Research on Risk Assessment and Intervention Methods of Freight

Vehicle Traffic Safety via Data Analysis

Bowen Wang

, Jingsheng Wang

1,*

, Tianyu Xia

and Dandan Ding

School of Traffic Management, People's Public Security University of China, Beijing, China

School of Information Technology and Network Security, People's Public Security University of China, Beijing, China

Keywords: Traffic Safety; Freight Vehicles; Vehicle Factors.

Abstract: In order to deeply explore the traffic safety risks of freight vehicles, the data of freight vehicles in Ningxia

Autonomous Region is taken as an example. Through the analysis of the driver's traffic behavior and the

condition of the freight vehicle, the XGBOOST model is established to help the relevant departments

reasonably understand the driver's traffic safety risks. The analysis results show that the classification

accuracy of the XGBOOST model is 86%, and the fitting effect is good. In an analysis of driver traffic

behavior, drivers were more likely to cause more serious fatal accidents when they misused their turn signals

and lights. Using the wrong turn signal is 1% more likely to result in a fatal accident than using the right turn

signal. In the analysis of vehicle conditions, the legal and safety conditions of trucks in Ningxia Autonomous

Region are not good, and the vehicles have potential safety hazards. Relevant departments should strengthen

safety issues such as scrapped vehicles, irregular inspections, poor lighting, poor braking, and faulty signaling

devices.

1 INTRODUCTION

Heavy trucks have the characteristics of large

transportation volume, extensive management, lack

of monitoring channels, unfixed transportation

routes, and high mobility of vehicles (Park 2019). In

the process of driving heavy freight vehicles, due to

their own factors such as large body size, frequent

overloading, speeding, and weak safety awareness of

some drivers, they are very easy to cause heavy and

very serious traffic accidents, which directly affect

the property interests and safety of the people

(McDonald 2019). According to statistics from the

Traffic Management Bureau of the Ministry of Public

Security, the current domestic truck ownership

accounts for about 8% of the motor vehicle

population, and the death toll from truck accidents

accounts for about 30% of the total accident deaths.

The truck accident rate per 10,000 vehicles is higher

than the national traffic accident rate per 10,000

vehicles. More than twice as high, the frequency and

severity of truck accidents are significantly higher

than other motor vehicles, and road transportation

safety is very severe. Frequent traffic accidents of

freight vehicles have brought huge challenges to the

supervision of freight traffic safety (Muratori 2017,

Ruesch 2016, Wang 2021, Wang 2021). People are

increasingly realizing that it is urgent to prevent and

reduce freight vehicle traffic accidents and implement

road traffic safety management scientifically and

effectively (Taylor 2019, Wang 2021, Wang 2021).

Therefore, in order to dig deeper into the traffic

safety risks of freight vehicles, take the data of freight

vehicles in Ningxia Autonomous Region as

examples, analyze the traffic behaviors of drivers and

the status of freight vehicles to help relevant

departments rationally understand the traffic safety

risks of drivers.

2 ANALYSIS OF ILLEGAL DATA

ON FREIGHT VEHICLES

2.1 Data Source and Preprocessing

Page Setup

The two data sets studied in this paper come from the

databases of the public security traffic management

departments of Ningxia Autonomous Region. First,

the acquired data is preprocessed. In this article,

Wang, B., Wang, J., Xia, T. and Ding, D.

Research on Risk Assessment and Intervention Methods of Freight Vehicle Trafﬁc Safety via Data Analysis.

DOI: 10.5220/0011345700003437

In Proceedings of the 1st International Conference on Public Management and Big Data Analysis (PMBDA 2021), pages 373-378

ISBN: 978-989-758-589-0

373

duplicate samples and samples with more vacancies

are deleted.

2.2 Data Analysis

2.2.1 Irregular Use of Turn Signals

Vehicles, especially trucks with a large volume and

carrying capacity, are more likely to cause traffic

accidents when turning, changing lanes, U-turning,

overtaking, sudden braking, sudden start and other

special behaviors than when driving in a straight line.

As one of the traffic signals, the normal use of the

vehicle turn signal can remind other road users that

the vehicle is about to turn, change lanes, and turn

around, so that other road users can take measures

such as slowing down and avoiding in advance to

reduce traffic accidents. Taking the Ningxia

Autonomous Region as an example, the result of

analyzing the use of the turn signal of the accident

vehicle is shown in fig.1. There were a total of 305

truck traffic accidents, of which 200 were accidents

with the correct use of vehicle steering lights,

accounting for 66% of the total number of accidents.

There were 105 accidents with incorrect use of turn

signals, accounting for 34% of the total number of

accidents. In accidents with incorrect use of turn

signals, there are situations such as turning on the left

turn signal when the vehicle is going straight, and

double flashing when turning left and right. The

wrong signal is transmitted to other traffic

participants, which is a hidden danger to the

occurrence of traffic accidents.

Figure 1: The use of vehicle turn signal.

Further analysis of the use of the turn signal of the

accident vehicle, the distribution of accidents

corresponding to the correct and wrong use of the turn

signal is shown in fig.2 and fig.3 respectively. In the

data of accident vehicles, there were 97 fatal

accidents out of 200 accidents when the turn signal

was used correctly, accounting for 49% of the total

number of accidents when the turn signal was used

correctly. When the turn signal is used incorrectly,

there are 53 fatal accidents in 105 accidents,

accounting for 50% of the total number of accidents

when the turn signal is used incorrectly. The use of

the wrong turn signal is 1% higher than the

probability of a fatal accident when the turn signal is

used correctly.

34%

66%

Vehicle turn signal usage

Wrong use of vehicle turn signal Correct use of vehicle turn signal

PMBDA 2021 - International Conference on Public Management and Big Data Analysis

374

Figure 2: Distribution of accident types when vehicle turn signals are used correctly.

Figure 3: Distribution of accident types when vehicle turn signals are used incorrectly.

2.2.2 Irregular Use of Lighting

The use of low-beam, high-beam, and position lights

of vehicles can provide lighting and reminders for

vehicles under poor road lighting conditions, meeting

vehicles, etc., but the vehicle's low-beam, high-beam,

and position lights are wrong The use will not only

produce poor lighting effects for the driver, but also

its adverse effects on other road traffic users is also

one of the factors leading to traffic accidents. Take

Ningxia Autonomous Region as an example to

analyze the lighting usage of the truck in the accident,

and the result is shown in fig.4. There were a total of

305 truck accidents, of which 215 were accidents with

correct vehicle lighting, accounting for 70% of the

total number of accidents. There were 90 accidents

involving incorrect use of lighting, accounting for

30% of the total number of accidents. In the accidents

of incorrect use of lighting, there are situations in

which vehicles use high beams illegally when there

are street lights at night, and do not use lights when

there are no street lights at night. Wrong use of high

beam headlights can cause instant blindness to the

opposing driver. Do not use lights under poor lighting

conditions, which are all hidden dangers that lead to

traffic accidents.

34%

66%

Vehicle turn signal usage

Wrong use of vehicle turn signal Correct use of vehicle turn signal

46%

50%

Type distribution of accidents when vehicle turn signals are used

incorrectly

Property damage accident Wounding accident Fatal accident

Research on Risk Assessment and Intervention Methods of Freight Vehicle Trafﬁc Safety via Data Analysis

375

Figure 4: Use of vehicle lighting.

2.3 Vehicle Status

Taking Ningxia Autonomous Region as an example,

the result of the analysis of the legal status of the

accident vehicle is shown in fig.5. There were 305

truck accidents, of which 293 were accidents with

vehicles in a legal status, accounting for 96% of the

total number of accidents. There were 12 accidents in

which vehicles were illegal, accounting for 4% of the

total number of accidents.

Figure 5: Analysis of the legal status of vehicles.

Further analysis of the illegal status of the vehicle,

the illegal status of the vehicle is shown in fig.6.

Among the 12 accidents where vehicles were illegal,

1 vehicle was scrapped, 3 vehicles were not inspected

on time, and 8 vehicles were other reasons.

Figure 6: Analysis of illegal conditions in vehicles.

34%

66%

Vehicle turn signal usage

Wrong use of vehicle turn signal Correct use of vehicle turn signal

96%

Legal status of the vehicle

Abnormal condition Normal condition

Illegal condition of the vehicle

Scrapped Other reasons

Unscheduled inspection

PMBDA 2021 - International Conference on Public Management and Big Data Analysis

376

The results of the analysis of the safety status of

accident vehicles in Ningxia Autonomous Region are

shown in fig.7. There were 305 truck traffic accidents,

of which 278 accidents were in a legal state,

accounting for 91% of the total number of accidents.

There were 27 accidents in which vehicles were

illegal, accounting for 9% of the total number of

accidents.

Figure 7: Analysis of the legal status of vehicles.

Further analysis of the unsafe condition of the

vehicle, the unsafe condition of the vehicle is shown

in fig.8. Among the 27 accidents in which vehicles

were in unsafe conditions, one was due to leakage of

oil/liquid/gas from the vehicle, one was due to failure

of vehicle lighting and signaling devices, and 11 were

due to poor vehicle braking or failure.

Figure 8: Analysis of illegal conditions in vehicles.

2.4 Model Establishment

Using Python, using Jupyter lab as a tool, call the

XGBClassifier function of the xgboost package to

build the XGBOOST model. All parameters use

default values.

The accident type is used as the label, and the

value types are: property damage accident and

casualty accident. The legal status of the vehicle, the

safety status of the vehicle, the usage of turn signals,

lighting conditions and other factors are input into the

model as features.

It is concluded that the classification accuracy of

the model on the test set reaches 86%. The model has

excellent fitting effect and can be used to identify the

severity of traffic accidents

3 CONCLUSION

This article is to dig deeper into the traffic safety risks

of freight vehicles. Taking the freight vehicle data of

Ningxia Autonomous Region as an example, the

following conclusions can be drawn by analyzing the

91%

The safety of the vehicle

Abnormal condition Normal condition

Unsafe conditions in the vehicle

Other reasons

Leaking oil/liquid/gas

Failure of lighting and

signaling devices

Poor braking

Research on Risk Assessment and Intervention Methods of Freight Vehicle Trafﬁc Safety via Data Analysis

377

traffic behavior of drivers and the status of freight

vehicles:

(1) In the driver's analysis of traffic behavior, it is

more likely to cause more serious fatal accidents

when the driver misuses the turn signal and lighting.

(2) In the analysis of the vehicle status, the legal

status and safety status of trucks in Ningxia

Autonomous Region are not good, and there are

hidden safety hazards in the vehicles. The relevant

departments need to strengthen the disposal of

vehicles, unscheduled inspections, poor lighting, poor

braking, and signaling devices.

(3) The classification accuracy of the XGBOOST

model on the test set reaches 86%. The model has

excellent fitting effect and can be used to identify the

severity of traffic accidents

REFERENCES

McDonald N., Yuan Q., Naumann R. (2019). Urban freight

and road safety in the era of e-commerce. J. Traffic

injury prevention,20(7): 764-770.

Muratori M., Holden J., Lammert M., et al. (2017).

Potentials for platooning in US highway freight

transport. National Renewable Energy Lab, Golden,

CO (United States).

Park, J., Abdel-Aty, M., Wang, L., et al. (2019). Influence

of multiple freeway design features on freight traffic

safety. J. Journal of Advanced Transportation, 2019.

Ruesch M., Schmid T., Bohne S., et al. (2016). Freight

transport with vans: Developments and measures. J.

Transportation Research Procedia, 12: 79-92.

Taylor, S. G., Russo, B. J., James, E., (2018). A comparative

analysis of factors affecting the frequency and severity

of freight-involved and non-freight crashes on a major

freight corridor freeway. J. Transportation research

record, 2672(34): 49-62.

Wang B., Wang J., Zhu Y.，et al. (2021). Research on Short-

term Traffic Flow Prediction Model Based on ARMA-

SVR. J. Journal of Highway and Transportation

Research and Development,2021,38(11):126-133.

Wang B., Wang J., Wang B., et al. 2021. Research on

Quantitative Method of Traffic Safety Credit Score

Based on Ridge-Logistic Regressio. International

Conference on Man-Machine-Environment System

Engineering. Springer, Singapore: 746-752.

Wang B., Wang J., Zhang Z., et al. 2021. Traffic Flow

Prediction Model Based on Deep Learning.

International Conference on Man-Machine-

Environment System Engineering. Springer, Singapore:

739-745.

Wang, B., Wang, J., Zhu, Y., et al. ,Analysis of Driver

Causative Factor Tendency Based on K －means ++. J.

Science Technology and Engineering, 2021,21(36):

15695-15699.

PMBDA 2021 - International Conference on Public Management and Big Data Analysis

378