Keywords

1 Introduction

People, vehicles, roads, and the environment are the four elements of traffic safety, among which people have a significant impact on safe driving. According to the traffic accident statistics of various countries in the world, road traffic accidents caused by human factors are as high as 80% to 90%, and road traffic accidents caused by drivers themselves account for more than 70% [1]. By analyzing the psychological factors of drivers and combining them with questionnaire surveys, Yang Yu et al. proposed improving the psychological quality of drivers in order to achieve driving safety [2]. Wu Di et al. analyzed the traffic accidents in Anhui Province in 2019. Among the 22 large road traffic accidents with more than 3 deaths, those caused by the illegal behavior of drivers accounted for the majority [3].

Driving behavior has a significant impact on traffic safety [4]. Yan Ge et al. studied the association between impulsive behavior and violations using data from 299 Chinese drivers. The results show that the driver’s impulsivity is positively correlated with the driver’s positive behavior and some common violations. The other three dimensions of dysfunction are negatively correlated with positive driving behavior, and positively correlated with abnormal driving behavior and fines [5]. Zhang Mengge et al. established an association model between road conditions and abnormal driving behavior based on current research status of driving behavior at home and abroad, combined with data of abnormal driving behavior from the Internet of Vehicles OBD, thereby establishing a research idea for identifying road traffic safety risks.

Many scholars have paid attention to the correlation between the driver’s personal characteristics and driving behavior [6]. Lourens et al. deduced from the Dutch database that there is a relationship between violations and traffic accidents in different types of annual mileage and that there is no difference in the degree of involvement of male and female drivers in accidents. The rate of accidents among young drivers is the highest [7]. Wang et al. employed the Eysenck Personality Questionnaire (EPQ) and the Symptom Self-Rating Scale (SCL-90-R) to assess the personality and mental health of truck drivers, as well as investigate the link between mental health and personal traits. These findings provide a theoretical foundation for truck driver selection and intervention strategies for high-risk drivers, which will help to better manage road traffic safety construction and reduce road traffic injuries.

The Logistic regression model has been used by many researchers to investigate the association between a driver’s personal characteristics and traffic safety behavior. Lin Qingfeng et al. built a Logistic regression model to analyze the relationship between motor vehicle driver attributes, non-motor vehicle driver attributes, motor vehicles, non-motor vehicles, roads, and the environment, and the relationship between the driver’s fault and the severity of the accident. The results show that the severity of motor vehicle accidents is significantly related to seven variables, including the motor vehicle driver’s driving age, motor vehicle safety status, road alignment, and the alignment and motor vehicle driver’s fault [8]. Tian Sheng et al. utilized Pearson correlation analysis and multiple regression model analysis to survey 1,800 primary and middle school children in Guangzhou, and the results showed that education, awareness, attitude, and personal variables influence young people’s traffic safety practices [9].

The current study has conducted a pretty extensive investigation into the relationship between driver behavior and traffic safety. However, its concentration is primarily on ordinary drivers, with little investigation into the features of truck drivers. This article investigates the impact of truck drivers’ personal characteristics on violations, investigates the relationship between the two, and searches for appropriate personal characteristics for truck drivers in order to provide a theoretical foundation and reference for professional truck driver selection.

2 Research Methods

2.1 LOGISTIC Regression Model

The Logistics regression model is a classification model that investigates the link between classification outcomes and affecting factors. It can be defined as the likelihood of influencing factors on a specific outcome. The Logistic regression model is an important model for assessing personal traffic behavior in the field of road traffic. It can analyze the impact of one or more influencing factors on a non-numerical classification result, and more accurately and comprehensively describe the decision-making behavior of individuals or groups, has achieved relatively rich research results. This paper applies it to the field of truck transportation safety analysis, employing a binary logistic regression model and a truck driver’s driving behavior selection model based on the model theory. The model is constructed and calibrated using personal information collected from truck drivers via online questionnaire surveys.

The driving dependent variable y of the model is a binary variable with values of 1 and 0, and x is a risk factor that affects y. Let the probability of y = 1 under the condition of x be:

$$ P = P\left( {{\text{y = }}1{\text{|x}}} \right){ = }\frac{{{\text{e}}^{{\upalpha +\upbeta {\text{x}}}} }}{{1{\text{ + e}}^{{\upalpha +\upbeta {\text{x}}}} }}{ = }\frac{{\exp \left( {\upalpha +\upbeta {\text{x}}} \right)}}{{1 + \exp \left( {\upalpha +\upbeta {\text{x}}} \right)}} $$
(1)

This article mainly adopts the binary logistic regression model, and its mathematical model is:

$$ P = P\left( {{\text{y = }}1{\text{|x}}} \right){ = }\frac{{{\text{e}}^{{\alpha + \beta {\text{x}}}} }}{{1{\text{ + e}}^{{\alpha + \beta {\text{x}}}} }}{ = }\frac{{\exp \left( {\upalpha +\upbeta _{1} {\text{x}}_{1} { + }\upbeta _{2} {\text{x}}_{2} +\upbeta _{2} {\text{x}}_{2} + \cdot \cdot \cdot\upbeta _{{\text{K}}} {\text{x}}_{{\text{X}}} } \right)}}{{1 + \exp \left( {\upalpha +\upbeta _{1} {\text{x}}_{{1}} +\upbeta _{2} {\text{x}}_{2} + \cdot \cdot \cdot\upbeta _{{\text{K}}} {\text{x}}_{{\text{X}}} } \right)}} $$
(2)

2.2 Questionnaire Design and Survey

Questionnaire Design.

The author designs a questionnaire based on some phenomena existing in reality and combines them with existing related research. According to Song Xiaolin et al.’s examination of connected accidents, men were responsible for a higher proportion of road accidents caused by speeding than women [10]. Lourens et al. found that age is related to drivers’ violations [7]. Chuang and Wu found that sleep problems can cause stress in professional drivers [6]. Salar Sadeghi Gilandeh found that driving behavior is related to road conditions [5]. Gender, age, education level, years of employment, personality, household registration, driver’s license level, and other factors are combined in this article to create a questionnaire with a total of 18 factors, including the truck driver’s gender, age, education level, years of employment, personality, household registration, driver’s license level, and so on.

From a psychological point of view, the driver’s personality is divided into depressive qualities (sensitive, frustrated, withdrawn, indecisive, slow recovery from fatigue, slow response), and bloody (calm, tolerant, focused and hardworking, patient and hardworking. But inflexibility, lack of enthusiasm, conservatives), mucus quality (enthusiasm, ability, adaptability, wit, lack of focus, changeable emotions, lack of patience), bile quality (excited, short-tempered, straightforward, enthusiastic, But the mood is lower when the energy is exhausted).

Data Acquisition and Processing.

In order to improve the accuracy of the data, this survey uses the real-name system to fill in the blanks. In order to meet the universality, we chose to put the questionnaire online and send the link to the truck driver through the truck company in Anhui Province to collect the questionnaire. Truck drivers are required to fill out the questionnaire objectively and impartially. The business managers will answer the questions that the driver has. Finally, a total of 1354 papers have been filled out. There is no invalid questionnaire due to the driver’s personal reasons, and the effective questionnaire is 100%.

3 Establishment and Improvement of Driving Violation Behavior Model

3.1 Descriptive Statistical Analysis

Truck drivers are the subjects of this study. According to statistics, a total of 1354 people were investigated, including 938 people who violated regulations and 416 people who did not. There are 1331 male drivers and 23 female drivers (Table 1).

Table 1. Driver’s statistical information.
Full size table

3.2 Reliability Analysis

In this paper, the Cranbach α coefficient is used to analyze the reliability of the questionnaire through SPSS 23.0 software, and the calculation result is α = 0.143 (Table 2).

Table 2. Driver’s statistical information.
Full size table

The SPSS 23.0 software was used to analyze the validity of the questionnaire, and the results are shown in Table 3. The KMO coefficient is 0.680, which is greater than 0.50, and the Sig value is 0.00, which is less than 0.05. Therefore, factor analysis can be performed.

Table 3. Kmo and Bartlett test.
Full size table

3.3 Logistic Model Analysis

The Choice of Dependent and Independent Variables.

Based on whether truck drivers violate the regulations, the total number of people is planned to be classified into two types: violation and non-violation. The value of the dependent variable Y is shown in the table below. As shown in Table 4, according to the questionnaire data, all items are set as independent variables (X).

Table 4. Dependent variable.
Full size table

Initially, we used the SPSS 23.0 software to perform binary logistic regression analysis on 18 factors, with a significance level of = 0.05 and the forward LR method (forward stepwise regression method based on maximum likelihood estimation). First, use the score test method to screen the independent variables. According to whether the p value corresponding to the score value meets the given significance level, the variables that meet the requirements are initially selected as shown in Table 5.

Table 5. Score test result.
Full size table

Determine the significance of all the influencing factors according to the preliminary test, and then gradually substitute all the influencing factors into the equation. When the parameter estimation value changes by less than 0.001, the estimation is terminated at the 7th iteration, and the following results are initially obtained, as shown in Table 6.

Table 6. Model (if item is removed).
Full size table

Model Checking.

In this comprehensive test of the binary logistic regression model coefficients, one line of the model outputs the likelihood ratio test results of whether all the parameters in the logistic regression model are 0, as shown in Table 7. Where the significance level is less than 0.05, it means that the OR value of at least one of the included variables in the fitted model is statistically significant, that is, the model is overall meaningful.

Table 7. Comprehensive test of model coefficients.
Full size table

In this paper, Hosmer and Lemeshow tests are used to test the goodness of fit of the model, and the calculated significance level is 0.781 > 0.005, which indicates that the model fits well, as shown in Table 8.

Table 8. Comprehensive test of model coefficients.
Full size table

After preliminary fitting model calculations, six factors including personality, driver’s license level, daily driving time, whether there is a fixed transportation route, vehicle ownership, and whether there is an occupational disease are selected from the analysis results, and SPSS 23.0 software is used to target these six factors. Perform binary Logistic regression analysis, select the significance level α = 0.05, and use the input method. The final result is consistent with Table 6. In the comprehensive test of model coefficients, the significance level is less than 0.05, indicating that the model is meaningful in general. In the Hosmer and Lemeshow test, the significance level is 0.731 and greater than 0.05, indicating that the model fits well. It can be seen that the truck driver’s personality, driver’s license level, daily driving time, whether there is a fixed route, the ownership of the vehicle, and whether there is an occupational disease have a significant impact on the driver’s traffic violations.

4 Discuss

Based on the data from the questionnaire survey, a binary logistics model for truck drivers is established for comprehensive analysis. In this section, the author will discuss the relevant results of other scholars on the factors that affect drivers’ traffic violations, and compare the results of this article to get more information and practical suggestions.

According to previous related research, personality is divided into depressive, bloody, mucous, and bile (easily excited, short-tempered, straightforward, enthusiastic, but depressed when energy is exhausted). According to previous related studies, the driver’s personality changes from depression to bile, and the driving speed is getting faster and faster. The number of people with bloody and mucous personalities is the highest among them [12, 13], and this survey confirms this.The situation is roughly the same. The bloody personality has the most people in this article, with 643 people, accounting for 47.49% of the total number of people, 432 of whom have broken the rules, accounting for 67.19%; the mucus personality has 326 people, accounting for 24.08% of the total number of people, and 231 of whom have broken the rules. People accounted for 70.86%; 308 people with biliary personalities accounted for 22.75% of the total, with 225 of them having 73.05% violations; and depressive personalities affected 77 people, or 5.69% of the total, with 48 of them having major depression.Violations made up 62.34% of the total.The significant difference between drivers with bloody and biliary personalities is bigger, implying that drivers with biliary personalities are more prone to committing infractions while driving, and that drivers with biliary personalities require special attention at work. To strengthen their self-control and avoid traffic offenses caused by high-speed driving, such people must be supervised.

The driver’s license level is quite different in the model of the truck driver’s personal attributes and violation behavior (significance = 0.014). The investigated truck driver obtained primarily A2 driver’s licenses, with a total of 867 people, accounting for 64.03% of the total.Among them, 602 people have violated regulations, accounting for 69.43%; the second is the B2 driver’s license type, with a total of 380 people, accounting for 28.06% of the total, of which 254 people have violated the regulations, accounting for 66.84%; and the C driver’s license type has a total of 45 people, accounting for 3.32% of the total, of which 26 people have violated the rules, accounting for 57.78%. With the trend toward larger vehicles, truck drivers with A2 licenses have increasingly become the mainstream. At present, driving a tractor requires an A2 driver’s license, which must be increased on the basis of obtaining a B2 driver’s license. It is not possible to directly apply for the test, and a motor vehicle that is driven during the internship period is not allowed to tow a trailer. Due to the high cost of taking photos, it is also one of the reasons why it is difficult to attract young practitioners to enter. At present, some auto manufacturers have introduced automatic tractors, but they have to apply for an A2 driver’s license.

In the past, a large number of relevant studies have shown that fatigue driving is one of the important causes of traffic accidents [14]. There are also many reasons for fatigue driving. Among them, the driver’s perceptual reaction time and the ability to maintain attention increase with the driver’s drowsiness. Sleep is reduced [15], and daily driving time is also one of the important factors that make people fatigued. This article divides the daily driving time into 8 h or less, 8–10 h, 10–12 h, and 12 h or more. There were 632 people under 8 h, accounting for 46.68% of the total, of which 230 offenders accounted for 36.39%; there were 397 people under 8–10 h, accounting for 29.32% of the total, of which 157 offenders accounted for 39.55%; 190 people in 10–12 h, accounting for 14.03% of the total number of people, of which 65 offenders accounted for 34.21%; and 135 people over 12 h, accounting for 9.97% of the total number, accounting for 9.97% of the total number, of which 102 offenders People accounted for 75.56%. The special working environment of truck drivers makes them generally work longer hours and be labor-intensive. 53.32% of truck drivers drive 8 h or more per day, and there is a risk of fatigue driving, which may lead to violations.

Whether there is a fixed transportation route is quite different in the model of a truck driver’s personal attributes and violation behaviors (significance = 0.050). There are 730 people with fixed transportation routes, accounting for 53.91% of the total, of which 488 people are in violation. It accounted for 66.85%; there were 624 people without fixed transportation routes, accounting for 46.09% of the total number, of which 448 people who violated regulations accounted for 71.79%. There is a higher rate of violations without fixed transportation routes, which may be due to driving on an unfixed road section, leading to traffic accidents due to unfamiliar road conditions when driving. It shows that different driving environments have a greater impact on the driver.

In the model of a truck driver’s personal attributes and violation behavior, vehicle ownership and whether there is an occupational disease are very different, and the significance is 0.000.The survey shows that 76.88% of truck drivers report that their vehicles are self-owned vehicles, 39.81% of which are currently in the process of repaying their loans, and only 23.12% of truck drivers drive vehicles that belong to their employer or fleet. Self-employed truck drivers are still more common, with back-loan drivers taking up more space. There are 502 people without arrears in their own vehicles, accounting for 37.07% of the total, of which 357 people are in violation of the rules, accounting for 71.12%; 539 people are in arrears with their vehicles, accounting for 39.81% of the total, and among them, 417 are in violation of the rules. People accounted for 77.37%; there were 313 hired drivers, accounting for 23.12% of the total number, of which 162 offenders accounted for 51.76%. At present, there is a “0” down payment model in the truck sales market. Financial companies use ultra-low threshold “0” down payment or low down payment methods to attract a large number of truck drivers to enter the freight market. Financial companies turn the down payment burden into high monthly payments and high fees (maintenance, etc.), which increases purchase costs. At the same time, drivers are required to attach their vehicles to the anchoring company and charge higher anchorage fees, insurance premiums, and inspection fees, which further increases the driver’s burden. Affiliated companies can obtain a large number of vehicle input invoices and transfer them to other markets. At the same time, when the loan expires, the driver asks to transfer the vehicle out, generally facing the problem of a high transfer-out fee. The survey shows that 56.13% of truck drivers suffer from one or more occupational diseases such as stomach disease, cervical spondylosis, and back pain due to long-term driving. A total of 760 people have occupational diseases. Among them, 559 people who violate regulations account for 73.55%. The health problems of truck drivers are worth causing. focus on. 43.87% of truck drivers did not have the above-mentioned health problems because of their low working years or short driving time each day. There were 594 people without occupational diseases, of which 377 people who violated regulations accounted for 63.47%.

People usually think that age and driving age are very related to drivers’ violations. Leixing et al. found that as the driver’s age changes, his driving behavior will change accordingly, which will affect driving safety [16]. Fang Yuerong believes that drivers between the ages of 40 and 52 have relatively slow driving speeds, more stable driving behaviors, and safer driving [17]. The research in this article found that age and driving age have no obvious relationship with whether truck drivers have traffic violations.

5 Conclusion

This article investigated the personal attributes and violations of truck drivers and obtained 1354 traffic violation data samples. The driver’s infraction data was mined and evaluated using the logistics model, and the following findings were drawn:

  1. (1)

    Whether truck drivers will violate the rules is significantly related to six variables: personality, driver’s license level, daily driving time, whether there is a fixed transportation route, vehicle ownership, and whether there is an occupational disease. Among them, personality, daily driving time, whether there is a fixed transportation route, and vehicle ownership are positively related to violations.

  2. (2)

    Further data analysis shows that this group of people who are bile, drive more than 12 h a day, have no fixed transportation routes and have loans for their own vehicles are most likely to have violations during the driving process, which can be further improved in the future. Investigate and research this part of the group. When hiring drivers, relevant departments can conduct personality tests. They can strengthen management and coaching for this portion of the group among the existing truck drivers.

6 Practical Implications and Directions for Further Research

In this study, there is no guarantee that the data filled in by the surveyed persons when filling out the questionnaire is authentic. Some people have personal subjective emotions when filling out the questionnaire, which leads to a certain deviation in the data filled in. Therefore, in the future research work, it is necessary to adjust the existing survey methods.

The data obtained from the questionnaire survey in this article has certain deficiencies. Among them, there are too few female drivers and they are not representative. The sample data is not enough, it can only represent part of the truck drivers in Anhui, and cannot distinguish the personal attributes of the drivers in the plain area and the mountain forest area. The dependent variables used in this model are divided into two types of violations and non-violations. In future research, violations can be divided into high-risk violations and low-risk violations in more detail, so that truck drivers in different regions can be studied in detail.