparel into specific visual design elements, significantly improving user satisfaction with the optimized design schemes compared with conventional outdoor apparel design approaches (mean score > 1). In summary, the integrated pathway for user requirement translation and design generation proposed in this study is applicable not only to light outdoor apparel design but also shows potential for broader application across other apparel categories and related product design domains.

Keywords

lightweight outdoor apparel design AHP KE AIGC functional apparel design urban travel

Introduction

Since 2022, global enthusiasm for outdoor activities has risen significantly. The increasing popularity of healthy lifestyles and the proliferation of varied leisure options have markedly enhanced participation in outdoor sports and expanded the overall market size.¹ Data from the Outdoor Industry Association indicates that involvement in outdoor activities has consistently increased in both Europe–North America and the Asia–Pacific region, hence propelling the advancement of associated categories such as equipment and apparel. Influenced by global trends, China’s tourism and outdoor markets have demonstrated significant revival. According to the Ministry of Culture and Tourism reports that the domestic tourism prosperity index has consistently stayed elevated since 2022, with the most important indicators in 2024 already exceeding those of 2019. At the same time, travel consumption in China has grown increasingly localized and concentrated on short-distance journeys, with intra-provincial travel comprising 81.24% of the total. Tourism scenarios have evolved from conventional activities like hiking and camping to encompass urban and proximate short-distance travel. The rise in local and nearby travel has spurred concurrent expansion in the outdoor wear sector.² The “2024–2025 China Apparel Industry Development Report” indicates that outdoor apparel constitutes 40% of China’s outdoor goods market. This proportion is projected to sustain double-digit compound growth until 2025, presenting substantial opportunities for product design and innovation.

Lightweight outdoor apparel has rapidly emerged as a distinct segment of the outdoor clothing market since 2022. Unlike traditional outdoor apparel, it primarily serves light outdoor activities such as short urban hikes and weekend excursions to nearby areas, with its core feature being an enhanced “adaptability to multiple urban scenarios.” Despite the swift transition of outdoor activity consumption toward “lightweight” experiences like micro-vacations and short-distance urban hiking, research dedicated to lightweight outdoor clothing for urban travel is still scarce. Most current research emphasizes traditional outdoor apparel utilized in extreme settings or particular high-intensity sports, concentrating on factors such as material and functional integration, comfort, and thermo-humidity regulation, and smart technology.^3
–5 The China Customer Satisfaction Index (C-CSI) research indicates that, as of 2024, user satisfaction with outdoor apparel brands stands at 67.8%, markedly below the industry average of 75.4%. This suggests that contemporary outdoor apparel businesses have not adequately fulfilled consumer requirements. Therefore, this study aims to address the mismatch between lightweight outdoor apparel design and user needs in urban travel scenarios in China by exploring a user-oriented improved design pathway.

The Analytic Hierarchy Process (AHP) provides a systematic and empirical method for discovering and refining user requirements across several product categories, including apparel, to solve the research gaps and issues. The fundamental principle involves deconstructing intricate user requirements into many tiers and criteria. A judgment matrix is created through paired comparisons to assess the relative significance of each condition.⁶ In the analysis of user requirements for outdoor apparel such as hardshell jackets, Wen and Lee used AHP to construct a judgment matrix and calculate the importance weights of various design elements, thereby establishing a design priority system.⁷ Nonetheless, although AHP can proficiently quantify and organize user demand criteria, the methodology frequently depends on expert subjective assessment and may not entirely encompass users’ genuine psychological and emotional needs.⁸ In contrast to AHP, Kansei Engineering (KE) effectively captures users’ intrinsic emotions, rendering it particularly appropriates for products such as apparel that fulfill both functional and emotional requirements. Through the extraction and analysis of Kansei words, KE converts these impressions into distinct design aspects, yielding solutions that are more closely aligned with user experience and preferences.

Recent research indicates that the AHP-KE approach can be effectively utilized in design-related domains. For instance, Gao et al. employed the AHP-KE methodology to examine the age-friendliness of community spaces.⁹ However, systematic research on lightweight outdoor apparel in the context of urban travel is still rare in the field of apparel design. Existing studies offer valuable methodological references for this work and highlight the significance of expanding both application areas and research scenarios in this study.¹⁰ On this basis, the study incorporates AIGC technology by integrating intelligent generation tools into the design process. User feedback is integrated to create a comprehensive cycle from need identification to product development and user validation, which improves the efficiency of design solutions and their congruence with user needs.¹¹

In summary, this study introduces an innovative research framework that combines AHP, KE, and AIGC. It systematically analyzes multi-level user requirements for lightweight outdoor apparel in urban travel contexts. Using the semantic differential method in Kansei Engineering, the study identifies key design features preferred by users, aiming to optimize and improve existing products. The main issues to be addressed in this study are as follows:

Q1: What are the multi-level user requirement criteria for lightweight outdoor apparel in urban travel contexts?

Q2: How can the multi-level user requirement criteria for lightweight outdoor apparel in urban travel be effectively translated into specific design elements that reflect user preferences?

Q3: How can design optimization and user feedback be used to verify the effectiveness of lightweight outdoor apparel in meeting multi-level user requirements?

Literature review

Current research status of lightweight outdoor apparel

With the steady recovery of the global tourism market, demand for lightweight outdoor apparel has grown significantly. Nevertheless, a study of pertinent literature indicates that research concerning apparel for urban travel or short-distance contexts is quite few. Most of the current research continues to concentrate on conventional apparel for the outdoors and is designed for extreme conditions or certain high-intensity pursuits, such as mountaineering and skiing. Research issues primarily encompass material and functional integration, comfort and thermal-moisture regulation, and the integration of smart technology.

For example, Liu et al. recognized innovation in lightweight materials as a primary research topic in outdoor wear. They underscored that new materials must possess low weight while including various functions, including insulation, waterproofing, breathability, and antibacterial characteristics, to satisfy the requirements of intricate and evolving environmental conditions.¹² In the area of comfort prediction and data-driven design, Zhang et al. introduced machine learning and environmental modeling to develop a apparel thermal comfort prediction model based on variables such as temperature, climate, and gender. This approach helps guide apparel layering choices for different user groups in specific environments and provides data support for personalized product recommendations and functional optimization.¹³ In recent years, advancements in artificial intelligence technology have led to the integration of intelligent systems in apparel for the outdoors design. These systems incorporate functionalities including heart rate monitoring, UV detection, fall alerts, temperature regulation, and emergency calling, frequently via detachable modules. This trend indicates that lightweight equipment is increasingly incorporating advanced sensing and human-machine interface technologies.¹⁴

The results mentioned above show that lightweight outdoor wear, as a sector of outdoor apparel, remains under-researched systematically. Theoretical and empirical research in this domain is still constrained. The majority of research has concentrated on material features or functional characteristics, with less systematic examination of consumer needs across several levels. There is little investigation into the attributes of lightweight outdoor apparel that meet user needs in urban travel scenarios. This study seeks to present a user-centric research approach for the improved design of lightweight outdoor gear intended for urban travel. It methodically evaluates multi-tiered user requirements and converts them into distinct design attributes that embody user preferences, aiming to optimize and improve current market items.

Application of the Analytic Hierarchy Process (AHP) in the field of design

The Analytic Hierarchy Process (AHP) is a multi-criteria decision-making technique utilized to evaluate the weighting and prioritization of user requirements for various product attributes. It serves to decompose intricate decision problems into several tiers, including goals, criteria (or attributes), sub-criteria, and alternatives. Through pairwise comparisons of these factors with experts or pertinent decision-makers, AHP determines the weights of each aspect and offers a scientific foundation for ultimate decision-making. In comparison to other prevalent multi-criteria decision-making methods in design, AHP has distinct advantages in the analysis of user need criteria: (1) It possesses a distinct and well-articulated hierarchical framework; (2) it offers a systematic assessment procedure and guarantees uniformity in outcomes; (3) it facilitates logical reasoning and causal analysis across many objectives; (4) it integrates both qualitative and quantitative methodologies. This renders it especially efficacious for activities such as user requirement organization, weight assessment, and scheme selection.^15
–19

Research in apparel design indicates that AHP can disaggregate intricate issues into various hierarchical tiers, hence minimizing the likelihood of overlooking critical elements. It is ideally suited for the methodical analysis of the diverse features of various types of clothes. Liu et al. employed AHP to develop a hierarchical index system for evaluating the digital restoration outcomes of excavated garments, offering a scientific and effective approach for analyzing digital restoration in historical apparel.²⁰ Xue et al. employed the AHP method to establish a requirement index system for firefighters concerning new multifunctional interactive firefighting apparel. They subsequently validated the model through subjective testing, illustrating the efficacy of AHP in prioritizing and identifying user requirement criteria.²¹ Even so, the AHP technique possesses specific shortcomings, primarily stemming from its substantial dependence on expert subjective opinion, which may result in biases in the evaluation outcomes. In the realm of lightweight outdoor apparel research, only depending on expert opinions to determine requirement criteria and allocate weights frequently inadequately represents the genuine needs and preferences of customers. This methodology may neglect users’ genuine expectations regarding the precise attributes of lightweight outdoor apparel, therefore diminishing the research’s practical significance and the market relevance of product design.²²

Application of Kansei Engineering in the field of design

To address these issues, Kansei Engineering (KE), a method that focuses on the relationship between user emotional responses and product design, has gained increasing attention in the field of design. The basic principle of Kansei Engineering is to use scientific methods to convert users’ subjective emotional needs—such as feelings, impressions, and preferences—into measurable Kansei factors. These factors are then mapped to specific design elements of the product, providing data-driven and user-oriented support for design decisions.²³ KE has been utilized in domains such as rosewood furniture and product shape design. Through the extraction of Kansei words and their conversion into design qualities, research has successfully aligned consumer preferences with product characteristics.^24
–26 In the field of apparel, Jiang et al. used the semantic differential approach derived from Kansei Engineering to evaluate user views about the combination of Yungang Grottoes Bodhisattva necklaces with contemporary qipao. This methodology discerned the most favored design alternatives and facilitated subsequent design innovation. The methodology enhanced the congruence between the product and consumer expectations while also augmenting the theoretical and methodological basis of garment design.²⁷

Therefore, the integration of AHP with KE amalgamates the systematic benefits of AHP in multi-tier requirement analysis and weight allocation. Simultaneously, it leverages the distinctive value of KE in quantifying user kansei attributes and converting them into design components. This approach efficiently translates multi-level user needs for lightweight outdoor apparel in urban travel into particular design elements that align with user preferences. This establishes a robust theoretical and practical basis for product optimization and facilitates additional validation of the design in this research.

Application of algorithm-based image generation tools in fashion design practice

In recent years, rapid advances in computational intelligence have driven the design field toward greater automation and data-driven approaches, leading to the increasingly widespread application of algorithm-based image generation tools in design practice. Compared with conventional design approaches that rely on designers’ experience and manual exploration, computational intelligence offers new technical support and innovative pathways for design practice through algorithmic analysis, automated generation, and interactive feedback.²⁸

Among the current mainstream algorithm-driven image generation tools, Midjourney, Stable Diffusion, DALL·E, and Adobe Firefly have all been applied in design practice. Among these, Stable Diffusion offers advantages in model customization and structural control, but requires a relatively advanced technical environment and operational expertise. In contrast, DALL·E demonstrates more reliable semantic understanding and generation stability, yet exhibits certain limitations in expressing complex visual styles. Adobe Firefly provides value in image editing and integration within the design workflow; however, its capability for generating intricate apparel designs remains at an early stage of development.^29
–31 Overall performance comparisons indicate that Midjourney demonstrates stronger adaptability in visual style expression and rapid multi-solution generation. Furthermore, related studies have shown that it can efficiently support design proposal development and diversified exploration in both apparel and broader product design practice.³² For example, when creating styles for themed fashion series, participants can use Midjourney to iteratively generate mood boards and style variations by inputting images or text. In product design practice, researchers have also proposed Midjourney-based methods for generating color images that match target visuals.^33,34

These studies indicate that Midjourney holds considerable potential for supporting design ideation and visual concept generation, and they provide a technical foundation for incorporating the platform into the improved design practice of lightweight outdoor apparel in this study.

Methods

Research model construction

To identify the multi-level user requirements for lightweight outdoor apparel in urban travel, this study proposes an AHP-KE-AIGC research model. The model is also used to translate these requirements into specific design features that reflect user preferences. Furthermore, it aims to verify the effectiveness of design optimization and user feedback in meeting these requirements. Based on this framework, the research is divided into four stages:

(1) Extraction and weighting of multi-level user requirements based on literature review and expert interviews.

(2) Application of Kansei Engineering to collect Kansei words and representative samples and to conduct user subjective evaluation studies.

(3) Analysis of user evaluation data and extraction of design elements based on Principal Component Analysis.

(4) Generation of improved design solutions using AIGC and validation of user acceptance.

The research model is depicted in Figure 1. The integration of AHP and KE offers a robust methodological framework with both theoretical and practical benefits. This approach has been effectively utilized to assess Green Express package recycling initiatives. By evaluating both environmental impact and user satisfaction, it can accurately determine the ideal design.³⁵ Research in product creation indicates that this strategy can effectively find essential design aspects favored by users, applicable to both pens and seats. It offers dependable quantitative evidence for enhancing product attractiveness and market competitiveness.³⁶ Despite the extensive use and validation of the AHP-KE technique in domains like furniture and product design, systematic research in the apparel sector—particularly concerning clothing products—remains scarce. This gap provides methodological guidance and avenues for theoretical advancement in this work.

Figure 1.

AHP-KE-AIGC research model.

Based on this, the study further incorporates AIGC (AI-generated content) technology as an important support tool in the design process to enhance both design efficiency and solution generation. Unlike traditional manual drawing and experience-based judgment, AIGC can intelligently redesign lightweight outdoor apparel for urban travel based on existing design elements and styles. It also supports a range of tasks, such as style variation and fabric coordination. AIGC not only expands the application scope of the AHP-KE method but also builds an efficient feedback mechanism “from data to design” through intelligent recognition and rapid iteration. This approach effectively improves the development efficiency and quality of improved designs for lightweight outdoor apparel.^37,38

Construction of user requirement criteria

The AHP necessitates the decomposition of the problem into a hierarchical framework. This study produced a user demand hierarchy model for enhancing the design of lightweight outdoor apparel for urban travel, utilizing a literature analysis and expert interviews to address both design attributes and the genuine needs of the target user demographic. This model categorizes the primary user requirements into three tiers: goal level, criteria level, and sub-criteria level.³⁹ See Table 1.

Table 1.

Hierarchical structure of user requirements for the improved design of lightweight outdoor apparel for urban travel.

Goal level	Criteria level	Sub-criteria level
Overall User Requirements for the Improved Design of Lightweight Outdoor Apparel for Urban Travel	B1 Garment Structure and Craftsmanship Requirements	High-Capacity Pockets C1
		Non-Slip Band Design C2
		Seamless Technology C3
		Taping Technology C4
		Multifunctional Collar Design C5
		Detachable Design C6
	B2 Color and Pattern Requirements	High-Contrast Colors C7
		Natural and Eco-Friendly Colors C8
		Geometric Patterns C9
		Patterns with Natural Elements C10
	B3 Material Requirements	Lightweight Materials C11
		Waterproof and Windproof Materials C12
		Breathable Materials C13
		Quick-Drying Materials C14
		Antibacterial Materials C15
	B4 Usage Requirements	Comfort C16
		Convenience C17
		Safety C18
		Adaptability C19

Next, an online questionnaire survey was conducted. 22 experts in related fields were invited to complete the AHP questionnaire. The participants included university faculty and industry professionals with experience in apparel design, functional apparel development, and user research. In this process, the study used a 1–9 scale to assign quantitative scores to each pairwise comparison of factors. In particular, the factors being compared were labeled as i and j and scored according to the scale shown in Table 2. All returned expert questionnaires were checked for validity before data entry. The review included completeness of the judgment matrices, a consistency ratio (CR) below 0.1, and verification that the expert backgrounds matched the research field. Only questionnaires meeting these criteria were considered valid and included in the analysis. This ensured the scientific accuracy of the weight calculations and the consistency of the judgment logic.⁴⁰ A total of 22 questionnaires were distributed in this study. After excluding invalid responses, 20 valid questionnaires were obtained.

Table 2.

Scale criteria for assigning values and interpreting the judgment matrix.

Scale	Scale meaning
1	Element i is equally important as element j
3	Element i is slightly more important than element j
5	Element i is more important than element j
7	Element i is much more important than element j
9	Element i is definitely more important than element j.
2, 4, 6, 8	The importance of element i relative to element j is between the previous levels of assessment.
Reciprocal	If the importance of element i relative to element j is represented by a_ij, then the importance of element j relative to element i is represented by a_ji, where a_ji is the reciprocal of a_ij.

Based on the established evaluation index system, this study further developed the hierarchical pairwise comparison matrix $A_{n \times n} = {[a_{i j}]}_{n \times n}$ . This matrix is used to compare and assess the factors within each level.

A_{n \times n} = [\begin{matrix} 1 & a_{12} & a_{13} & \dots & a_{1 n} \\ a_{21} & 1 & a_{23} & \dots & a_{2 n} \\ a_{31} & a_{32} & 1 & \dots & a_{3 n} \\ ⋮ & ⋮ & ⋮ & ⋮ & ⋮ \\ a_{n 1} & a_{n 2} & a_{n 3} & \dots & 1 \end{matrix}]

(1)

To incorporate the opinions of multiple experts, this study adopts the geometric mean method. The scoring matrices from m experts (m = 1,2, . . . k) were multiplied element-wise, and then the m-th root was taken. This produced the integrated judgment matrix $I_{n \times n}$ , in which each element is the geometric mean of the corresponding elements from the m matrices. The formula is as follows:

I_{n \times n} = {(\prod_{k = 1}^{m} a_{i j}^{k})}^{\frac{1}{m}} = [\begin{matrix} i_{11} & i_{12} & \dots & i_{1 n} \\ i_{21} & i_{22} & \dots & i_{2 n} \\ ⋮ & ⋮ & ⋮ & ⋮ \\ i_{n 1} & i_{n 2} & \dots & i_{n n} \end{matrix}]

(2)

The calculation of index weights is a core step in the Analytic Hierarchy Process. In this study, the geometric mean method was used to calculate the weights. First, the n-row vectors in the integrated judgment matrix I are multiplied, with the product denoted as K_i. The calculation formula is as follows:

K_{i} = \prod_{i = 1}^{n} i_{i}

(3)

Next, each component of the resulting vector K_i is taken to the n-th root to obtain the vector, which is used to estimate the weights. M_i is the unnormalized priority vector calculated by the root method. The calculation formula is as follows:

M_{i} = \sqrt[n]{K_{i}}

(4)

Afterward, the vector M_i is normalized to produce the final weight vector. The calculation formula is as follows:

F_{i} = \frac{M_{i}}{\sum_{i = 1}^{n} M_{i}} = [\begin{matrix} F_{1} \\ F_{2} \\ \dots \\ F_{n} \end{matrix}]

(5)

Consistency testing of the judgment matrix is a key step in ensuring that the weight assignments within the matrix are reasonable and logically consistent. When the judgment matrix is of order greater than 2, the consistency index (CI) and consistency ratio (CR) must be calculated. If the CR is less than 0.1, the judgment matrix exhibits acceptable consistency, and its weight assignments are considered valid. Conversely, if the CR fails to meet this condition, the matrix elements should be adjusted until consistency is achieved.

The calculation formula for the consistency index (CI) is as follows:

C I = \frac{λ_{m a x} - n}{(n - 1)}

(6)

Where $λ_{m a x}$ denotes the approximate maximum eigenvalue of the judgment matrix, calculated as follows:

λ_{m a x} = \sum_{i = 1}^{n} \frac{{(I W)}_{i}}{{(n W)}_{i}}

(7)

The consistency ratio (CR) depends on the consistency index (CI) and the average random consistency index (RI). It is expressed as:

C R = \frac{C I}{R I}

(8)

The general reference values for the random consistency index (RI) are presented in Table 3.^41,42

Table 3.

Average random consistency index (RI) for judgment matrices.

Matrix Order	1	2	3	4	5	6	7	8	9	10	11	12
RI	0	0	0.52	0.89	1.12	1.26	1.36	1.41	1.46	1.49	1.52	1.54

Construction of the User Evaluation Questionnaire

The weighted ranking obtained from the AHP analysis pinpoints the requirements most important to users. These requirements become the primary design variables prioritized in the Kansei Engineering stage. During the Kansei semantic-matching stage, design indicators with higher weights are mapped first to users’ Kansei semantics. This approach ensures that the design process fully reflects core user requirements, reduces subjectivity and ambiguity in determining the design direction, and increases the scientific rigor and accuracy of design decisions.

Establishment of Kansei words

Prior to the construction of the user evaluation questionnaire, this study conducted a systematic process of Kansei word acquisition, filtering, and clustering. First, an initial collection of Kansei words was performed by retrieving relevant textual data using “light outdoor apparel” as the core keyword, with the search period spanning from January 2022 to December 2024. Data were collected through both online and offline sources. For online data acquisition, Python-based web crawling techniques were employed to extract authentic user reviews from major Chinese e-commerce and social media platforms, including Taobao, JD.com, Xiaohongshu, and Douyin. For offline data acquisition, textual information was extracted from academic literature and product manuals using the PyPDF2 library. Ultimately, an original text corpus of approximately 30,000 words was established, providing a data foundation for subsequent word frequency analysis.⁴³

Subsequently, the textual data were preprocessed using ROST Content Mining Software. The preprocessing procedures included the removal of invalid symbols and emojis, the elimination of garbled characters, and the deletion of stop words (e.g. functionally meaningless terms such as “is” and “a”). Tokenization and part-of-speech filtering were then performed using the software’s built-in functions. In this study, only adjectives and adverbs were retained, resulting in 39 word-frequency groups for further analysis. Following this, a focus group was convened to conduct expert evaluation. Kansei words with high semantic similarity were consolidated, while terms lacking meaningful content or exhibiting low relevance to light outdoor apparel were eliminated. Through this refinement process, 25 groups of Kansei words closely aligned with the high-weight factors identified by the AHP analysis were obtained.⁴⁴

Next, a semantic structure analysis was conducted on the 25 groups of Kansei words (50 terms in total). Based on the results of in-depth interviews with three experts in the field of fashion design and six consumers, and in conjunction with an auxiliary rating scale employed during the interviews, the semantic relationships among the Kansei words were examined. The auxiliary scale adopted a 7-point Likert format ranging from −3 to 3 (−3, −2, −1, 0, 1, 2, 3). The ratings were used to measure the degree of semantic similarity between any two Kansei words, where higher values indicate greater semantic proximity, a value of 0 represents a neutral relationship with no apparent similarity or opposition, and negative values indicate a tendency toward semantic opposition. Semantic similarity scores were assigned for all 50 Kansei words, resulting in the construction of a 50 × 50 similarity matrix, as shown in equation (9). Each element of the matrix was calculated as the average of all evaluators’ ratings, with the calculation formula provided in equation (10).

S = [\begin{matrix} S_{11} & \dots & S_{1, 50} \\ ⋮ & ⋱ & ⋮ \\ S_{50, 1} & \dots & S_{50, 50} \end{matrix}]

(9)

S_{i j} = \frac{1}{N} \sum_{k = 1}^{N} R_{i j k}

(10)

Here, S_ij represents the final averaged semantic similarity score; N denotes the total number of evaluators (N = 9); and R_ijk is the raw rating given by the k-th evaluator for Kansei words i and j, where i, j∈{1, 2, . . ., 50}.

Finally, the study visualized the semantic structure relationships among the collected data through Multi-Dimensional Scaling (MDS), transforming the similarity matrix into a two-dimensional Euclidean distance model. Before applying MDS, the similarity scores were linearly converted into distance data using the following formula: D_ij = 4 − S_ij, where D_ij denotes the distance between words i and j, and S_ij represents the corresponding average semantic similarity score. The constant 4 is based on the seven-point Likert scale ranging from −3 to 3, ensuring an accurate mapping from similarity to distance. Based on the 50 transformed variables, this study performed MDS in SPSS Statistics 29, following the path “Analyze → Scale → ALSCAL,” to visualize the semantic structure relationships among the 50 Kansei words (see Figure 2). According to the relative positions and proximities of the Kansei words in the semantic space, the words were classified into semantic groups, eventually forming nine semantic groups. These results provided a quantitative basis for selecting representative pairs of antonymous Kansei words from each cluster. The nine finalized pairs of Kansei words established in this study are presented in Table 4. Positive Kansei words were used as evaluative descriptors, with their negative counterparts serving as contrasts; for example, “balanced–imbalanced” represents a pair of opposing Kansei words. These Kansei words effectively capture the subjective preferences of target users for lightweight outdoor apparel and can provide guidance for specifying particular design elements in the design scheme.^44,45

Figure 2.

Semantic group diagram of Kansei words.

Table 4.

Kansei word set.

No.	Kansei word pairs
1	Imbalanced—Balanced
2	Stiff—Soft
3	Rough—Exquisite
4	Heavy—Lightweight
5	Slim-fit—Loose-fitting
6	Poorly layered—Well-layered
7	Tight—Comfortable
8	Prone to wear—Durable
9	Lacking in detail—Rich in detail

Selection of typical apparel samples and questionnaire development

Upon determining the essential Kansei words, this study chose representative product samples for user assessment. 20 samples of lightweight outdoor apparel for urban travel with significant sales were gathered from sources including WGSN trend reports, fashion news, prominent e-commerce platforms, and official brand websites. The review group comprised 2 outdoor apparel designers, 5 customers, and 1 industry person experienced in functional garment development. The panel evaluated and categorized the samples, ultimately selecting 5 final samples (see Figure 3). To guarantee the precision and dependability of the research findings, all chosen samples conformed to the basic design criteria, exhibited comparable pricing ranges, and encompassed various brands, styles, and material attributes. These conditions established a robust foundation for comparison.⁴⁶ Next, a user evaluation questionnaire was developed based on the chosen Kansei words. The questionnaire employed a 7-point Likert scale (−3, −2, −1, 0, 1, 2, 3), with −3 to 3 representing strongly disagree, disagree, somewhat disagree, neutral, somewhat agree, agree, and strongly agree. This format was employed to ascertain users’ explicit sentiments regarding the Kansei dimensions of various garment samples. Figure 4 illustrates an example of the questionnaire.⁴⁷

Figure 3.

Research samples.

Figure 4.

Questionnaire example.

This research focused on young consumers aged 20–30 with an interest in urban travel. The evaluation questionnaire was disseminated on the internet platform Wenjuanxing, and participants were chosen by random sampling. Before participation, respondents were apprised of the questionnaire’s objective and substance. They were mandated to verify that they had undertaken at least one urban short-distance journey in the preceding year to ensure the relevance and validity of the responses. Upon collection, the questionnaires were evaluated as follows: responses with absent items were deemed invalid, and reverse-coded items were utilized to assess answer consistency. Questionnaires exhibiting evident discrepancies were omitted. Only surveys that satisfied these criteria were used in the statistical analysis to guarantee data quality. This study issued a total of 105 questionnaires. Following the exclusion of 5 invalid responses, 100 valid questionnaires were obtained.

Analysis of user evaluation data and extraction of design elements

This study employed Principal Component Analysis (PCA) as a key approach for dimensionality reduction in the processing of Kansei data. The objective is to combine numerous highly collinear Kansei descriptive terms into a limited number of fundamental Kansei elements. This method efficiently detects the essential Kansei attributes of customers about lightweight outdoor apparel for urban travel, facilitating more focused design applications.⁴⁸ The study initially employed SPSS software to evaluate the appropriateness of the gathered scale data, utilizing the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity to ascertain the data’s acceptability for principal component analysis. Upon verifying the prerequisites, components with eigenvalues exceeding one were extracted. The Varimax technique was employed for orthogonal rotation to enhance the clarity of the factor loadings. The total variance accounted for by the retrieved component factors must reach 70% to guarantee sufficient explanatory power and representation of Kansei features. The rotating component matrix and commonality served as the foundation for delineating design features in subsequent stages.

The principal component analysis results revealed the essential requirement dimensions and critical design components anticipated by customers for lightweight outdoor gear intended for urban travel. The evaluation findings from legitimate questionnaires were systematically grouped and evaluated to acquire more representative quantitative data. User evaluations for each Kansei word pair were determined for every garment sample. These scores represent the comprehensive kansei attributes of all respondents concerning the presentation of each kansei word pair across various samples. The sample exhibiting the greatest user acceptability score in each fundamental Kansei dimension was chosen to inform the final design process.

Methods for improved design generation and user validation

In the improved design stage, this study introduced Artificial Intelligence Generated Content (AIGC) technology as a design translation tool for converting the results of Analytic Hierarchy Process (AHP) and Principal Component Analysis (PCA) into concrete design outcomes. To reduce randomness in the generation process and enhance semantic consistency, a combined approach integrating structured prompt construction, visual reference constraints, and iterative generation was employed in the design practice.

Specifically, during the generation process, representative existing apparel samples were introduced as visual reference inputs to constrain the generated outcomes in terms of overall style, proportional relationships, and user preference orientations, thereby reducing deviations toward non-target styles. These visual references functioned solely as morphological constraints and did not directly determine the final design solutions. In the prompt construction process, Kansei words with higher weights identified through principal component analysis were adopted as dominant factors to define the overall style and characteristic orientation of the apparel. Meanwhile, sub-objective indicators derived from the Analytic Hierarchy Process were incorporated as auxiliary factors to further specify design constraints related to silhouette, structure, craftsmanship, materials, and color. The aforementioned factors were integrated into a structured prompt description following a fixed logical sequence to ensure the stability and controllability of the generated semantics. Given the inherent randomness of AIGC-generated outputs, multiple rounds of iterative generation were conducted using the structured prompts, while consistency between the visual reference samples and the prompt content was maintained throughout the process. The generated results were manually screened, with the evaluation focusing on the degree of semantic consistency between the design outcomes and the target Kansei words, as well as their suitability within lightweight outdoor apparel usage contexts. Only those design solutions that simultaneously met the predefined constraints in terms of semantic characteristics and design logic were retained for subsequent analysis.^32,49

Through the above approach, the AIGC generation process was transformed from a random generation mechanism into an assistive design tool jointly regulated by Kansei characteristics, visual constraints, and design conditions, thereby enhancing the stability and consistency of the generated outputs in terms of semantic expression and design logic.

To examine the efficacy of the improved design solutions in fulfilling user requirements, 20 target users from various provinces were solicited to review the produced designs. A 5-point Likert scale (−2, −1, 0, 1, 2) was employed for assessment, where −2 indicates very unsatisfied, −1 denotes dissatisfied, 0 represents neutral, 1 signifies satisfied, and 2 reflects very satisfied. The evaluation dimensions aligned with the criteria-level requirement indicators defined in the previous AHP study. Targeted inquiries assessed if each improved design solution fulfilled user expectations across various criteria parameters. Participants were requested to respond to the following four questions for each improved design: Q1: Are you satisfied with the design and structure of this lightweight outdoor apparel? Q2: Are you satisfied with the material of this lightweight outdoor apparel? Q3: Are you satisfied with the colors and patterns of this lightweight outdoor apparel? Q4: Does this lightweight outdoor apparel meet your needs? The final evaluation results consist of three parts: (1) the total score of each design solution across all evaluation criteria; (2) the average score for each evaluation criterion; and (3) the overall average score for each design solution. Collectively, these results demonstrate both the rationality of the improved design solutions and their acceptance among users.

Results

AHP results for user requirement weights

In the multi-level AHP framework developed in this study, user requirements were divided into three levels: goal, criteria, and sub-criteria. This structure systematically presents the core concerns for lightweight outdoor apparel for urban travel from the user’s perspective. The hierarchy of requirements is shown in Table 1. As shown in the table, the model consists of four primary indicators at the criteria level: B1 Garment Structure and Craftsmanship Requirements, B2 Color and Pattern Requirements, B3 Material Requirements, and B4 Usage Requirements, along with 19 sub-criteria (C1. . .C19). This study analyzed 20 valid questionnaires and obtained the following results:

For the primary user requirement indicators for lightweight outdoor apparel in urban travel—B1 Garment Structure and Craftsmanship Requirements, B2 Color and Pattern Requirements, B3 Material Requirements, and B4 Usage Requirements—the judgment matrix was first aggregated. According to formulas (2) through (8), the product vector, the unnormalized priority vector, and the normalized weight vector were calculated in sequence. The maximum eigenvalue λ_max and the consistency check were also computed. The results for the weights and consistency checks of the matrix are summarized in Table 5.

Table 5.

Weight results for B1–B4 under overall user requirement a in the improved design of lightweight outdoor apparel for urban travel.

A	B1	B2	B3	B4	Evaluation criteria	Weight value	λ_max	CI
B1	1	3.3213	1.4736	2.5529	B1	0.4103	4.0202	0.0067
B2	0.3011	1	0.3119	0.5373	B2	0.1034
B3	0.6786	3.2061	1	2.0816	B3	0.3184
B4	0.3917	1.8612	0.4804	1	B4	0.1679

CR: 0.0076 < 0.1 Consistency Check Passed.

Similarly, for the secondary indicators in the B1 Garment Structure and Craftsmanship Requirements group—including C1 High-Capacity Pockets, C2 Non-Slip Band Design, C3 Seamless Technology, C4 Taping Technology, C5 Multifunctional Collar Design, and C6 Detachable Design—the matrices were aggregated and calculated in sequence according to formulas (2) through (8). The weight results and consistency check for the matrix are summarized in Table 6.

Table 6.

Weight results for C1–C6 under B1 garment structure and craftsmanship requirements.

B1	C1	C2	C3	C4	C5	C6	Evaluation criteria	Weight value	λ_max	CI
C1	1	3.223	4.9555	4.3112	2.4659	2.3035	C1	0.3497	6.395	0.079
C2	0.3103	1	5.6957	3.4649	0.4251	0.3051	C2	0.1244
C3	0.2018	0.1756	1	0.5628	0.2343	0.2519	C3	0.042
C4	0.232	0.2886	1.7768	1	0.2485	0.2614	C4	0.0575
C5	0.4055	2.3522	4.2682	4.0248	1	0.4725	C5	0.1819
C6	0.4341	3.2776	3.9698	3.8252	2.1164	1	C6	0.2445

CR: 0.0627 < 0.1 Consistency Check Passed.

For the secondary indicators in the B2 Color and Pattern Requirements group—including C7 High-Contrast Colors, C8 Natural and Eco-Friendly Colors, C9 Geometric Patterns, and C10 Patterns with Natural Elements—the matrices were aggregated and calculated in sequence according to formulas (2) through (8). The weight results and consistency check for the matrix are summarized in Table 7.

Table 7.

Weight results for C7–C10 under B2 color and pattern requirements.

B2	C7	C8	C9	C10	Evaluation criteria	Weight value	λ_max	CI
C7	1	0.3176	0.4182	0.5422	C7	0.1126	4.0504	0.0168
C8	3.1482	1	2.0651	3.0164	C8	0.4576
C9	2.3911	0.4842	1	1.9871	C9	0.2678
C10	1.8443	0.3315	0.5032	1	C10	0.1619

CR: 0.0189 < 0.1 Consistency Check Passed.

For the secondary indicators in the B3 Material Requirements group—including C11 Lightweight Materials, C12 Waterproof and Windproof Materials, C13 Breathable Materials, C14 Quick-Drying Materials, and C15 Antibacterial Materials—the matrices were aggregated and calculated sequentially according to formulas (2) through (8). The weight results and consistency check for the matrix are summarized in Table 8.

Table 8.

Weight results for C11–C15 under B3 material requirements.

B3	C11	C12	C13	C14	C15	Evaluation criteria	Weight value	λ_max	CI
C11	1	1.8421	1.8601	2.181	1.9965	C11	0.3216	5.0559	0.014
C12	0.5428	1	0.6967	1.8631	1.6592	C12	0.1932
C13	0.5376	1.4353	1	2.0926	1.6986	C13	0.2292
C14	0.4585	0.5367	0.4779	1	0.703	C14	0.1138
C15	0.5009	0.6027	0.5887	1.4226	1	C15	0.1423

CR: 0.0125 < 0.1 Consistency Check Passed.

For the secondary indicators in the B4 Usage Requirements group—including C16 Comfort, C17 Convenience, C18 Safety, and C19 Adaptability—the matrices were aggregated and calculated sequentially according to formulas (2) through (8). The weight results and consistency check for the matrix are summarized in Table 9.

Table 9.

Weight results for C16–C19 under B4 usage requirements.

B4	C16	C17	C18	C19	Evaluation criteria	Weight value	λ_max	CI
C16	1	1.6288	1.9828	1.6873	C16	0.3627	4.0306	0.0102
C17	0.6139	1	1.9836	1.5591	C17	0.2787
C18	0.5043	0.5041	1	0.6693	C18	0.1525
C19	0.5927	0.6414	1.4941	1	C19	0.2061

CR: 0.0115 < 0.1 Consistency Check Passed.

The above calculation results were summarized to obtain the relative weights of each indicator. By multiplying the relative weights at each level, the comprehensive weight for each indicator was determined. The comprehensive weight reflects the hierarchical ranking of the lowest-level indicators concerning the overall goal. The detailed results are presented in Table 10. As shown in Table 10, users assigned the highest weight to B1 Garment Structure and Craftsmanship Requirements, with a weight value of 0.4103. This was followed by B3 Material Requirements, accounting for 0.3184. B4 Usage Requirements and B2 Color and Pattern Requirements ranked next. Based on the comprehensive weight rankings, the improved design of lightweight outdoor apparel for urban travel should place greater emphasis on garment structure and craftsmanship, especially on C1 High-Capacity Pockets and C5 Multifunctional Collar Design. Next in importance are C11 Lightweight Materials and C13 Breathable Materials in the category of material requirements. In addition, under usage requirements, special attention should be given to C16 Comfort. The demand for color and pattern requirements was relatively low. Among the secondary indicators in this category, C8 Natural and Eco-Friendly Colors ranked highest, indicating a preference for such color and pattern features in lightweight outdoor apparel for urban travel.

Table 10.

Summary of indicator weights.

Goal level	Criteria level	Relative weight	Sub-criteria level	Relative weight	Overall weight
Overall User Requirements for the Improved Design of Lightweight Outdoor Apparel for Urban Travel	B1 Garment structure and craftsmanship requirements	0.4103	High-Capacity Pockets C1	0.3497	0.1435
			Non-Slip Band Design C2	0.1244	0.0510
			Seamless Technology C3	0.0420	0.0172
			Taping Technology C4	0.0575	0.0236
			Multifunctional Collar Design C5	0.1819	0.0746
			Detachable Design C6	0.2445	0.1003
	B2 Color and pattern requirements	0.1034	High-Contrast Colors C7	0.1126	0.0116
			Natural and Eco-Friendly Colors C8	0.4576	0.0473
			Geometric Patterns C9	0.2678	0.0277
			Patterns with Natural Elements C10	0.1619	0.0167
	B3 Material requirements	0.3184	Lightweight Materials C11	0.3216	0.1024
			Waterproof and Windproof Materials C12	0.1932	0.0615
			Breathable Materials C13	0.2292	0.0730
			Quick-Drying Materials C14	0.1138	0.0362
			Antibacterial Materials C15	0.1423	0.0453
	B4 Usage requirements	0.1679	Comfort C16	0.3627	0.0609
			Convenience C17	0.2787	0.0468
			Safety C18	0.1525	0.0256
			Adaptability C19	0.2061	0.0346

Results of user evaluation data analysis and extraction of design elements

The AHP analysis facilitates the establishment of a preliminary requirement index system for the improved design of lightweight outdoor apparel intended for urban travel. Nevertheless, these demand indications cannot currently be easily converted into specific design elements. The study employed the Kansei Engineering approach, concentrating on the two most significant indicators identified in the AHP analysis: B1 Structure and Craftsmanship Requirements and B3 Material Requirements. The nine pairs of Kansei words chosen for examination were subsequently organized and classified according to their grammatical characteristics into two categories: garment construction and craftsmanship, and materials.⁵⁰ Table 11 presents the categorized Kansei words, establishing a theoretical foundation for the extraction of design aspects from typical commercial samples in the ensuing analysis.

Table 11.

Kansei Word pairs for lightweight outdoor apparel in urban travel.

No.	Garment structure and craftsmanship category	No.	Material category
1	Imbalanced—Balanced	6	Heavy—Lightweight
2	Rough—Exquisite	7	Stiff—Soft
3	Slim-fitting—Loose-fitting	8	Prone to wear—Durable
4	Poorly layered—Well-layered	9	Tight—Comfortable
5	Lacking in detail—Rich in detail

Following the exclusion of invalid responses, 100 valid questionnaires were gathered for this study, and the original data were input into SPSS Statistics 29 for analysis. Principal component analysis was performed on the Kansei words questionnaire data to uncover the fundamental evaluation characteristics in users’ Kansei replies to lightweight outdoor gear for urban travel. The initial KMO and Bartlett’s test results indicated a KMO value of 0.871, an approximate chi-square value of 1098.622, 36 degrees of freedom, and a significance level of p < 0.001. These findings demonstrate strong structural validity and affirm that the data are appropriate for principal component analysis (refer to Table 12). Principal components with eigenvalues over one were extracted, and the Varimax approach was employed to enhance factor differentiation. Two primary components were subsequently identified, with the cumulative variance explained achieving a reasonable threshold of at least 70%. Moreover, the commonality values for all Kansei word pairings exceeded 0.8, further demonstrating that the derived principal components adequately account for the variation in the original data. The factor structure was evident and exhibited great dependability.⁵¹

Table 12.

KMO and Bartlett’s Test.

KMO Measure of sampling adequacy		0.871
Bartlett’s test of sphericity	Approximate Chi-square	1098.622
	df	36
	p-Value	<0.001

According to the ordering of rotating factor loadings, Factor 1 was linked to adjective pairs such as “slim-fitting–loose-fitting” and “poorly layered–well-layered,” which pertain to the dimensions of garment construction and craftsmanship. Factor 2, denoted as “stiff–soft,” pertained to the material dimension. This outcome signifies that the two dimensions previously discerned via AHP weighting were precise. Consequently, lightweight outdoor apparel for urban travel must include the following unique characteristics: it should be loose-fitting, exhibit clear layering in garment structure and craftsmanship, and be composed of soft materials. The results of the PCA are presented in Table 13.

Table 13.

PCA results.

Component number	Eigenvalue	Individual variance explained (%)	Cumulative variance explained (%)	Variable name	Rotated component matrix		Communality
Component number	Eigenvalue	Individual variance explained (%)	Cumulative variance explained (%)	Variable name	Factor 1	Factor 2	Communality
1	4.430	49.224	49.224	Imbalanced—Balanced	0.924	0.155	0.878
2	3.601	40.008	89.232	Rough—Exquisite	0.912	0.178	0.863
3	0.215	2.384	91.615	Slim-fitting—Loose-fitting	0.930	0.194	0.902
4	0.205	2.278	93.894	Poorly layered—Well-layered	0.930	0.104	0.876
5	0.147	1.629	95.522	Lacking in detail—Rich in detail	−0.928	0.147	0.882
6	0.137	1.523	97.045	Heavy—Lightweight	−0.186	0.938	0.915
7	0.099	1.102	98.147	Stiff—Soft	−0.143	0.944	0.912
8	0.088	0.975	99.123	Prone to wear—Durable	−0.201	0.927	0.900
9	0.079	0.877	100.000	Tight—Comfortable	−0.243	0.918	0.902

Subsequently, the samples that exhibited optimal performance on the Kansei word pairs “slim-fitting–loose-fitting,” “poorly layered—well-layered,” and “stiff—soft” were chosen as reference objects for further improvement. A secondary analysis of the original data was conducted to derive user rating scores for each Kansei word pair across the five garment samples (refer to Table 14). Table 14 indicates that the garment sample with the highest score for the Kansei word pair “slim-fitting—loose-fitting” was S4, which had a score of 0.200 for loose-fitting. Sample S1 got the highest ratings for both “unclear layering—distinct layering” (distinct layering: 0.150) and “stiff—soft” (soft: 0.150). Due to Factor 1 accounting for the greatest proportion of variance, and S4 ranking highly in both Factors 1 and 2, samples S1 and S4 were chosen as the focal points for further design optimization. Given that other variables in Factors 1 and 2 exhibit substantial loadings, the three highest-ranked variables from each factor were selected as pivotal components for the ensuing design practice. The optimum design must have loose-fitting, well-layered, and proportionately balanced garment structures and craftsmanship, utilizing fabrics that are soft, lightweight, and durable.

Table 14.

Evaluation of Kansei word pairs for samples.

Sample ID	Imbalanced—Balanced	Rough—Exquisite	Slim-fitting—Loose-fitting	Poorly layered—well-layered	Lacking in detail—Rich in detail	Heavy—Lightweight	Stiff—Soft	Prone to wear—Durable	Tight—Comfortable
S1	−0.060	0.080	0.150	0.150	0.040	0.200	0.150	0.040	−0.040
S2	0.080	−0.180	0.090	−0.100	−0.080	0.090	−0.040	−0.050	0.300
S3	0.040	−0.090	−0.430	−0.290	−0.120	0.080	0.070	0.270	−0.010
S4	−0.300	−0.150	0.200	−0.200	−0.160	−0.210	0.080	−0.030	−0.020
S5	0.000	0.080	−0.200	0.090	0.100	−0.090	0.000	0.060	0.020

Generation of improved design solutions and user feedback results

In the improved design scheme for Lightweight Outdoor Apparel, this study employed the Midjourney platform to conduct Artificial Intelligence Generated Content (AIGC) design practice. During the generation process, samples S1 and S4 were used as visual references, and structured prompts were constructed based on the results of prior Analytic Hierarchy Process (AHP) and Principal Component Analysis (PCA) to reduce randomness in the generation process.

Among these, the Kansei words “loose-fitting,” “distinctly-layered,” “well-balanced,” “lightweight,” “soft-textured,” and “durable” were used as dominant factors to define the overall style of the apparel and the Kansei characteristics of user preferences. Meanwhile, the sub-objective indicators from the previous hierarchy (C1–C6, C11–C15) were employed as auxiliary factors to supplement design constraints related to specific apparel elements, craftsmanship, materials, and colors. These factors were then integrated into prompt language following a fixed logical sequence (see Table 15) to ensure the stability and consistency of the generated semantics.

Table 15.

Summary of prompt commands.

Leading factors	Secondary factors	Design theme	Design background, observation perspective, other parameters
loose-fitting, well-layered, balanced, lightweight, soft-textured, durable	High-Capacity Pocket Design, Detachable Design, Natural Colours, Eco-Friendly Colours . . .	Lightweight outdoor apparel design, outdoor apparel design, urban travel	Front perspective, 4k quality, realistic style, white background, –ar 1:1, -chaos 100, iw 3

Considering the inherent randomness of AIGC-generated outputs, multiple iterative generations were performed while maintaining consistency between the reference images and the prompt content. During the screening process, the improved design schemes were required to satisfy the following criteria: the garments should adopt a loose fit to ensure mobility and comfort during outdoor activities; craftsmanship should emphasize fine workmanship and clear structural details to enhance perceived quality. Lightweight and durable fabrics were selected to reduce travel burden while providing sufficient flexibility and adaptability. In addition, guided by prior Analytic Hierarchy Process indicators, colors were predominantly natural and eco-friendly, aligning with urban travel contexts. Based on these considerations and a comprehensive comparison, four designs exhibiting superior performance in both Kansei characteristics and design feasibility were selected as the final improved design outcomes (see Figure 5).³²

Figure 5.

Improved design proposals.

To verify whether the AHP–KE–based improved designs for lightweight outdoor apparel in urban travel truly address consumers’ core needs, this study adopted a Likert-scale survey and therefore administered it to 20 consumers drawn from different provinces, As illustrated in Figure 6, an example of the design evaluation is provided, while the corresponding results are presented in Table 16. For each improved design case, 4 specific questions were formulated. Q1: Are you satisfied with the design and structure of this lightweight outdoor apparel? Q2: Are you satisfied with the material of this lightweight outdoor apparel? Q3: Are you satisfied with the colors and patterns of this lightweight outdoor apparel? Q4: Does this lightweight outdoor apparel meet your needs? Because the mean score for every case exceeded the satisfaction benchmark of 1 on all four questions. These findings thus provide strong evidence for both the effectiveness and the feasibility of applying the present approach to the improved design of outdoor apparel.

Figure 6.

Example of design evaluation.

Table 16.

Design case evaluation.

Question	Case 1			Case 2			Case 3			Case 4
Question	Total score	Average score	Overall average score	Total score	Average score	Overall average score	Total score	Average score	Overall average score	Total score	Average score	Overall average score
Q1	29	1.450	1.162	23	1.150	1.262	24	1.200	1.337	26	1.300	1.412
Q2	18	0.900		20	1.000		28	1.400		35	1.750
Q3	24	1.200		32	1.600		19	0.950		32	1.600
Q4	22	1.100		26	1.300		36	1.800		20	1.000

Discussion

Research findings

By integrating the Analytic Hierarchy Process with a Kansei-engineering model and further leveraging AI-generated content technology, this study conducted an in-depth analysis of the hierarchical user requirements and Kansei characteristics of lightweight outdoor apparel for urban travel. Consequently, a series of improved design proposals were developed in line with the predefined design objectives. The key findings are as follows:

Clarification of requirement weights. By constructing a hierarchical requirement model with the AHP method, this study identified the principal user requirement indicators and their corresponding weight distribution for lightweight outdoor apparel in urban travel. More precisely, the results show that garment structure and craftsmanship (B1) carry the highest weight (0.4103), followed by material requirements (B3) at 0.3184, then usage requirements (B4) at 0.1679, and finally color and pattern requirements (B2) at 0.1034. Overall, the findings reveal that when selecting lightweight outdoor apparel for urban travel, consumers attach much greater importance to practicality and functionality, whereas, by contrast, color and pattern receive relatively limited attention.

Clear Kansei preferences. By applying the semantic-differential method in Kansei Engineering, the study screened and analyzed the Kansei words supplied by consumers and, in doing so, clarified their affective preferences for both garment structure and craftsmanship, and materials. Specifically, consumers tend to favor garments that are “loose-fitting,” “well-layered,” and “balanced” with respect to structure and craftsmanship; whereas, for materials, they prefer fabrics that are “lightweight,” “soft,” and “durable.” Consequently, these findings provide explicit design targets for the subsequent improvement phase.⁵²

Integrated approach and user validation. Combining the indicators identified above, the study generated three improved design proposals using the AIGC platform, which were then validated through user evaluation and feedback. The results show that these design proposals received high ratings in terms of structural suitability, material selection, and overall Kansei compatibility. This demonstrates that the integrated AHP–KE–AIGC design pathway is not only logically rigorous and practically feasible but also achieves strong user acceptance. Moreover, this approach effectively addresses the disconnect often found in traditional design processes between needs identification and design target transformation. As a result, it enhances both the scientific foundation and user alignment of apparel design, while expanding the application boundaries of Kansei Engineering in practical product development.³³

Advantages of the AHP–KE integrated model. By combining AHP and KE methods, this study established a comprehensive evaluation model. Notably, this model overcomes the subjectivity and limitations often associated with the application of AHP in the textile and apparel field, while the KE approach translates consumers’ Kansei needs into concrete design targets, thereby improving both the accuracy and efficiency of the design process. Furthermore, the integration of AIGC technology further enhances the level of innovation and personalization in design.

Research significance

Compared with previous research, this study achieved a breakthrough in both methodological integration and practical application. In particular, at the theoretical level: (1) Enrichment of user-driven design theory in apparel. This study proposed a user multi-level demand-oriented framework for apparel design analysis, systematically integrating both AHP and Kansei Engineering (KE) methods. Unlike conventional research pathways that focus primarily on functionality or material performance, the present study emphasizes the integration of user Kansei experience and real-world usage scenarios, thereby broadening the theoretical scope of apparel design research. (2) Establishment of the AHP–KE–AIGC integrated design paradigm. For the first time in the field of lightweight outdoor apparel design for urban travel, this study organically combined AHP, KE, and AIGC technologies, thereby establishing a closed-loop methodological chain of “requirement decomposition—Kansei mapping—intelligent generation—user feedback.” In this way, the proposed paradigm offers theoretical innovation for the functional apparel sector and provides significant theoretical guidance for the synergistic application of multi-criteria decision-making, Kansei Engineering, and artificial intelligence. (3) Expansion of the application boundaries of the AHP–KE method in the apparel field. Although the AHP–KE approach has been widely validated in areas such as furniture and product design, its systematic application in apparel design—especially for the emerging category of lightweight outdoor apparel for urban travel—remains limited. Therefore, this study provides both practical cases and theoretical foundations for the implementation of the AHP–KE method in apparel design.⁵³

At the practical level: (1) Provision of a scientific decision-making basis for lightweight outdoor apparel design practice. This study established a hierarchical index system for user requirements in lightweight outdoor apparel for urban travel, and, drawing on real user evaluations, extracted core design elements. Consequently, the findings offer quantitative references for designers and companies to optimize product functions, structure, and materials, thereby improving the alignment between products and market demand. (2) Promotion of intelligent design tools in the apparel field. By introducing AIGC (AI-generated content) into the design process, this study enabled the rapid generation and diversified expression of design solutions, which in turn significantly improved both efficiency and innovation in apparel design. Therefore, the approach provides effective technical support for the apparel industry to better respond to rapidly changing market demands. (3) Facilitation of user-centered innovation in apparel products. By placing user requirements and Kansei preferences at the core, this study advances the integrated innovation of functionality and effective experience. As a result, it offers both theoretical guidance and methodological reference for future apparel development in urban travel and other emerging scenarios, thus providing substantial value for industry adoption. (4) Enhancement of market competitiveness and user satisfaction in the apparel industry. By scientifically identifying and translating core user requirements, this study has promoted more precise positioning and innovative upgrading of lightweight outdoor apparel products. Accordingly, these advancements help strengthen brand competitiveness and increase consumer satisfaction, thereby supporting the high-quality development of the apparel industry.^54,55

Limitations and future perspectives

Nevertheless, this study has certain limitations. First, the research primarily targets young urban consumers of lightweight outdoor apparel in China, a group that is relatively representative within the Chinese market. Given the differences in outdoor lifestyles, esthetic preferences, and clothing size systems across countries, the findings are mainly applicable to the Chinese context, and their generalizability to other cultural settings and consumer groups requires further investigation.

Second, this study adopted Midjourney as the AIGC-based design generation tool, primarily due to its comprehensive advantages in visual style expression and rapid multi-solution generation. However, the platform remains constrained in terms of fine-grained design control and structural accuracy, as its outputs are influenced by prompt construction and algorithmic optimization mechanisms, which may limit the precise representation of complex functional structures. Future research may integrate multi-platform generation technologies or incorporate parametric design methods to further enhance design controllability and professional adaptability.

Therefore, future research is needed to further validate and refine this integrated pathway across diverse user groups and multi-scenario design demands, thereby promoting the continued advancement of user requirement translation and intelligent design generation methods.

Conclusion

This study focused on the core user requirements for lightweight outdoor apparel in urban travel and established a systematic design pathway that integrates the Analytic Hierarchy Process (AHP), Kansei Engineering (KE), and AI-generated content (AIGC) technologies. Through this approach, the research explored the integration of functional criteria and user Kansei preferences in apparel design and verified the effectiveness of this pathway in design optimization.

First, by applying the AHP method, this study systematically established a hierarchical index system for user requirements in lightweight outdoor apparel for urban travel. In this way, the priorities of structure and craftsmanship, materials, and usage requirements in user decision-making were clarified, thereby providing a quantitative basis for subsequent design work. Second, with the aid of semantic differential scales and principal component analysis (PCA) within the KE approach, key Kansei preference features such as “well-layered,” “loose-fitting,” “balanced,” “lightweight,” “soft,” and “durable” were identified. Accordingly, the mapping between users’ subjective kansei characteristics and design elements was established. Subsequently, by incorporating AIGC technology, three improved design proposals were generated on the Midjourney platform based on multimodal prompts. Four cases were then evaluated through user questionnaires, which verified their performance in terms of structure, materials, and overall style. The results showed that all four cases achieved average scores above the benchmark across core indicators, thereby confirming the effectiveness of this pathway in meeting user requirements and addressing Kansei preferences. This study not only proposed an innovative AHP–KE–AIGC integrated pathway at the methodological level but also demonstrated a user-oriented product development process in design practice.

Future research may be further extended in the following aspects. First, the sample size can be expanded by incorporating more diverse population data and extending cross-regional design studies across multicultural contexts, thereby promoting the broader application of the methodology in international markets and enhancing the generalizability of the model. Second, the precision of AIGC in representing design details and structural features can be improved, while exploring its integration with parametric design and multi-platform collaborative generation technologies, in order to further refine the integrated framework of user requirement translation and intelligent design generation.

Footnotes

ORCID iD

Qian Qiu

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is funded by the Macau Foundation, Project No. I00242-2308-190.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Ahmad

Akhtar

Anam

, et al. Recent developments in materials and manufacturing techniques used for sports textiles. Int J Polym Sci 2023; 2023(1): 2021622.

Chen

Naylor

, et al. Festival or sport? Chinese motivations to a modern urban hiking event. Sage Open 2021; 11(1): 2158244020986155.

Arumugam

Tok

AIY

Lipik

A novel method to produce sustainable wind resistant and water repellant fabric for outdoor sport clothing. J Ind Text 2023; 53: 15280837231184256.

Wang

Liu

Clothing adjustment in outdoor environment: A new clothing model based on temperature change. Build Environ 2021; 206: 108395.

Burtscher

Kofler

Gatterer

, et al. Effects of lightweight outdoor clothing on the prevention of hypothermia during low-intensity exercise in the cold. Clin J Sport Med 2012; 22(6): 505–507.

Lin

Wang

Chen

, et al. Using AHP and TOPSIS approaches in customer-driven product design process. Comput Ind 2008; 59(1): 17–31.

Wen

Lee

SW.

Study on the demand of hardshell jacket users based on KANO-AHP model. Korea Inst Des Res Soc 2024; 9(4): 135–150.

Yılmaz Kaya

Dağdeviren

. Selecting occupational safety equipment by MCDM approach considering universal design principles. Hum Factors Ergon Manuf Serv Ind 2016; 26(2): 224–242.

Gao

Wang

, et al. Research on the aging-suitability of community outdoor spaces in Zhengzhou based on the analytic hierarchy process and semantic differential method. Front Public Health 2024; 12: 1452120.

10.

Morrissey

Rossi

RM.

Clothing systems for outdoor activities. Text Prog 2013; 45(2-3): 145–181.

11.

Apichonbancha

Lin

Chuang

CL.

Integration of principal component analysis with AHP-QFD for improved product design decision-making. Appl Sci 2024; 14(14): 5976.

12.

Liu

Huang

Zhang

, et al. Thermal, waterproof, breathable, and antibacterial cloth with a nanoporous structure. ACS Appl Mater Interfaces 2018; 10(2): 2026–2032.

13.

Zhang

Wan

Zheng

, et al. A new model for evaluating dynamic clothing thermal comfort. Int J Therm Sci 2025; 215: 109946.

14.

Lee

Baek

Developing a smart multifunctional outdoor jacket with wearable sensing technology for user health and safety. Multimed Tools Appl 2021; 80(21–23): 32273–32310.

15.

Chen

Chuang

, et al. A Hybrid MADM model for product design evaluation and improvement. Sustainability 2020; 12(17): 6743.

16.

Lin

Application of fuzzy delphi method (FDM) and fuzzy analytic hierarchy process (FAHP) to criteria weights for fashion design scheme evaluation. Int J Clothing Sci Technol 2013; 25(3): 171–183.

17.

Lin

Jerusalem

MA.

Integrated MCDM for evaluating fashion design schemes. Int J Clothing Sci Technol 2016; 28(6): 880–892.

18.

Marjanovic-Halburd

Nasiri

, et al. Assessment of building-integrated green technologies: a review and case study on applications of multi-criteria Decision Making (MCDM) method. Sustain Cities Soc 2016; 27: 106–115.

19.

Yeh

Pai

Liao

CW.

Using a hybrid MCDM methodology to identify critical factors in new product development. Neural Comput Appl 2014; 24(3–4): 957–971.

20.

Liu

Zhao

Zhu

Research on digital restoration of plain unlined silk gauze gown of Mawangdui Han dynasty tomb based on AHP and human–computer interaction technology. Sustainability 2022; 14(14): 8713.

21.

Xue

Jin

Jiang

Design and evaluation of a novel interactive firefighter clothing with multiple functionalities. Text Res J 2023; 93(1-2): 221–241.

22.

Desai

Bidanda

Lovell

MR.

Material and process selection in product design using decision-making technique (AHP). Eur J Ind Eng 2012; 6(3): 322–346.

23.

López

Murillo

González

. Systematic literature reviews in Kansei engineering for product design—a comparative study from 1995 to 2020. Sensors 2021; 21(19): 6532.

24.

Guo

Nagamachi

, et al. A proposal of the event-related potential method to effectively identify kansei words for assessing product design features in kansei engineering research. Int J Ind Ergon 2020; 76: 102940.

25.

Lei

Zhu

, et al. Wooden furniture design based on physiological-psychological measurement technology and Kansei engineering: taking Ming-style chair as an example. BioResources 2024; 19(3): 6304–6324.

26.

Liu

Chen

, et al. An improved Kansei engineering method based on the mining of online product reviews. Alex Eng J 2023; 65: 797–808.

27.

Jiang

Gan

Hong

, et al. Application of Kansei engineering in the innovative design of traditional fashion elements. Ind TEXTILA 2024; 75(3): 289–301.

28.

AI-driven computational creativity in fashion design: a review. Text Res J 2025; 95(5-6): 658–675.

29.

Yang

Jin

Analyzing the adoption of AIGC tools in fashion design: an S-O-R framework integrating task-technology fit. PLoS One 2025; 20(10): e0335522.

30.

Jin

Yoon

Self

, et al. Understanding fashion designers’ behavior using generative AI for early-stage concept ideation and revision. Arch Des Res 2024; 37: 25–45.

31.

Wang

Tao

, et al. EEG assessment of artificial intelligence-generated content impact on student creative performance and neurophysiological states in product design. Front Psychol 2025; 16: 1508383.

32.

Zhang

Liu

Unlocking the potential of artificial intelligence in fashion design and e-commerce applications: the case of midjourney. J Theor Appl Electron Commer Res 2024; 19(1): 654–670.

33.

Hsiao

An AIGC-empowered methodology to product color matching design. Displays 2024; 81: 102623.

34.

Kim

SJ.

Generative artificial intelligence in collaborative ideation: Educational insight from fashion students. IEEE Access 2024; 12: 49261–49274.

35.

Xiong

Wei

Liang

, et al. Research on design evaluation method for green packaging recycling. In: Int Conf Mod Educ Inf Manag ICMEIM. 2020, pp.470–474.

36.

Sihombing

Ahmad

MMHM

Yuhazri

, et al. The product design preferences towards personality style. Int J Public Sect Perform Manag 2024; 14(3/4): 447–464.

37.

Lin

Jiang

Deng

, et al. Comparing AIGC and traditional idea generation methods: evaluating their impact on creativity in the product design ideation phase. Think Skills Creat 2024; 54: 101649.

38.

Yin

Zhang

Liu

The exploration of integrating the midjourney artificial intelligence generated content tool into design systems to direct designers towards future-oriented innovation. Systems 2023; 11(12): 566.

39.

Zhou

Shi

Outdoor clothing design for traffic safety based on big data and artificial intelligence. J Adv Transport 2022; 2022: 8697421–8697513.

40.

Zhou

, et al. Research on the design of accompanying products for the mentally retarded elderly based on AHP. Math Probl Eng 2022; 2022: 1–12.

41.

Sun

Yang

Evaluation and design of reusable takeaway containers based on the AHP–FCE model. Sustainability 2023; 15(3): 2191.

42.

Xie

Zhu

Ding

, et al. AHP and GCA combined approach to green design evaluation of kindergarten furniture. Sustainability 2024; 16(1): 1.

43.

Giordano

Primerano

The use of network analysis to handle semantic differential data. Qual Quant 2018; 52(3): 1173–1192.

44.

Zheng

Comparing AI and human emotional responses to color: a semantic differential and word-color association approach. Color Res Appl 2025; 50: 286–300.

45.

Gong

Guo

Xie

Using kansei engineering for the design thinking framework: Bamboo pen holder product design. Sustainability 2022; 14(17): 10556.

46.

Cash

Isaksson

Maier

, et al. Sampling in design research: Eight key considerations. Des Stud 2022; 78: 101077.

47.

Anjani

Kühne

Naddeo

, et al. PCQ: preferred comfort questionnaires for product design. WORK J Prev Assess Rehabil 2021; 68: S19–S28.

48.

Huang

Chen

Khoo

LP.

Products classification in emotional design using a basic-emotion based semantic differential method. Int J Ind Ergon 2012; 42(6): 569–580.

49.

Hsiao

Tang

, et al. A generative-AI-based design methodology for car frontal forms design. Adv Eng Inform 2024; 62: 102835.

50.

Ren

Guo

, et al. A consumer-oriented Kansei evaluation model through online product reviews. J Intell Fuzzy Syst 2023; 45(6): 10997–11012.

51.

Temnitskiy

AL.

How to teach a sociology student good data analysis using SPSS?

Sotsiol Issled 2024; (9): 159–168.

52.

Zhagn

Tang dynasty chair feature design based on Kansei evaluation and eye tracking system. WOOD Res 2020; 65(1): 161–174.

53.

Liu

Cheng

Liao

, et al. Research on harmonious design of chairs based on the Kano model and analytic hierarchy process. BioResources 2024; 19(3): 5535–5548. https://doi.org/10.15376/biores.19.3.5535-5548

54.

Wang

Jin

, et al. Identifying popular products at an early stage of sales season for apparel industry. INFORMS J Appl Anal 2024; 54(3): 282–296.

55.

Vieira

Osório JMA, Mouta

, et al. Kansei engineering as a tool for the design of in-vehicle rubber keypads. Appl Ergon 2017; 61: 1–11.