Keywords

1 Introduction

With the development of mobile information, map navigation interaction in dynamic environment becomes more and more complex. Facing constantly changing technologies and interactive systems, the user’s maladjustment in the interaction process is becoming more and more obvious [1]. Due to the current interaction mode of multi-sensory channel fusion and the application of artificial intelligence technology, the interaction mode, operation process, interface layout and information level of map navigation have been changed, resulting in the user’s unclear information, unclear positioning and frequent errors in the use process [1]. Facing the current situation, scholars at home and abroad have proposed many solutions such as deconstruction of human-computer interface, reconstruction of information system, and establishment of evaluation system. However, the current interaction design field has not yet established standard for the application of artificial intelligence technology in map navigation. Therefore, in the field of map navigation interaction at mobile terminal, the influencing factors of user behavior change and interaction availability. Therefore, the establishment of a systematic and scientific evaluation system is still a matter of concern.

At present, scholars at home and abroad have summed up different evaluation systems according to different specific evaluation objects. Pucillo [2] combines the theory of user experience and interactive availability, and puts forward the model of user experience availability. This is of great help to the research on the evaluation system of map navigation interactive usability in the paper.

Through literature review and questionnaire survey, this paper selected the factors that affect the interactive usability on the subjective emotion, information cognition and objective operation level, established the corresponding evaluation model, explored the influence degree of the factors on the interactive usability of mobile map navigation, and measured the influence degree with a systematic evaluation system, so as to improve the users’ experience and interaction quality.

In the first section, this paper summarizes the social status of mobile terminal navigation and the problems of existing solutions. In the second section, based on the efforts made by domestic and foreign scholars in this field, we get the user’s behavior mode in the map navigation context and the interactive usability impact level of the corresponding behavior stage. In the third section, we select the evaluation indexes that have a great impact on the interactive usability, and in the fourth section, we evaluate the index weight by experiment and questionnaire scoring, in the fifth section, we calculate the evaluation index score by fuzzy comprehensive evaluation, and in the sixth section, we get the final conclusion.

2 Related Work

2.1 User Behavior Model in the Context of Map Navigation

The user behavior is easily affected by the thinking pattern, which is based on the past experience or rules to quickly evaluate the existing situation. Therefore, in order to eliminate the influence of emotional factors and establish an objective user behavior model, Card et al. [3] proposes a GOMS model, namely goals, operations, methods and selection rules. Based on the four stages of an action in motivational psychology, Norman [4] establishes a user behavior model of human-computer dialogue, which divided human behavior into three parts: goal, execution and evaluation.

In the context of map navigation, the user perceives the environmental information with the navigation system, defines the behavior sequence, executes the action, and completes the action and evaluation finally [5]. At present, researchers have studied user behavior models from various perspectives. From the perspective of user experience, Yang [6] summarizes the logical characteristics of user behavior and the corresponding interface design. Based on the logic of user behavior, Zheng [7] summarizes four design principles of user behavior: consistent behavior, related behavior, similar behavior and behavior contrary. Richard and Jan [8] emphasize the attention to behavior itself in the interaction process, and propose a complete user behavior research process. Wang et al. [9] propose that when users want to complete a certain task, first they need to determine the behavior motivation and have behavior trigger, then they can change the behavior according to the change of using scenes.

To sum up, based on the previous research on user behavior, this paper divided the user behavior into four stages: goal, method, operation and evaluation, which are responding to determination of task, determination of method, sequential execution of action and evaluation of operation effect. The above four stages response to the user determination of destination, selection of navigation mode, entry of navigation mode and the consistency of navigation mode and results (Fig. 1).

Fig. 1.
figure 1

The user behavior map of navigation

Full size image

2.2 The Interactive Availability of Map Navigation

At present, the international organization for Standardization (ISO) describes usability as the effectiveness, efficiency and subjective satisfaction of a specific user when using a specific product to complete a specific task in a specific use environment [10]. Nielsen, founder of usability engineering, believes that usability includes learnability, memorability, efficiency, error rate and satisfaction. Nielsen and his colleagues put forward 10 main usability principles: visibility of system state, system should be consistent with the real world, the control and autonomy of users, system consistency and standardization, help users diagnose and repair errors, prevent errors, rely on identification rather than memory, flexibility of operation and minimize design [11] (Fig. 2).

Fig. 2.
figure 2

Ideas of the paper

Full size image

In a word, this paper combined the four stages of the user in the navigation process to divide the usability target into the subjective emotional level, the information cognitive level and the objective operation level from the subjective and objective interaction usability perspectives.

3 The Formulation of Map Navigation Interactive Availability Evaluation Index

In this paper, when selecting the evaluation indexes of mobile terminal map navigation, we follow the principles of hierarchy, integrity and feasibility.

Based on the above user behavior model, usability goal and usability design principle in the context of mobile terminal map navigation, this paper evaluated the usability of mobile terminal map navigation from the emotional level, cognitive level and objective level. This paper determined the first level evaluation index of navigation interactive usability as feasibility, learnability, security, visibility, memorability and consistency, effectiveness and fault tolerance. There are secondary evaluation indexes under each evaluation objective in the first level. The paper defined the first level as criterion layer, and determined the second level as scheme layer (Table 1).

Table 1. Description of scheme layer characteristics
Full size table

4 Evaluation Index Weight

4.1 Experiment

Experimental Purpose

The subjects were asked to score the evaluation index according to the specific feelings.

Experimental Equipment

Mobile Baidu map navigation app 10.22.0.960 version.

Participants

Ten participants were invited to participate in the experiment. The age range was from 20 to 30 years old, and the sex ratio was 4:1. The use process was observed in the same use scenario. All subjects had normal hearing, vision and good rest. And a single test lasted about 30 min. In order to avoid the influence of learning effect on the experimental results, the participants were fully familiar with the navigation operation of mobile phones before the experiment. Each subject was not allowed to discuss the problems related to the experiment.

Experimental Process

Set the same navigation destination on Baidu map navigation app, and arrange the participants to complete the same navigation task in the same use scenario. After completing the assigned task, the participants will score the elements according to the relative importance scale of AHP factors. Before the formal experiment, the subjects had 10 min to be familiar with the equipment and scoring rules. During the experiment, the assistant is not allowed to prompt the participants unless they ask for help.

Experimental Result

After the subjects completed the assigned navigation task, they were asked to evaluate the first level and second level rating indicators in the approach of AHP. According to the relative importance of the evaluation indicators and the corresponding scale, a judgment matrix was constructed. Finally, the average value was obtained, which is based on the scores of invited people, and a judgment matrix of the average value was constructed (See Table 2). Due to space limitation, only the judgment matrix of the average value of the first level evaluation index is shown here. In this paper, after comprehensive processing, the weight value and other data will be calculated according to the judgment matrix.

Table 2. Judgment matrix of first level evaluation index
Full size table

4.2 Data Analysis

Calculating Eigenvector and Index Weight of Judgment Matrix

Sum the indexes of each column of judgment matrix A.

$$ A_{j} \, = \,\sum\nolimits_{i = 1}^{{n_{2} }} {a_{ij} \,\,{\text{i}}\, = \,1,2, \ldots ,{\text{n}}_{1} ;\,\,{\text{j}}\, = \,1,2, \ldots , {\text{n}}_{2} } $$
(1)

In the formulation: n1 is the number of square matrix rows; n2 is the number of square matrix arrays.

The new matrix B is obtained by normalizing each index of judgment matrix A.

$$ {\text{B}}\, = \,\frac{{a_{ij} }}{{A_{j} }}\,\, i\, = \,1,2, \ldots n_{1} \,\,j\, = \,1,2, \ldots n_{2} $$
(2)

Sum the indexes of each row of matrix B, and get the eigenvector \( {\text{T}}_{\text{i}} \) as follows:

$$ T_{i} = \sum\nolimits_{j = 1}^{{n_{2} }} {b_{ij} \,\, i\, = \,1,2, \ldots n_{1} \,\,j = 1,2, \ldots n_{2} } $$
(3)

After the eigenvector \( {\text{T}}_{\text{i}} \) is normalized, the index weight \( {\text{W}}_{\text{i}} \) is obtained as follows:

$$ W_{i} \, = \,\left| {T_{i} } \right|/\sum \left| {T_{i} } \right| \,\, i\, = \,1,2, \ldots n_{1} $$
(4)

After the above steps, the following levels of interactive usability evaluation index weight values are obtained (Table 3).

Table 3. Weight value and ranking of primary evaluation indexes
Full size table

Weight analysis of each evaluation index: B2 learnability, B7 effectiveness and B8 fault tolerance are the most important indexes that affect the interactive availability of map navigation on the mobile. B4 visibility and B3 security also have a great impact on the interactive availability of map navigation on the mobile terminal. Among the first level evaluation indexes, the weight value of learnability is the highest, which reflects the users subjective demand for learnability of the interactive system; the second is effectiveness and the third is fault tolerance, which reflect the objective data of the user when using the product objectively; the weight difference among the visibility, security indexes and the top three indexes is small, which indicates that good visibility and security can also greatly improve the interactive usability; B6 consistency, B5 memorability and B1 feasibility account for a relatively small proportion, which have little influence on the interactive usability of map navigation on the mobile, but they are also indispensable factors (Table 4).

Table 4. Weight and ranking of secondary evaluation indexes of interactive availability
Full size table

Numerical analysis of the weight of each evaluation index: C41 feedback timeliness, C71 task completion steps and C61 information comprehensibility are the most important indexes that affect the interactive availability of map navigation on the mobile terminal. C51 information density suitability and C22 learning fatigue also have an important influence on the interactive availability of mobile map navigation. Among the first level evaluation indexes, the weight value of feedback timeliness is the highest, which reflects that users want to get system feedback in time when they are cognizing system information; the second is the operation steps of task completion, which reflects the objective data of users when they are operating products objectively; the weight value of information density suitability and learning fatigue index is not much smaller than the top three indexes’ weight value, which indicates that information density suitability and learning fatigue also plays an important role in improving the interactive usability of map navigation on the mobile; the weight values of Comfort in Use and psychological burden are moderate, reflecting the users subjective emotional demands in the interaction process; the relatively small ones are C72 task completion time, C42 system state understanding process, C83 error correction method and C33 help function, which are useful for map guidance on the mobile. The impact of mobile map navigation interactive availability is small, but it should not be ignored when designers want to improve the interactive usability.

Consistency Test of Judgment Matrix

Calculate λ maxof Judgment Matrix A.

$$ {\varvec{\uplambda}}_{ \hbox{max} } \, = \,\frac{1}{{n_{3} }}\sum \left( {A \cdot W} \right)_{i} /W_{j} \quad n_{3} \in Z^{ + } $$
(5)

In the formulation: n3 is the order; Z+ is the positive integer set; A · W is the multiplication of judgment matrix A and weight set W, W = (W1, W2, Wn1), the result is a column vector.

Calculate the Consistency Index CI of Judgment Matrix A.

The consistency index is calculated by CI. The smaller the CI is, the greater the consistency is. The degree of inconsistency of A is measured by the value of λ - n. The consistency indicators are defined as:

$$ {\text{CI}}\, = \,\frac{\lambda \, - \,n}{n\, - \,1} $$
(6)

When CI is close to 0, there is satisfactory consistency; when CI is larger, there is more inconsistency; when CI = 0, there is complete consistency.

In order to measure the size of CI, the random consistency index RI is introduced.

$$ {\text{RI}}\, = \,\frac{{CI_{1} + CI_{2} + \cdots + CI_{n} }}{n} $$
(7)

Considering that the deviation of consistency may be caused by random reasons, when checking whether the judgment matrix has satisfactory consistency, it is necessary to compare CI with random consistency index RI to obtain the test coefficient CR, which is as follows:

$$ {\text{CR}}\, = \,\frac{CI}{RI} $$
(8)

Generally, if CR is in line with CR < 0.1, the judgment matrix is considered to pass the consistency test, otherwise, it will not have satisfactory consistency.

According to the above consistency test steps, we can get (Tables 5 and 6):

Table 5. Consistency test of primary evaluation indexes
Full size table
Table 6. Consistency test of secondary evaluation indexes
Full size table

According to the above data, the weights of the evaluation indexes calculated by AHP in this paper have passed the consistency test. The CI value of the first level evaluation indexes approaches to 0, which has satisfactory consistency and meets the requirements of CR < 0.1. The CI value of the second level evaluation indexes of the third group and the eighth group is close to 0, with satisfactory consistency. The CI value of the rest second level evaluation indexes is 0, with consistency. What’s more, the CR value meets the requirements of CR < 0.1, passing the consistency test.

5 Fuzzy Evaluation Method to Verify Evaluation Indexes

Establish the evaluation object factor domain U = {U1, U2, Un5}, n5 are the number of factors. Create comment set V = {V1, V2, VX}, X is the number of comments.

In this paper, we set the comment set V = {V1, V2, V3, V4, V5} and use the 5-point Likert scale to assign 1–5 points to it. 5 points correspond to excellent, 4 points correspond to good, 3 points correspond to medium, 2 points correspond to general, and 1 point corresponds to poorness. V value results are shown in the Table 7:

Table 7. Fuzzy comprehensive evaluation and assignment
Full size table

The fuzzy relation matrix R between the factor index Ui and the evaluation index Vj is established as follows:

$$ {\mathbf{R}} \, = \,\left[ {\begin{array}{*{20}c} {\begin{array}{*{20}c} {r_{11} } & {r_{12} } & \cdots \\ {r_{21} } & {r_{22} } & \cdots \\ \vdots & \vdots & \ddots \\ \end{array} \begin{array}{*{20}c} {r_{1j} } \\ {r_{2j} } \\ \vdots \\ \end{array} } \\ {\begin{array}{*{20}c} {r_{i1} } & {r_{i2} } \\ \end{array} \begin{array}{*{20}c} \cdots & {r_{ij} } \\ \end{array} } \\ \end{array} } \right] $$
(9)

In the above matrix R: \( r_{ij} \) is the membership degree of the i factor index to the j comment, j = 1, 2… X.

The fuzzy matrix of matrix R is calculated, and the membership matrix D of standard level to comment set is obtained as follows:

$$ {\text{D}}\,{ = }\,{\text{W}} . {\text{ R}} $$
(10)

The fuzzy and comprehensive score E [12] are:

$$ {\text{E}}\,{ = }\,{\text{D}} . {\text{ N}} $$
(11)

In the above formula, N is the set of expert scores.

In this paper, 10 experts are invited to carry out fuzzy evaluation on the evaluation indexes of map navigation interactive usability of mobile terminal. After the fuzzy evaluation matrix is obtained, the comprehensive score is calculated according to the formula of E [12] = D · N, and the full score is 50 points.

It can be seen from the above that the weight value set of the first level evaluation index of the criterion layer is as follows:

$$ W_{b} \, = \,\left( {\begin{array}{*{20}c} {0.025,\,0.333,\,0.077,\,0.115,\,0.035,} \\ {0.052,\,0.199,\,0.162} \\ \end{array} } \right) $$

The second level evaluation index weight sets of scheme level are as follows:

$$ W_{1} \, = \,\left( {0.667,\,0.333} \right) $$
$$ W_{2} \, = \,\left( {0.250,\,0.748} \right) $$
$$ W_{3} \, = \,\left( {0.350,\,0.584,\,0.061} \right) $$
$$ W_{4} \, = \,\left( {0.875,\,0.125} \right) $$
$$ W_{5} \, = \,\left( {0.800,\,0.200} \right) $$
$$ W_{6} \, = \,\left( {0.833,\,0.167} \right) $$
$$ W_{7} \, = \,\left( {0.857,\,0.143} \right) $$
$$ W_{8} \, = \,\left( {0.634,\,0.261,\,0.106} \right) $$

According to the above calculation steps, we can get the following formula.

The comprehensive evaluation matrix of the first level evaluation index is:

$$ \text{R}\, = \,\left[ {\begin{array}{*{20}c} {D_{1} } \\ {D_{2} } \\ {D_{3} } \\ {D_{4} } \\ {D_{5} } \\ {D_{6} } \\ {D_{7} } \\ {D_{8} } \\ \end{array} } \right]\, = \,\left[ {\begin{array}{*{20}c} {0.500} & {0.233} & {0.100} & {0.133} & {0.034} \\ {0.375} & {0.275} & {0.226} & {0.099} & {0.025} \\ {0.526} & {0.164} & {0.129} & {0.161} & {0.020} \\ {0.713} & {0.188} & {0.025} & {0.038} & {0.036} \\ {0.520} & {0.340} & {0.040} & {0.060} & {0.040} \\ {0.533} & {0.283} & {0.118} & {0.033} & {0.033} \\ {0.629} & {0.200} & {0.086} & {0.029} & {0.056} \\ {0.485} & {0.116} & {0.121} & {0.174} & {0.104} \\ \end{array} } \right] $$

The final result:

$$ \begin{aligned} & \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad {\text{D}}\, = \,{\text{W}}_{b} \, *\,{\text{R}} \\ & = \left( {\begin{array}{*{20}c} 0.025 & 0.333 & 0.077 & 0.115 \\ 0.035 & 0.052 & 0.199 & 0.162 \\ \end{array} } \right) \left[ {\begin{array}{*{20}c} {0.500} & {0.233} & {0.100} & {0.133} & {0.034} \\ {0.375} & {0.275} & {0.226} & {0.099} & {0.025} \\ {0.526} & {0.164} & {0.129} & {0.161} & {0.020} \\ {0.713} & {0.188} & {0.025} & {0.038} & {0.036} \\ {0.520} & {0.340} & {0.040} & {0.060} & {0.040} \\ {0.533} & {0.283} & {0.118} & {0.033} & {0.033} \\ {0.629} & {0.200} & {0.086} & {0.029} & {0.056} \\ {0.485} & {0.116} & {0.121} & {0.174} & {0.104} \\ \end{array} } \right] \\ & \quad \quad \quad \quad \quad \quad \quad \quad \quad = \left( {0.510\;\; 0.217\;\; 0.135\;\; 0.091\;\; 0.046} \right) \\ \end{aligned} $$

The final score is:

$$ {\text{E}}_{\text{f}} \, = \,{\text{D}}\cdot{\text{N}}\, = \,\left( \text{0.510\;\; 0.217\;\; 0.135\;\; 0.091\;\; 0.046} \right) \left[ {\begin{array}{*{20}c} {50} \\ 0 \\ 0 \\ 0 \\ 0 \\ \end{array} } \right]\, = \,25.5 $$

Similarly, the final score of each evaluation index can be obtained:

According to the above data results, we can conclude that the final evaluation result is excellent. According to the principle of maximum membership, the quality of the interactive usability evaluation system of mobile terminal map navigation established in this paper is excellent (Table 8).

Table 8. Final scores of primary and secondary evaluation indexes
Full size table

According to the above calculation of membership degree of evaluation indicators at all levels, the following results can be obtained:

  • According to the principle of maximum membership, the quality of the first level evaluation index is evaluated as excellent. The weight value of learnability ranks first, which shows that learnability takes a critical position on the interactive availability of mobile map navigation. According to the final calculation score of the secondary index, it can be seen that learning fatigue and system level complexity are essential evaluation factors in the interactive usability evaluation system.

  • The maximum membership of the first level evaluation index effectiveness is 0.629, and the ranking of weight value is the second, which proves that this index has an important influence on the interactive availability of map navigation on the objective operation level. According to the final calculation score of the secondary index, it can be seen that task completion steps have a significant influence on interactive usability.

  • The maximum membership of the first level evaluation indexes fault tolerance and visibility are 0.485 and 0.713 respectively. The judgment results are excellent and the weight value ranks high, which shows that fault tolerance and visibility also have an important impact on the interactive availability of map navigation. Combined with the final score of secondary indicators, it can be seen that the scores of error number, error correction time and feedback timeliness are high, which are important factors that cannot be ignored.

  • The first level evaluation index security has a maximum membership of 0.526, a moderate ranking of weight value. Besides, psychological burden and comfort in use indexes both have a high final score. This situation proves that security index still plays a key role in the interactive usability of map navigation. Interaction designers should consider the subjective feelings of users and pay attention to them in the interaction design.

  • The maximum membership degrees of consistency, memorability and feasibility are 0.533, 0.520 and 0.500 respectively. The weight values of these three evaluation indexes rank lower, and their scores are similar. Combined with the final scores of the secondary evaluation indexes, it can be seen that due to different factors such as users’ personal experience and status, the users’ intention achievement degree and task completion difficulty and information availability are different. In the process of interactive design, interactive designers should pay attention to these factors.

6 Conclusion

According to the calculation and verification results in this paper, the following conclusions can be drawn:

  • On the subjective emotional level, learnability and security have the greatest influence on the usability of interaction. Interaction designers should focus on reducing the users’ learning fatigue and psychological burden, and improving the users comfort;

  • At the level of information cognition, consistency and memorability have a great influence on the usability of interaction. Interaction designers should pay attention to the timeliness of system feedback, the comprehensibility of information and the suitability of information density;

  • In the objective operation level, the effectiveness and fault tolerance play a critical role on the interactive usability of map navigation on the mobile terminal. In the interactive design process, the interaction designer should focus on the task completion steps and the number of errors of the system, so as to provide a better interactive experience of map navigation for users.

Through the above literature review and questionnaire survey, this paper analyzed the interactive usability goal of mobile map navigation from the subjective and objective perspective, selected the factors that affect the interactive usability on the subjective emotion, information cognition and objective operation level, established the corresponding evaluation model, and explored the influence of the factors on the interactive usability of mobile map navigation on the three perspectives. The systematic evaluation system is used to measure the impact, so as to provide a meaningful optimization strategy for improving the interactive quality of user map navigation and provide better interactive experience for users in the navigation process of applying artificial intelligence technology. However, due to the small number of people in the questionnaire, the final evaluation system is not comprehensive enough, which needs further improvement and optimization.