Interpretative Structural Modeling Analyzes the Hierarchical Relationship between Mid-Air Gestures and Interaction Satisfaction

Abstract

Mid-air gestures as a new form of human–computer interaction has a wide range of satisfaction factors, for which the primary and secondary relationships and hierarchical relationships between factors are unclear. By examining usability definitions, collecting satisfaction questionnaires and user interviews, 30 observed variables were obtained and a scale was developed. A total of 310 valid questionnaires were collected, and six latent variables were summarized through factor analysis. The matrix quantitative analysis of latent variables based on interpretative structural model theory was used to construct a hierarchical model of influencing factors of satisfaction with mid-air gestures. The study shows that the influencing factors of mid-air gesture satisfaction can be divided into three levels. The first layer of attractiveness is the direct influencing factor on the surface and the goal of mid-air gesture design. In the second layer, Simplicity and Efficiency, Simplicity and Tiredness, and Tiredness and Friendliness interact with each other. Simplicity positively affects Friendliness, and Efficiency positively affects Tiredness. The third layer, Intuitiveness is the root layer influencing factor, which affects Simplicity. This study provides a theoretical basis for the design of mid-air gesture so that it can be designed and selected more objectively.

1. Introduction

Mid-air gestures have become more popular over the past decade as technology has evolved [1]. Mid-air gesture interaction is often considered to be the next generation of computer mouse [2,3], as future interactions are multifaceted. For humans, the hand is the general state-of-the-art controller [4,5]. Moreover, as an innate human skill, gesturing places a much lower cognitive burden on users [6]. Gesturing is a natural and intuitive form of interpersonal communication [7,8]. The advantage of gesture interaction is that the use of electronic devices is no longer limited to operating through contact screens, mice, or keyboards but can be completely removed from the operating medium.

Osen et al. divided human–computer interaction gestures into two categories: touch-screen gestures and mid-air gestures [9]. For a clear understanding, the mid-air gesture used in this article is shown in Figure 1, where the user is at a distance from the operating interface and interacts using the dynamic or static attitude of the hand. Mid-air gestures are not limited to the virtual operation interface or the actual screen. Touch-screen gesture interplay provides a simple and convenient interplay with contact devices, mostly because we have developed conventions on how to perform gestures. Many people have tried to design specifications for gestures [10,11]; however, to date, no clear design guidelines for how to design gesture-based interactions have yet been developed [12]. In contrast to touch interaction, which requires contact manipulation by way of customers and devices, and voice interaction, which requires a particular manner of listening, speaking, and high-precision focus [13], “gesture interaction” has the natural advantage of habitual human use. The input mode of gesture recognition is considered the most natural input and interaction mode [14]. Thus, it will be possible for this approach to become the most advantageous solution beyond the inconveniences of touch interaction and speech interaction. In contexts such as mixed reality [15], virtual reality [16], television [17], tablet computers [18], and smart homes [19], mid-air gestures are the most frequently explored and applied fields. Better-known products, such as Microsoft Kinect, Leap Motion, Microsoft HoloLens glasses, and Meta Quest Pro enable users to manipulate a blank interface, and they provide good insight into the future of human–computer interaction.

Technology and interaction design are interdependent. The development of mid-air gesture technology began with data gloves and has since progressed to include RGB cameras, radar, myoelectric sensors, 3D Time of Flight (ToF), and other technologies. Due to advancements in sensor and hardware technologies, mid-air gesture recognition has greatly improved in terms of recognition types and rates. In an ideal scenario, users could achieve human–computer interaction through mid-air gestures at any time, but this ideal state has not yet been reached, due to current technological limitations. The usability of gestures has become a major concern for researchers and designers [20]. Satisfaction is an important aspect of usability. To understand the relationship between gesture satisfaction factors in order to improve the usability of gestures and to formulate a theoretical basis for gesture design rules, we conducted two important studies. The focus was to extract the primary and secondary relationships and hierarchical relationships between gesture satisfaction factors and explicit factors. After employing three methodologies to gather factors that influence satisfaction in various articles, various and complex interaction methods using gestures and many factors involved in satisfaction were found, many of which are difficult to quantify directly. By conducting factor analysis to reduce the dimensionality of data, a small number of factors can be identified to explain most of the variance observed in the many observed variables [21]. Moreover, there are interactions and mutual influences among factors. Accordingly, interpretative structural modeling (ISM) theory was selected to study the influencing factors of gesture satisfaction. Therefore, the explanatory structure model was considered an appropriate method for studying the relationship between factors that influence mid-air gesture satisfaction [22]. Based on matrix quantitative analysis (as part of ISM theory), a structural interpretation model of gesture satisfaction was constructed. The influencing factors of gesture satisfaction were divided into three levels. According to the quantitative and qualitative research results obtained from this study, the theoretical basis of high-satisfaction gestures is summarized, which provides guidance for the design of gestures to ensure that gestures can be designed and selected more objectively.

2. Extraction of Related Factors

Previous studies have shown that air gestures are a new form of interaction, and the design and selection of gestures are subjective. Therefore, it is first necessary to first determine the factors that affect the satisfaction of air gestures, and then analyze the impact of each factor’s relationship. Ensuring the objectivity and comprehensiveness of the set of factors that influence satisfaction requires the utilization of multiple approaches, as a single approach alone is clearly insufficient. This study mainly used three methods to determine the satisfaction factors of air gesture interaction: previous questionnaires, usability definitions, and user interviews.

2.1. Availability Satisfaction Questionnaire Factor

The advent of the mouse and the graphical user interface in the mid-1980s, which opened up the computing world to the public, led to an increasing appreciation for usability evaluation. Interaction design has evolved over the past 30 years [23] and has been rapidly promoted and recognized by academia and society with the rapid development of mobile internet access. The ISO/IEC 9126 [24] usability definition has also been revised and refined, and user satisfaction now includes new usage scenarios as an important component of usability. User satisfaction varies across different usage scenarios and types, thus requiring a multidimensional examination. Currently, academia has produced numerous relevant research findings in this domain. By analyzing and synthesizing the existing literature through a comprehensive review of previous studies, we have gained a deeper understanding of prior research in this area and have identified notable variations in the literature. A total of 19 scales (

Table 1. Summary of questionnaire usability factors.

Survey Name	Developer	Number of Factors	Factor	Items	Reliability
SMEQ [25]	Zijlstra and Doorn (1985)	1	Task difficulty	150	Unreported
QUIS [26]	Chin, J. et al. (1988)	5	Overall Reaction, Screen terminology, System information, Learning, System capability	27	0.94
TAM [27]	Davis (1989)	2	Usefulness, Ease of use	12	0.98
ASQ [28]	Lewis (1991)	3	Task difficulty, Efficiency, Helpfulness	3	0.93
PSSUQ [29]	Lewis (1992)	3	System usefulness, Information quality, Interface quality	16	0.94
CSUQ [30]	Lewis (1995)	3	System quality, Information quality, Interface quality	19	0.95
SUMI [31]	J Kirakowski (1996)	5	Efficiency, Affect, Helpfulness, Control, Learnability	50	0.92
SUS [32]	Brooke (1996)	2	Usability, Learnability	10	0.92
WAMMI [33]	Kirakowski et al. (1998)	5	Attractiveness, Controllability, Efficiency, Helpfulness, Ease of learning	20	0.96
USE [34]	Lund (2001)	4	Usefulness, Ease of use, Ease of learning, Satisfaction	30	Unreported
NPS [35]	Reichheld (2003)	1	Willingness to recommend	1	Unreported
ER [36]	Albert and Dixon (2003)	1	Task difficulty	2	Unreported
UME [37]	McGee and Rich (2004)	1	Task difficulty	1	Unreported
UEQ [38]	Laugwitz et al. (2008)	6	Attractiveness, Perspicuity, Efficiency, Dependability, Stimulation, Novelty	26	0.8
SEQ [39]	Sauro and Dumas (2009)	1	Task difficulty	1	Unreported
UMUX [40]	Finstad (2010)	3	Effectiveness, Efficiency, Satisfaction	4	0.94
CES [41]	Dixon et al. (2010)	1	Task difficulty	1	Unreported
UMUX-LITE [42]	Lewis et al. (2013)	2	Usefulness, Ease of use	2	0.82
SUPRQ [43]	Sauro (2015)	4	Usability, Credibility, Appearance, Loyalty	10	0.86

) have been reviewed for this study.

2.2. Usability Definition Factors

In addition to the usability questionnaire, measurement factors are also important for the formulation and classification of questionnaires. Different researchers have different trade-offs for elements within usability definitions. This article collates six of the most mainstream usability definitions currently available, and

Table 2. Summary of usability definition factors.

Shackel [45]	Rengger [46]	ISO/IEC 9126 [49]	Nielsen [48]	ISO 9241-11 [44]	Parikh et al. [47]
Effectiveness	Effectiveness	Efficiency	Learnability	Effectiveness	Learnability
Learnability	Learnability	Functionality	Efficiency	Efficiency	Efficiency
Flexibility	Efficiency	Reliability	Errors	Satisfaction	Satisfaction
Attitude	Errors	Usability	Satisfaction
		Maintainability	Maintainability
		Portability

contains factors for each usability definition. There are three factors in the definition of ISO 9241-11 [44]; four factors in Shackel [45], Rengger [46], and Parikh et al. [47]; five factors in Nielsen [48]; and six factors in ISO/IEC 9126-1 [49]. All definitions contain a total of 14 factors.

2.3. User Interview Factors

Interviews are a commonly used method of user research that have a clear advantage in obtaining unique information and opinions on the research environment [50]. No matter how carefully we organize our language, the language itself contains many uncertainties and ambiguities. However, the goal of interaction design is to build a good interaction experience for the user and to design with the user in mind [51], so it is important to ensure that the user is involved in the process of collecting factors. Therefore, user interviews are not used as a unique method of collecting data. By using unstructured interviews [52], the interviewee is given the topic of satisfaction of mid-air gestures, and the interviewer is able to collect more objective variables as the interviewee is able to talk freely about the topic. For interviews to be smooth and the data collected to be reliable, the interviewee needs a certain amount of knowledge of mid-air gesture interaction and therefore must have experience using mid-air gestures.

The purpose of the interview is to explore the factors that prospective users perceive as influencing mid-air gesture interaction satisfaction and to complement the construction of a set of factors affecting satisfaction. As stated in Section 1, mid-air gestures are employed as an input method in a myriad of products, each featuring distinct hardware, functionality, and gesture posture. The methodology employed in this study involves allowing participants to select products freely, without any restrictions, as a means of comprehensively capturing a broad range of variables. At present, there is no consensus on the number of participants required for user interviews. Maria [53] calculated that the median number of samples found in more than 2000 qualitative interview doctoral dissertations is 31. Therefore, in this study, 31 people were recruited to take part in user interviews. The sample comprised 15 males and 16 females, aged 18 to 40 years. The interviews were conducted one-on-one, with each interview session lasting for approximately 20 min. Audio recording equipment was employed to capture the entire interview, while notes were taken to document the interviewer’s perspective. After all the interviews were completed, the notes from the auxiliary interviews were used to interpret and mark the language of the interviewees. Variables mentioned by all interviewees were then aggregated.

Table 3. Summary of influencing factors from user interviews.

User Interviews
Fatigue	Easy to remember	Experience
Feedback	Recognizable	High tolerance
Convenient	Fluency	Operating normally
Ease of use	Improve	Interface
Responsiveness	Usefulness	Effect
Natural	Mis-operation	Easy to trigger
Coherence between gestures	Everyday behavior	Quick identification
Customize	Habit	Steerable
Metaphor	Simple	Timely feedback
Interesting	Not strenuous	Similar to touch
Can do a lot of things	No bad meaning	Teach
Accuracy	Considers people with disabilities	High success rate
Distraction	Entertainment	Cool gestures

presents the screened results, comprising 39 factors.

2.4. Summary of Satisfaction Factors

This study employed three data collection methods, namely questionnaire surveys, definition summarization, and user interviews, to gather satisfaction factors associated with mid-air gestures. These factors were then analyzed for further exploration. The factors obtained from these three methods guaranteed the comprehensiveness of the factors. The factors collected are somewhat comprehensive overall but also overlap each other somewhat. Upon analysis, it becomes apparent that certain factors exhibit behavior towards achieving a common objective, and some factors partially overlapped or were similar in meaning. By filtering and induction, the final set of factors was completed as shown in

Table 4. Factors Affecting Interaction Satisfaction of Mid-air Gesture.

S/N	Factor	S/N	Factor	S/N	Factor
1	Fatigue	11	Ease of remembering	21	Easy to identify
2	Easy to trigger	12	Tolerance	22	Attractiveness
3	Novelty	13	Flexibility	23	Accuracy
4	Fast recognition	14	Interesting	24	High success rate
5	Interface quality	15	Use habits	25	Helpful
6	Learnability	16	Usefulness	26	Fluency
7	Generality	17	Gesture continuity	27	Strenuous
8	With operational feedback	18	Mis-operation	28	Cultural meaning
9	Metaphorical	19	Operate naturally	29	Have guidance
10	Universal	20	Comfort	30	Customize

. In this paper, the factors (N = 30) affecting the satisfaction of mid-air gestures are selected.

3. Methodology

3.1. Subjects

A questionnaire containing 30 questions was developed based on 30 factors that affect the satisfaction of gestures in open space. Factor analysis requires a large sample size: Rummel [54] recommended a minimum of 100, Guilford [55] argued for at least 200, Cattell [56] found the minimum desirable number to be 250 (believing the sample size should range from three to six times the number of questions). Gorsuch [57] recommended at least five times the number of questions, and Everitt [58] recommended 10 times to be the optimum. The questionnaire used in this study comprised 30 questions. A total of 349 completed questionnaires were recovered, from which repeated answers and missing data were deleted, resulting in 310 valid questionnaires being retained, with a response rate of 88.8%. The sample size met the requirements for factor analysis [59,60].

Table 5. Demographic profile for respondents (N = 310).

Category	Item	Frequency	%
Gender	Male	147	47.42
	Female	163	52.58
Age	20~24	52	16.77
	25~30	213	68.7
	31~35	34	10.97
	36~45	11	3.56
Experience using mid-air gestures	Used	155	50
	Have not used	155	50
Sum		310	100

lists the demographic characteristics of respondents using data in the final analysis. The proportion of male and female respondents was 47.42% and 52.58%. The proportion of respondents aged 25~30 was the largest at 68.7%, followed by those aged 20~24 (16.77%) and those aged 31~35 (10.97%). An equal number of people said they had/had not used mid-air gestures. Mid-air gestures can be classified into static and dynamic [61]. Even for the same gesture, differences in hardware can result in varying user experiences. The diversity and complexity of gesture types can also influence the data collected. To address this issue, the most frequently used gesture types in existing products were identified, and respondents who had used mid-air gestures were limited to seven types of gestures, as illustrated in Figure 2.

3.2. Measures

The questionnaire in this study used a 7-point Likert scale [62] (where 1 = Strongly disagree, 2 = Disagree, 3 = A little disagreement, 4 = Undecided, 5 = A little agreement, and 6 = Agree, 7 = Strongly agree), to indicate the subject’s response to each factor. Exploratory factor analysis [63] was used to verify the validity of the scale. The aggregation validity and fitting validity of the scale were evaluated using confirmatory factor analysis. Finally, by using interpretive structural modelling [64], the disorganized, irregular, and complex relationships between the various elements of empty gesture satisfaction were decomposed into a clear multilevel structural model after regionalization and cascading. Then, the influential relationships between the internal factors of the system were finally analyzed.

4. Results

4.1. Confirmatory Factor Analysis Results

SPSS software was used to analyze the data from 310 questionnaires. The overall reliability and validity of the 30-item questionnaire were analyzed. Bartlett’s test for sphericity (p = 0.000 < 0.05) and the Kaiser–Meyer–Olkin (KMO) measure value (=0.982 > 0.6) were calculated, showing that the sample size was adequate [65], as shown in

Table 6. KMO and Bartlett’s tests.

KMO and Bartlett’s Test
KM0		0.982
	Approx. Chi-Square	8088.676
Bartlett’s Test of Sphericity	df	435.000
	p	0.000

. All values satisfied the minimum acceptable score proposed by Kaiser [66]. For correlation between variables, factor analysis was considered valid and appropriate. The results indicated that the study data were suitable for factor analysis [67].

Confirmatory factor analysis was carried out to test the convergent validity, correlation, discriminant validity, and fit validity of each item in the scale proposed in this paper. Confirmatory factory analysis is a research method used to test whether the relationship between a factor and the corresponding measure item conforms to the theoretical relationship designed by the researcher and can be used for the scale analysis of questionnaires. Standard load coefficient values are commonly used to represent the correlation between factors and analysis items. The criteria proposed by Fornell and Larcker [68] were used to assess the convergence effectiveness of measurements. Their guidelines suggest a strong correlation when the statistical significance level p < 0.05 is greater than 0.6 for all standardized factor loads for calculated observation variables. In this study, the square root value of average variance extracted (AVE) was used to test the discriminant validity among factors, as shown in

Table 7. Load factor table.

Variable		Std. Estimate	AVE	CR
Standard value		>0.6	>0.5	>0.7
Simplicity	a1 Learnability	0.824	0.607	0.903
	a2 Ease of remembering	0.763
	a3 Usefulness	0.788
	a4 Generality	0.732
	a5 Continuity	0.788
	a6 Universal	0.776
Friendliness	b1 Tolerance	0.738	0.646	0.903
	b2 Mis-operation	0.802
	b3 Guidance	0.831
	b4 Operational feedback	0.823
	b5 Controllability	0.801
	b6 Extensibility	0.8
	b7 Helpful	0.828
Efficiency	c1 Easy to identify	0.808	0.651	0.903
	c2 Easy to trigger	0.79
	c3 Accuracy	0.803
	c4 Fluency	0.82
	c5 Fast recognition	0.814
	c6 High success rate	0.806
Intuitiveness	d1 Operate naturally	0.783	0.59	0.903
	d2 Metaphorical	0.768
	d3 Cultural meaning	0.724
	d4 Use habits	0.796
Attractiveness	e1 Interface quality	0.819	0.635	0.903
	e2 Customize	0.814
	e3 Interesting	0.769
	e4 Novelty	0.785
Tiredness	f1 Fatigue	0.838	0.742	0.903
	f2 Strenuous	0.872
	f3 Comfort	0.874

. The AVE and composite reliability (CR) values should be greater than 0.5 and 0.7, respectively. Table 7 is the result of validation analysis, and it is clear this scale has good convergence validity.

The model fit index was used for the overall model fit case. When evaluating the model fit of a measure (

Table 8. Results of model fit analysis.

Model Fit Index	x2/df	RMSEA	NFI	NNFI	IFI	TLI	CFI
Standard value	<3	<0.10	>0.9	>0.9	>0.9	>0.9	>0.9
value	1.772	0.050	0.918	0.958	0.962	0.958	0.962

), the chi-square ($x^2=1.772<3$), statistics and normalized fit index (NFI = 0.918 > 0.9), (NNFI = 0.985 > 0.9), comparison fit index (CFI = 0.962 > 0.962), Tucker–Lewis index (TLI = 0.958 > 0.9), incremental fit index (IFI = 0.962 > 0.9), and RMSEA were calculated. All indices met the minimum acceptable values proposed in the previous literature [69,70].

In this study, Mplus software was used to analyze the factors. The first order (Simplicity, Efficiency, Friendliness, Intuitiveness, Attractiveness, and Tiredness) construction of second-order constructions was measured at the level of statistical significance. The validation factor analysis results in Table 8 confirm that the standardized factor loads of the six first-order constructions were all greater than 0.6 at the significance level p < 0.05. Paths show significant relationships with p-values between factors less than 0.05. All assumptions are valid, and the relationship between the factors is shown in Figure 3.

4.2. Interpretative Structural Modeling

The ISM approach, which reveals intrinsic logical relationships between influencing factors, has been used in supply chain resilience research in several other areas. Previous studies have shown a relationship between six elements. Any system S consists of two or more interlinked, interacting elements $(S_{1 },S_2,\dots S_n)$, resulting in an organic whole. Through the previous analysis, we can see that there are six factors influencing the satisfaction of mid-air gesture, so six factors ($S_{1 },S_2,\dots S_6$) are selected to establish the set of influencing factors S:

$$S=\left\{S_1,S_2,\right.\cdots S_n|2\le n\le \left.6\right\}=\left\{\mathrm{Simplicity}, \mathrm{Efficiency}, \mathrm{L}, \mathrm{Tiredness}\right\}$$

(1)

Thereafter, we further discussed with experts how to study a reasonable structural relationship and quantify the relationship between factors affecting gestures. It was proposed to use the form of adjacency matrix A to represent the mutual correlation between them, which is defined in the following way:

$$A={(a_{ij})}_{ n\times n}$$

(2)

$$\alpha { }_{ij}=\left\{\begin{array}{c}1, S_{i}R{S}_j\\ 0, S_i\overline{R}{S}_j\end{array}\right.$$

(3)

where 1 indicates that there is a path between one element and another and 0 indicates that there is no path between one element and another. Thus, it is possible to establish an adjacency matrix based on Equations (2) and (3) for the factors influencing the satisfaction of air gestures:

$$A_{6\times 6}=\left[\begin{array}{c}0 1 1\\ 1 0 0\\ 0 0 0\\ 1 0 0\\ 0 0 0\\ 1 0 1\end{array}\right. \left.\begin{array}{c}0 1 1\\ 0 0 1\\ 0 0 1\\ 0 0 0\\ 0 0 0\\ 0 0 0\end{array}\right]$$

The reachability matrix refers to any transitive binary relationship between factors, and a square matrix that represents the relationship between two nodes that can be reached through any path length in a directed graph. Let I be a unit matrix of the same order as A. By the neighborhood matrix A, adopt the Boolean algebra algorithm [71] (i.e., 0 + 0 = 0, 0 + 1 = 1, 1 + 0 = 1, 1 + 1 = 1, 0 × 0 = 0, 0 × 1 = 0, 1 × 1 = 1). The corresponding reachable matrix M is obtained as follows:

$$M={(A+I)}^r$$

(4)

The reachable matrix M is obtained according to Equation (4):

$$M=\left[\begin{array}{c}1 1 1\\ 1 1 1\\ 1 1 1\\ 1 1 1\\ 0 0 0\\ 1 1 1\end{array}\right. \left.\begin{array}{c}0 1 1\\ 0 1 1\\ 0 1 1\\ 1 1 1\\ 0 1 0\\ 0 1 1\end{array}\right]$$

Based on the reachable matrix M, it is possible to calculate the reachable and antecedent sets and the intersection of each influencing factor and obtain the hierarchical results, as shown below:

$$L1=\left\{S_5\right\}$$

$$L2=\left\{S_1,S_2,S_3,S_6\right\}$$

$$L3=\left\{S_4\right\}$$

Accordingly, it is possible to map out the explanatory structure model of the factors influencing the gesture satisfaction, as shown in Figure 4:

Through ISM model analysis, it was found that six factors can be classified into three hierarchical relationships. The first layer is Attractiveness, which is the superficial influencing factor and most important aspect for users at this stage when interacting with gestures. Attractiveness is the goal of high-satisfaction mid-air gestures and is a direct factor of satisfaction. The second comprises Simplicity, Efficiency, Friendliness, and Tiredness. Among them, Simplicity and Efficiency, Simplicity and Tiredness, and Tiredness and Friendliness interact with each other. Simplicity positively affects Friendliness, and Efficiency positively affects Tiredness. The third layer, Intuitiveness, is the root layer influencing factor, which affects Simplicity. Other influencing factors must pass through Intuitiveness to have an impact on mid-air gesture satisfaction.

5. Conclusions

The relationship between the influencing factors of air gestures was established in this study through factor analysis and interpretative structural modeling. The relationship between the primary and secondary factors affecting the satisfaction of mid-air gestures is clearly shown. The results show that:

• The thirty observed variables can be summarized into six latent variables through factor analysis, namely Simplicity, Efficiency, Friendliness, Intuitiveness, Attractiveness, and Tiredness.
• In the interpretative structural modeling analysis, the hierarchical relationship shows that Intuitiveness is the basic influencing factor affecting the satisfaction of air gestures, and interaction designers need to focus on this when designing mid-air gestures.
• Attractiveness, as the first layer of the hierarchy, is the goal of gesture design. The main reason for this result is that air gestures are a very new form of human–computer interaction. Interaction designers need to attract users in many ways so that users accept air gestures. Intuitive mid-air gestures are simple for users, so they are attractive.
• This study can be used as a reference for relevant practitioners and provide a theoretical basis for designers’ gesture design and selection.

6. Limitations and Future Directions

Interpretative structural modeling provides a new approach to researching the influencing factors of air gesture satisfaction. However, this method can only approximately analyze the hierarchy of factors influencing mid-air gesture satisfaction from a qualitative perspective. To clarify the problem more accurately and design air gestures that satisfy users, the use of scientific quantitative analysis methods is required for future research. In addition to the development of the theory, mid-air gestures with higher satisfaction based on this theory should be designed, as they are also an important research direction to provide input methods for the development of the Metaverse.