Sage Journals: Discover world-class research

Abstract

This study was designed to build an evaluation system for secondary task driving safety by using method of Fuzzy Analytic Network Process (F-ANP). Forty drivers completed driving on driving simulator while interacting with or without a secondary task. Measures of fixations, saccades, and vehicle running status were analyzed. According to five experts' opinions, a hierarchical model for secondary task driving safety evaluation was built. The hierarchical model was divided into three levels: goal, assessment dimension, and criteria. Seven indexes make up the level of criteria, and the assessment dimension includes two clusters: vehicle control risk and driver eye movement risk. By method of F-ANP, the priorities of the criteria and the subcriteria were determined. Furthermore, to rank the driving safety, an approach based on the principle of maximum membership degree was adopted. At last, a case study of secondary task driving safety evaluation by forty drivers using the proposed method was done. The results indicated that the application of the proposed method is practically feasible and adoptable for secondary task driving safety evaluation.

1. Introduction

According to Sayer et al., drivers were observed engaged in secondary behaviors in approximately 34% of the clips. Conversation with passengers was the most common secondary behavior (15.3%), followed by grooming (6.5%) and using cellular phones (5.3%). Younger drivers were most likely to engage in secondary behaviors overall (42%) [1]. As the use of in-vehicle technologies becomes more popular, there is concern about a concomitant increase in driver distraction arising from their use.

Driver distraction can be defined as the diversion of attention from the driving task as compelled by an activity or event, inside or outside the vehicle, which competes for the attention of the driver. It usually results in a delayed recognition and processing of the information needed to accomplish the driving task safely [2]. A large body of evidence suggests that driving performance and driver visual degrade when drivers engage in secondary behaviors, such as using cellular phone, navigation system, and even conversation with passengers [3, 4].

Previous studies have shown that the effects of secondary task on driving performance are significant. Chisholm et al. found that the iPod tasks had larger variation in steering wheel adjustments than usual [5]. Jin et al. reported that the in-vehicle touch-screen telephone weakened driver's ability to control the vehicle [6]. Kun et al. research showed that the effects of secondary task driving on lane position, steering wheel angle, and the velocity of participants' cars were significant [7]. With regard to the effect of secondary task on driver eye movements, Young et al. found that drivers typically made a large number of short fixations to the in-vehicle device during the secondary task driving [8]. Pupil diameter was also significantly increased with doing the touch-screen secondary task while driving [9]. In addition, Konstantopoulos et al. also studied the horizontal and vertical spread of search angles in different driving conditions [10]. Saccade as another important measure of eye-movement behavior was less researched compared to fixation. Yang et al. found that in visual tasks, saccade duration and amplitude changed obviously [11]. Victor et al. reported that drivers made significantly fewer saccades with the cell phone task while driving [12]. Ahlstrom studied the effects of aircraft density on driver saccade behavior and found that means saccade distance decreased as the aircraft density increased [13].

As discussed above, secondary task driving is a significant and multifaceted road safety issue and has already threatened driving safety. However, few relative studies have systematically examined the driving safety levels with secondary task while driving, and the evaluation system for secondary task driving safety is also largely unexplored. Aimed at that, the overall goal of this paper is to examine the influence of different secondary tasks on measures of driver eye movements and vehicle control and to develop a method to evaluate the safety for secondary task driving.

2. Method

Fuzzy Analysis Network Process (F-ANP) is a new system analysis method based on traditional Analysis Network Process (ANP), combined with fuzzy comprehensive evaluation knowledge [14]. In the process of evaluation, this method not only considers the hierarchy of indicators, but also takes into account the relationship between the factors in each layer, therefore addressing the kind of assessment selection problem on evaluation index system with hybrid network hierarchy structure [15]. In this study, the F-ANP approach is applied to relieve the uncertainness of ANP method. The fuzzy comparison ratios are used to determine the assessment dimension and criteria weights from subjective judgments.

2.1. Structuring the Hierarchical Model for Secondary Task Driving Safety Evaluation, Including Goal, Assessment Dimension, and Criteria

First, this study determines the goal for the secondary task driving safety by focus group discussion with 5 experts on driving safety evaluation. In addition, we decide on 7 criteria and classify them into two assessment dimensions: the visual behavior safety and the vehicle control safety. The hierarchical model for secondary task driving safety evaluation is proposed in Figure 1.

Figure 1:

The hierarchical model for secondary task driving safety.

2.2. Calculating the Weights of Each Criterion

To reduce the decision-maker preferences, pair-wise comparison matrices are structured using triangular fuzzy numbers (l, m, u). Triangular fuzzy numbers are used to investigate the ambiguities involved in the linguistic-data assessment process. Because there are many criteria for the issue involving a suitable recycling option selection and because most of these are qualitative, it is difficult to assess these criteria quantitatively [16]. Fuzzy numbers reflect the relative strength of each pair of network elements [17]. So, in this study, we structure pair-wise comparison matrices using triangular fuzzy numbers.

2.2.1. Constructing the Fuzzy Judgment Matrix Using Triangular Fuzzy Numbers

The weight of each criterion is calculated through the following steps.

Step 1. Calculating the local weight matrix W₁₁ of criterion u₁₁. The triangular fuzzy number complementary judgment matrix with respect to criterion u₁₁ by triangular fuzzy numbers is calculated as shown in Table 1.

Calculating the comprehensive importance matrixes C₁₁, C₁₂, C₁₃, and C₁₄ of criterion u₁₁ according to the following formula [18]:

\begin{matrix} C_{1 i} = {\sum_{j = 1}^{n_{1}} p_{i j} \otimes [\sum_{i = 1}^{n_{1}} \sum_{j = 1}^{n_{1}} p_{i j}]}^{- 1} (i = 1,2, \dots, n_{1}) \\ (j = 1,2, \dots, n_{1}) . \end{matrix}

(1)

In formula (1), p_ij = (l_ij, m_ij, s_ij) is the triangular fuzzy number complementary judgment matrix. The calculated comprehensive importance matrixes of criterion u₁₁ are shown in Table 2.

Establishing the possible degrees for C_1i ≥ C_1k according to the following formula [19]. The results are demonstrated in Table 3:

V (C_{1 i} \geq C_{1 k}) = \{\begin{cases} 1 & (m_{i j}^{1 i} \geq m_{i j}^{1 k}) \\ \frac{s_{i j}^{1 i} - l_{i j}^{1 k}}{s_{i j}^{1 i} - l_{i j}^{1 k} + m_{i j}^{1 k} - m_{i j}^{1 i}} & (m_{i j}^{1 i} < m_{i j}^{1 k}, l_{i j}^{1 k} \leq s_{i j}^{1 i}) \\ 0 & (else) . \end{cases}

(2)

Obtaining the possible degrees of criterion u₁₁ is more important than other criteria:

W_{11}^{{(11)}^{'}} = {(d^{'} (u_{11}), d^{'} (u_{12}), d^{'} (u_{13}), d^{'} (u_{14}))}^{T} = {(1.000,0.650,0.921,0.911)}^{T} .

(3)

The weight matrix W₁₁⁽¹¹⁾ is determined after $W_{11}^{{(11)}^{'}}$ is normalized:

W_{11}^{(11)} = {(\frac{d^{'} (u_{11})}{\sum_{i = 1}^{4} d^{'} (u_{1 i})}, \frac{d^{'} (u_{12})}{\sum_{i = 1}^{4} d^{'} (u_{1 i})}, \frac{d^{'} (u_{13})}{\sum_{i = 1}^{4} d^{'} (u_{1 i})}, \frac{d^{'} (u_{14})}{\sum_{i = 1}^{4} d^{'} (u_{1 i})})}^{T} = {(0.287,0.187,0.265,0.262)}^{T} .

(4)

With the same method, W₁₁⁽¹²⁾, W₁₁⁽¹³⁾, and W₁₁⁽¹⁴⁾ are obtained. At last we can obtain the weight matrix W₁₁ of u₁₁ as follows:

\begin{matrix} W_{11}^{(12)} = {(0.233,0.317,0.230, 0.220)}^{T}, \\ W_{11}^{(13)} = {(0.192,0.235, 0.371,0.202)}^{T}, \\ W_{11}^{(14)} = {(0.169,0.277,0.231,0.323)}^{T}, \\ W_{11} = {(W_{11}^{(11)}, W_{11}^{(12)}, W_{11}^{(13)}, W_{11}^{(14)})}^{T} = (\begin{pmatrix} 0.287 & 0.233 & 0.192 & 0.169 \\ 0.187 & 0.317 & 0.235 & 0.277 \\ 0.265 & 0.230 & 0.371 & 0.231 \\ 0.262 & 0.220 & 0.202 & 0.323 \end{pmatrix}) . \end{matrix}

(5)

Table 1:

Triangular fuzzy number complementary judgment matrix with respect to criterion u₁₁.

u ₁₁	u ₁₁	u ₁₂	u ₁₃	u ₁₄

u ₁₁	(0.5, 0.5, 0.5)	(0.6, 0.7, 0.8)	(0.4, 0.6, 0.8)	(0.3, 0.5, 0.8)
u ₁₂	(0.2, 0.3, 0.4)	(0.5, 0.5, 0.5)	(0.2, 0.3, 0.5)	(0.3, 0.4, 0.7)
u ₁₃	(0.2, 0.4, 0.6)	(0.5, 0.7, 0.8)	(0.5, 0.5, 0.5)	(0.4, 0.5, 0.8)
u ₁₄	(0.2, 0.5, 0.7)	(0.3, 0.6, 0.7)	(0.2, 0.5, 0.6)	(0.5, 0.5, 0.5)

Table 2:

Comprehensive importance matrixes of criterion u₁₁.

Description	Results	Description	Results

$\sum_{j = 1}^{4} p_{1 j}$	(1.8, 2.3, 2.9)	C ₁₁	(0.176, 0.288, 0.500)
$\sum_{j = 1}^{4} p_{2 j}$	(1.2, 1.5, 2.1)	C ₁₂	(0.118, 0.188, 0.362)
$\sum_{j = 1}^{4} p_{3 j}$	(1.6, 2.1, 2.7)	C ₁₃	(0.157, 0.263, 0.466)
$\sum_{j = 1}^{4} p_{4 j}$	(1.2, 2.1, 2.5)	C ₁₄	(0.118, 0.263, 0.431)
$\sum_{i = 1}^{4} \sum_{j = 1}^{4} p_{i j}$	(5.8, 8.0, 10.2)	${[\sum_{i = 1}^{4} \sum_{j = 1}^{4} p_{i j}]}^{- 1}$	(0.098, 0.125, 0.172)

Table 3:

Possible degrees for C_1i ≥ C_1k.

Description	Results	Description	Results	Description	Results

V(C₁₁ ≥ C₁₂)	1.000	V(C₁₁ ≥ C₁₃)	1.000	V(C₁₁ ≥ C₁₄)	1.000
V(C₁₂ ≥ C₁₁)	0.650	V(C₁₂ ≥ C₁₃)	0.732	V(C₁₂ ≥ C₁₄)	0.765
V(C₁₃ ≥ C₁₁)	0.921	V(C₁₃ ≥ C₁₂)	1.000	V(C₁₃ ≥ C₁₄)	1.000
V(C₁₄ ≥ C₁₁)	0.911	V(C₁₄ ≥ C₁₂)	1.000	V(C₁₄ ≥ C₁₃)	1.000

Step 2. According to the above mentioned ways, we can obtain the weight matrixes W₂₂, W₁₂, and W₂₁ of u₂₂, u₁₂, and u₂₁, respectively.

\begin{matrix} W_{22} = (\begin{pmatrix} 0.474 & 0.211 & 0.357 \\ 0.340 & 0.532 & 0.276 \\ 0.186 & 0.257 & 0.367 \end{pmatrix}), \\ W_{12} = (\begin{pmatrix} 0.227 & 0.262 & 0.188 \\ 0.240 & 0.318 & 0.304 \\ 0.312 & 0.244 & 0.279 \\ 0.221 & 0.176 & 0.229 \end{pmatrix}), \\ W_{21} = (\begin{pmatrix} 0.415 & 0.198 & 0.331 & 0.254 \\ 0.329 & 0.474 & 0.192 & 0.468 \\ 0.256 & 0.328 & 0.477 & 0.278 \end{pmatrix}) . \end{matrix}

(6)

2.2.2. Calculating the Weights of Assessment Dimension Factors

Method of triangular fuzzy numbers is also used to calculate the weights of assessment dimension factors. The weight matrix A of assessment dimension factorsis shown as follows:

A = (\begin{pmatrix} a_{11} & a_{12} \\ a_{21} & a_{22} \end{pmatrix}) = (\begin{pmatrix} 0.574 & 0.449 \\ 0.426 & 0.551 \end{pmatrix}) .

(7)

2.2.3. Calculating the Hypermatrix W and Weighted Super Matrix $\bar{W}$

Consider

W = (\begin{pmatrix} W_{11} & W_{12} \\ W_{21} & W_{22} \end{pmatrix}) \bar{W} = (\begin{pmatrix} a_{11} W_{11} & a_{12} W_{12} \\ a_{21} W_{21} & a_{22} W_{22} \end{pmatrix}) .

(8)

In the hypermatrix Wand weighted super matrix $\bar{W}$ , a₁₁W₁₁, a₁₂W₁₂, a₂₁W₂₁, and a₂₂W₂₂ are shown as follows:

\begin{matrix} a_{11} W_{11} = (\begin{pmatrix} 0.165 & 0.134 & 0.110 & 0.097 \\ 0.107 & 0.182 & 0.135 & 0.159 \\ 0.152 & 0.132 & 0.213 & 0.133 \\ 0.150 & 0.126 & 0.116 & 0.185 \end{pmatrix}), \\ a_{12} W_{12} = (\begin{pmatrix} 0.102 & 0.118 & 0.084 \\ 0.108 & 0.143 & 0.136 \\ 0.140 & 0.110 & 0.125 \\ 0.099 & 0.079 & 0.103 \end{pmatrix}), \\ a_{21} W_{21} = (\begin{pmatrix} 0.177 & 0.084 & 0.141 & 0.108 \\ 0.140 & 0.202 & 0.082 & 0.199 \\ 0.109 & 0.140 & 0.203 & 0.118 \end{pmatrix}), \\ a_{22} W_{22} = (\begin{pmatrix} 0.261 & 0.116 & 0.197 \\ 0.187 & 0.293 & 0.152 \\ 0.102 & 0.142 & 0.202 \end{pmatrix}) . \end{matrix}

(9)

2.2.4. Calculating the Global Weight of Each Criterion

$G = \lim_{x \to \infty} {\bar{W}}^{\infty}$ is the global weight of each criterion. By using Matlab tool, the global weight matrix of each criterion is determined as follows:

G = \lim_{n \to \infty} {\bar{W}}^{\infty} = (0.116,0.140,0.144,0.120,0.157,0.187,0.148) .

(10)

2.3. Defining the Remark Set of Secondary Task Driving Safety

In this study, five levels of secondary task driving safety are classified: very high, high, average, low, and very low.

2.4. Constructing the Membership Functions and Classifying the Secondary Task Driving Safety

The membership function of a fuzzy set is a generalization of the indicator function in classical sets. In this study, the relationship between the membership degree and the driving safety score is not in a monotony trend. So it's suitable to solve this problem using middle type membership function patterns. Therefore, sine curve is selected as the membership function. The fuzzy analytical expression is as follows:

r_{1} (u) = \{\begin{cases} 1 & 0 \leq u \leq 0.1, \\ \frac{1}{2} [\sin (\frac{u - 0.1}{0.2} + \frac{1}{2}) π + 1] & 0.1 < u < 0.3, \\ 0 & u \geq 0.3, \end{cases}

(11)

r_{2} (u) = \{\begin{cases} 0 & u \leq 0.1, \\ \frac{1}{2} [\sin (\frac{u - 0.3}{0.2} + \frac{1}{2}) π + 1] & 0.1 < u < 0.5, \\ 0 & u \geq 0.5, \end{cases}

(12)

r_{3} (u) = \{\begin{cases} 0 & u \leq 0.3, \\ \frac{1}{2} [\sin (\frac{u - 0.5}{0.2} + \frac{1}{2}) π + 1] & 0.3 < u < 0.7, \\ 0 & u \geq 0.7, \end{cases}

(13)

r_{4} (u) = \{\begin{cases} 0 & u \leq 0.5, \\ \frac{1}{2} [\sin (\frac{u - 0.7}{0.2} + \frac{1}{2}) π + 1] & 0.5 < u < 0.9, \\ 0 & u \geq 0.9, \end{cases}

(14)

r_{5} (u) = \{\begin{cases} 0 & u \leq 0.7, \\ \frac{1}{2} [\sin (\frac{u - 0.9}{0.2} + \frac{1}{2}) π + 1] & 0.7 < u < 0.9, \\ 1 & 1 \geq u \geq 0.9 . \end{cases}

(15)

In these formulas, r₁ to r₅ represent the degree of secondary task driving safety belonging to “very low”, “low”, “average”, “high”, and “very high,” respectively, and u is the value of criterion. Due to the maximum membership grade principle, the levels of secondary task driving safety could be determined.

3. Experimental Verification

3.1. Description of the Experiment

A simulated driving experiment is conducted to examine the effects of four secondary tasks on driver eye movements and vehicle running status.

3.1.1. Participants

Forty participants (29 females, 11 males, mean age 32.5 years) between 20 and 52 years of age were recruited from Jilin University and surrounding communities. All participants hold a valid class C1 driver's license for more than 1 year and have an annual mileage from 5000 km to 76000 km. The participants are also required to be in good physical and mental health and not to be taking any medication or drugs that would affect their driving performance. Learning from the study of Guo [20], based on drivers driving experience, we classify them into three groups: experienced driver (mileage is equal or greater than 5 million kilometers), general experienced driver (mileage is between 2 and 5 million kilometers), and inexperienced driver (mileage is equal or less than 2 million kilometers).

3.1.2. Apparatus and Materials

The experiment platform built by intelligent vehicle group in Jilin University includes an eye movements tracking system (SmartEye 6.0), a Besturn B50 car, an embedded driving simulator, and four secondary task devices (an in-vehicle radio, a computer (iPad 4), a fixed telephone, and a Bluetooth headset).

(1) The SmartEye 6.0 Eyes Tracking System. Driver eye movements while driving are captured by SmartEye 6.0 (Smart Eye AB, Första Långgatan 28, Gothenburg, Sweden). It uses four cameras mounted in front of windscreen to capture the driver's eye movements and three scenes' cameras to record the simulated traffic environment. Video analysis is performed using SmartEye-Analysis, an offline data review and reduction analysis program. As compared to head-mounted eye movement tracking systems, SmartEye 6.0 has the advantage of being separately mounted and therefore has less influence on driving performance. In addition, 60 Hz is adopted as the sample frequency in SmartEye 6.0.

(2) Driving Simulator and Secondary Task Devices. The driving simulator is a fixed-base, static driving simulator equipped with a Besturn B50 car. The display system projects onto a screen in front of the vehicle. The experiment apparatus are shown in Figure 2.

Figure 2:

Experimental equipment.

(3) Road Environment. The highway from Changchun to Siping is selected as the driving road environment, for which a real-time 3D model is built using the visual simulation software Multigen Creator and Vega. This highway section has two lanes in each direction, separated by a grassy median. The width of traffic lane and grassy median are 3.75 m and 3 m, separately. The road is in good condition without damage on surface texture and the weather is clear.

(4) Secondary Tasks. There are four secondary tasks adopted in this study: conducting a conversation using Bluetooth headset, talking with laboratory assistant, tuning the radio to a local radio station, and touching the computer or telephone menu to a certain song. The radio task was a press-button task and needed ten operation steps and took about 7–10 s to complete when tested alone, with turning to a local radio station as the end. Computer or telephone task should be completed by touching screen and need more than five completion steps. The conversations of conducting a speech-based Bluetooth headset conversation and talking with laboratory assistant include fifteen simple questions, such as how long have you got the driving license? The number of questions is random.

In these secondary tasks, conducting a speech-based Bluetooth headset conversation and talking with laboratory assistant are cognitive tasks while tuning the radio to a local radio station and touching the computer or telephone menu to a certain song are visual tasks.

3.1.3. Experimental Sessions

The experiment is conducted in two sessions.

Training session: participants practice with secondary tasks and driving in the simulator first separately and then together. The practice session ends when the participants feel they can operate the secondary tasks safely while driving.

Test session: on the premise of security, drivers should complete five drives, a drive without secondary task (baseline) and four drives with secondary tasks.

3.2. Data Analysis

3.2.1. Constructing the Single Factor Assessment Matrix for Each Driver

In order to describe how the matrix constructed, we will take the driver 1 as an example. First of all, the data collected from driving simulator should be processed using the following formula [21 –23]:

u_{11}^{'} = \frac{u_{i} - \min (u_{i})}{\max (u_{i}) - \min (u_{i})} .

(16)

In formula (16), max ⁡ (u_i) is the maximum of criteria value, min ⁡ (u_i) is the minimum of criteria value, and u_i is the criteria value of the present driver.

Second, substituting the ${u_{11}^{'}}^{}$ into formulas (13), (14), (15), and (16), the single factor assessment matrix for each driver is determined. In accordance with this method, driver 1's single factor assessment matrix under normal driving condition is obtained and shown in the matrix R₁:

R_{1} = [\begin{bmatrix} 0 & 0 & 0 & 0.19 & 0.81 \\ 0.18 & 0.82 & 0 & 0 & 0 \\ 0.19 & 0.81 & 0 & 0 & 0 \\ 0 & 0 & 0.67 & 0.33 & 0 \\ 0 & 0 & 0.67 & 0.33 & 0 \\ 0 & 0 & 0 & 0 & 1 \\ 0 & 0 & 0.33 & 0.67 & 0 \end{bmatrix}] .

(17)

In this way, the single factor assessment matrixes of drivers from 1 to 40 under different driving conditions are obtained.

3.2.2. Determining the Driving Safety Levels

The driving safety judgment sets are determined by multiplying the global weight of each criterion G and single factor assessment matrix R. With the maximum membership grade principle, the driving safety levels are divided. So, driver 1's driving safety judgment set under normal driving condition is B₁ as follows:

B_{1} = G \cdot R_{1} = {(\begin{pmatrix} 0.116 \\ 0.140 \\ 0.144 \\ 0.120 \\ 0.157 \\ 0.187 \\ 0.148 \end{pmatrix})}^{T} \times [\begin{bmatrix} 0 & 0 & 0 & 0.19 & 0.81 \\ 0.18 & 0.82 & 0 & 0 & 0 \\ 0.19 & 0.81 & 0 & 0 & 0 \\ 0 & 0 & 0.67 & 0.33 & 0 \\ 0 & 0 & 0.67 & 0.33 & 0 \\ 0 & 0 & 0 & 0 & 1 \\ 0 & 0 & 0.33 & 0.67 & 0 \end{bmatrix}] = (0.053,0.233,0.234,0.213,0.281) .

(18)

The maximum value of B₁ is 0.281. According to the membership function, the score 0.281 belongs to “very high” level of driving safety. Similarly, the driving safety levels of other drivers under different driving conditions could be calculated.

3.3. Results

Based on the evaluation results, we classify and analyze the stata statistically according to driving conditions, driver's driving experience, and types of secondary task.

Tables 4 –7 are statistical data of the influence of driving conditions, types of secondary task, and driver's driving experience on driving safety.

Table 4:

Statistics of drivers' driving safety levels under different driving conditions.

Driving condition	Percent of driving safety level
Driving condition	Very high	High	Average	Low	Very low

Normal	42.5%	22.5%	27.5%	7.5%	0%
Bluetooth headset conversation	2.5%	15%	40%	37.5%	7.5%
Talking with laboratory assistant	5%	5%	45%	40%	5%
Radio tuning	0%	2.5%	17.5%	60%	20%
Touch-screen touching	0%	5%	7.5%	47.5%	40%

Table 5:

Comparison results of driving safety levels under normal driving and secondary task driving conditions.

Driving condition	Percent of driving safety level
Driving condition	Very high	High	Average	Low	Very low

Normal	42.5%	22.5%	27.5%	7.5%	0%
Secondary task driving	1.3%	6.9%	27.5%	46.2%	18.1%

Table 6:

Influences of secondary task type on driving safety levels.

Driving condition	Percent of driving safety level
Driving condition	Very high	High	Average	Low	Very low

Normal	42.5%	22.5%	27.5%	7.5%	0%
Cognitive tasks	2.5%	10%	42.5%	38.8%	6.2%
Visual tasks	0%	3.75%	12.5%	53.75%	30%

Table 7:

Influence of driver's driving experience on driving safety levels.

Driving condition	Driving experience	Percent of driving safety level
Driving condition	Driving experience	Very high	High	Average	Low	Very low

Normal driving	Rich	63.6%	18.2%	18.2%	0%	0%
	General	40%	25%	30%	5%	0%
	Bad	22.2%	22.2%	33.4%	22.2%	0%

Secondary task driving	Rich	4.5%	15.9%	45.5%	31.8%	2.3%
	General	0%	3.75%	25%	55%	16.25%
	Bad	0%	2.8%	11.1%	44.4%	41.7%

3.3.1. Secondary Task Driving Has Significant Influence on Driving Safety

From Tables 4 and 5, we can draw the conclusion that secondary task driving has significant influence on driving safety. “Average” and above of driving safety levels under normal driving condition (92.5%) are higher than driving with Bluetooth headset conversation (57.5%), laboratory assistant talking (55%), radio tuning (20%), and touch-screen touching (12.5%). Overall, during the secondary task driving, “average” and above of driving safety levels account for only 35.7%, while “low” and “very low” of driving safety levels are responsible for 64.3%. Through analysis for test data, it can be seen that secondary task driving, regardless of secondary task type, has significant influence on driving safety.

3.3.2. Visual Tasks Have a Greater Impact on Driving Safety Than Cognitive Tasks

It is evident from the results in Table 6 that driving with visual tasks (“low” and “very low” account for more than 83%) are more dangerous than with cognitive tasks (“low” and “very low” account for about 45%). Additionally, the security of driving with cognitive tasks is also lower than normal driving.

3.3.3. Driver's Driving Experience Has Significant Influence on Driving Safety

The statistical analysis indicates that, under normal driving condition, driving safety level of experienced drivers is the highest (“average” and above account for about 100%). The same situation applies to secondary task driving. Experienced drivers driving with secondary task can still keep good driving safety (“average” and above account for about 65.9%) while inexperienced drivers cannot (“average” and above account for only 13.9%).

4. Conclusions

This study aimed to determine the driving safety level of secondary task driving by using Fuzzy Analytic Network Process approach. The proposed approach takes into account the driver's visual behavior safety and vehicle control safety with their subcriteria (assessment dimension). The interdependencies between criteria are also considered within the proposed approach. The weights of the criteria and subcriteria are determined by using triangular fuzzy numbers. After determining the weights of the criteria and constructing the single factor assessment matrix for each driver, the maximum membership grade principle is applied to divide the safety level of secondary task driving. Finally, the classification is implemented.

Data of forty drivers engaging in driving simulation are used for verifying the applicability of the secondary task driving safety evaluation system. The analysis results showed that secondary task driving, secondary task types, and driver's driving experience have influence on the driving safety, obviously. Secondary task driving can decrease the driving safety significantly.

In conclusion, the proposed approach is useful and effective to classify the safety level of the secondary task driving. Secondary task driving has negative effects on driving safety and drivers should enhance their consciousness away from interacting with secondary tasks while driving. The government should pay more attention to this issue and take action to prevent traffic accident caused by these secondary tasks.

Conflict of Interests

The authors do not have any conflict of interests with the content of the paper.

Footnotes

Acknowledgments

This research was supported partly by the Research Foundation of Shandong Jiaotong University (2014), the Doctoral Scientific Research Foundation of Shandong Jiaotong University, the International Cooperation Project of Science and Technology of Jilin Province under Grant (20130413056GH), the Tsinghua University Open Fund of State Key Laboratory of Automotive Safety and Energy (KF14182), the Key Scientific and Technological Project of Changchun Technology Bureau (13KG05), and Jilin University Alternates Development Scheme of The National Science Fund for Distinguished Scholars.

References

Sayer

J. R.

Devonshire

J. M.

, and Flannagan

C. A.

, The Effects of Secondary Tasks on Naturalistic Driving Performance, The University of Michigan, Transportation Research Institute, Ann Arbor, Mich, USA, 2005.

Jin

Xian

Niua

, and Biec

, “Research on safety evaluation model for in-vehicle secondary task driving,” Accident Analysis & Prevention, 2014.

Jin

Xian

Jiang

Niu

, and Yang

, “Research on evaluation model for secondary task driving safety based on driver eye movements,” Advances in Mechanical Engineering, vol. 2014, Article ID 624561, 8 pages, 2014.

Wang

Sun

, and Guo

, “A region tracking-based vehicle detection algorithm in nighttime traffic scenes,” Sensors, vol. 13, no. 12, pp. 16474–16493, 2013.

Chisholm

S. L.

Caird

J. K.

, and Lockhart

, “The effects of practice with MP3 players on driving performance,” Accident Analysis and Prevention, vol. 40, no. 2, pp. 704–713, 2008.

Jin

L.-S.

Xian

H.-C.

Hou

H.-J.

Niu

Q.-N.

, and Lü

H.-H.

, “Analyzing effects of touch-screen on driving performance,” Journal of Beijing Institute of Technology, vol. 21, no. 2, pp. 6–10, 2012.

Kun

Paek

, and Medenica

, “The effect of speech interface accuracy on driving performance,” in Proceedings of the 8th Annual Conference of The International Speech Communication Association, vol. 1, pp. 2332–2335, August 2007.

Young

K. L.

Mitsopoulos-Rubens

Rudin-Brown

C. M.

, and Lenné

M. G.

, “The effects of using a portable music player on simulated driving performance and task-sharing strategies,” Applied Ergonomics, vol. 43, no. 4, pp. 738–746, 2012.

Jin

Niu

Hou

Xian

Wang

, and Shi

, “Driver cognitive distraction detection using driving performance measures,” Discrete Dynamics in Nature and Society, vol. 2012, Article ID 432634, 12 pages, 2012.

10.

Konstantopoulos

Chapman

, and Crundall

, “Driver's visual attention as a function of driving experience and visibility. Using a driving simulator to explore drivers' eye movements in day, night and rain driving,” Accident Analysis and Prevention, vol. 42, no. 3, pp. 827–834, 2010.

11.

Yang

McDonald

, and Zheng

P. J.

, “The effects of increasing workload on drivers' eye movement,” in Proceedings of the 12th World Conference on Transport Research (WCTR '10), Lisbon, Portugal, 2010.

12.

Victor

T. W.

Harbluk

J. L.

, and Engström

J. A.

, “Sensitivity of eye-movement measures to in-vehicle task difficulty,” Transportation Research Part F: Traffic Psychology and Behaviour, vol. 8, no. 2, pp. 167–190, 2005.

13.

Ahlstrom

, “Work domain analysis for air traffic controller weather displays,” Journal of Safety Research, vol. 36, no. 2, pp. 159–169, 2005.

14.

Chen

and Hu

, “The application of fuzzy analytic network process in supplier selection based on the connection number,” in Proceedings of the International Conference on Artificial Intelligence and Education (ICAIE '10), pp. 554–559, Hangzhou, China, October 2010.

15.

Wang

, and Li

, “RFID-based vehicle positioning and its applications in connected vehicles,” Sensors, vol. 14, no. 3, pp. 4225–4238, 2014.

16.

Liu

Zhang

W. G.

, and Zhang

L. H.

, “Consistency analysis of triangular fuzzy reciprocal preference relations,” European Journal of Operational Research, vol. 235, no. 3, pp. 718–726, 2014.

17.

Wang

Zhang

, and Ban

, “Lane-changing model with dynamic consideration of driver's propensity,” International Journal of Modern Physics C, vol. 26, no. 2, 19 pages, 2015.

18.

Tuzkaya

Gulsun

Kahraman

, and Ozgen

, “An integrated fuzzy multi-criteria decision making methodology for material handling equipment selection problem and an application,” Expert Systems with Applications, vol. 37, no. 4, pp. 2853–2863, 2010.

19.

Isalou

A. A.

Zamani

Shahmoradi

, and Alizadeh

, “Landfill site selection using integrated fuzzy logic and analytic network process (F-ANP),” Environmental Earth Sciences, vol. 68, no. 6, pp. 1745–1755, 2013.

20.

Guo

, Study on effects of traffic environment and driving experience on driver's eye movement and workload [Ph.D. dissertation], Chang'an University, Xi'an, China, 2009.

21.

Tian

G. D.

Chu

J. W.

Liu

Y. M.

Zhao

, and Xu

, “Expected energy analysis for industrial process planning problem with fuzzy time parameters,” Computers and Chemical Engineering, vol. 35, no. 12, pp. 2905–2912, 2011.

22.

Tian

Liu

Tian

, and Chu

, “Evaluation model and algorithm of product disassembly process with stochastic feature,” Clean Technologies and Environmental Policy, vol. 14, no. 2, pp. 345–356, 2012.

23.

Tian

Zhou

Chu

, and Liu

, “Probability evaluation models of product disassembly cost subject to random removal time and different removal labor cost,” IEEE Transactions on Automation Science and Engineering, vol. 9, no. 2, pp. 288–295, 2012.

Classifying Secondary Task Driving Safety Using Method of F-ANP

Abstract

1. Introduction

2. Method

2.1. Structuring the Hierarchical Model for Secondary Task Driving Safety Evaluation, Including Goal, Assessment Dimension, and Criteria

2.2. Calculating the Weights of Each Criterion

2.2.1. Constructing the Fuzzy Judgment Matrix Using Triangular Fuzzy Numbers

2.2.2. Calculating the Weights of Assessment Dimension Factors

2.2.3. Calculating the Hypermatrix W and Weighted Super Matrix W -

2.2.4. Calculating the Global Weight of Each Criterion

2.3. Defining the Remark Set of Secondary Task Driving Safety

2.4. Constructing the Membership Functions and Classifying the Secondary Task Driving Safety

3. Experimental Verification

3.1. Description of the Experiment

3.1.1. Participants

3.1.2. Apparatus and Materials

3.1.3. Experimental Sessions

3.2. Data Analysis

3.2.1. Constructing the Single Factor Assessment Matrix for Each Driver

3.2.2. Determining the Driving Safety Levels

3.3. Results

3.3.1. Secondary Task Driving Has Significant Influence on Driving Safety

3.3.2. Visual Tasks Have a Greater Impact on Driving Safety Than Cognitive Tasks

3.3.3. Driver's Driving Experience Has Significant Influence on Driving Safety

4. Conclusions

Conflict of Interests

Footnotes

Acknowledgments

References

2.2.3. Calculating the Hypermatrix W and Weighted Super Matrix $\bar{W}$