Study of Collaborative Management of Intelligent Highway Tunnel

HAN Zhi, GUAN Feifei

(1. China Merchants Chongqing Communications Technology Research & Design Institute Co., Ltd.,Chongqing 400067, China; 2. Chongqing Jiaotong University, Chongqing 400074, China)

0 Introduction

Intelligent highway[1]is a hot topic of current social concerns, and it is same for intelligent tunnel. The intelligence highway should include traffic organization intelligence, vehicle intelligence, facility intelligence, and service intelligence as the final goal. At present, the researches of intelligent highway are focused mainly on car networking, road networks, and unmanned automatic driving technology[2-4]. For the informatization and intelligence of highway tunnels, current researches are focused on information construction[5]and intelligent monitoring[6]. In addition, intelligent ventilation and intelligent lighting are also hot topics in the research of intelligent tunnel. For example, the project "Research on key technology of tunnel construction and management in Zhongnan Mountain of Qinling" focuses on intelligent ventilation. The patent "A highway tunnel lighting control method" proposes a three-element control method[7-8]for intelligent lighting. When it comes to intelligent tunnels, little attention is paid to the differences between tunnels and roads. Comparing to road sections, tunnels are semi-closed and the lateral net width is relatively smaller, which gives tunnels five distinct characteristics that are different from highway sections: (1) traffic capacity is lower than that of road sections, (2) operation risk is higher than that of road sections, (3) the number of facilities is larger than that of road sections, (4) accident and disasters are more severe than that of road sections, and (5) management is more difficult than in road sections.

As a result, national and local governments in China have carried out numerous studies related to "safety, efficiency, economy" for nearly 30 years, such as: the National 863 Program "Tunnel LED lights" , the Ministry of Transportation Project in the west of China "Research of Shanghai-Chengdu west highway′s safety and energy saving technology", the Ministry of Transportation′s joint research project "Guang-Wu Highway safety and energy saving technology research", the key project of Chongqing Science and Technology Commission "Research and demonstration of multi-level road network in the core area of Chongqing city" (project No. : cstc2015shms-ztzx30014). These studies have solved some common problems and engineering issues in the industry, promoting technological progress. However, due to the limitations of technology, there are still five problems need to addressed: (1) information collection dimension is insufficient; (2) information processing depth is insufficient; (3) the information transmission scope is not wide; (4) information integration platform is not intelligent; (5) information application is incomplete.

With the development and application of the Internet of Things, Cloud Computing and Big Data, it is possible to solve the above problems and make the tunnel more intelligent. However, to truly realize the intelligent tunnel, we need to realize the intelligence of traffic organization, the intelligence of facilities and the intelligence of vehicles. At present, it is impossible to realize the intelligence of vehicles, but realize the intelligence of traffic organization and the intelligence of facilities partially. In this paper, starting from intelligent connotation, the characteristics of intelligence is discussed firstly. Secondly, from the key nodes of highway traffic, a collaborative intelligent tunnel control of network topology structure is proposed, which refers to the information interaction between people, cars, roads and environment, making decisions based on information, and achieving optimal management. Then, considering the factors involved in collaborative control, an information system of intelligent tunnel is built. Finally, mathematical models are recommended according to the requirements of solving key problems in information fusion.

1 The connotation of intelligence

"Intelligence" belongs to the category of cognition, although some people argue that "Intelligence" is one of the unfathomed issues remaining for the past 500 000 years of human history, we can conclude that "intelligence" should contain the following features to realize the collaborative management of intelligent tunnel from the narrow definition of intelligence.

1.1 Identification and perceptual ability

Identification refers to the classification and characterization of objects and conditions according to a particular rule in the application of norm conflicts, e.g. the road structure, whether or not the equipment is in good working condition, the traffic condition is safe or unimpeded, and the condition of the tunnel environment is up to standard. Perception refers to skills of detection, induction, observation, etc. for characters and images, such as vehicle position, equipment and structure condition, driver′s driving status, obstacles in the tunnel, vehicle′s performance and the type of goods being transported. Through identification and perception, risk and danger can be reduced, and the safety, efficiency and economy of the tunnel during operation can be improved.

1.2 Learning ability

Automatic self-learning machines, e.g. "alpha-go" is able to become much more intelligent than before by learning the game of go for 3 months. Another machine like "alpha zero" is able to defeat "alpha-go" with 100∶0 by automatically exploring, gaining experience and enhancing its skills according to a set of rules in three days of learning.

1.3 Predictive ability

Predictive ability is about extrapolating the future according to methods and rules, so as to understand the development process and results in advance. Such as, through the prediction of the state of facilities and structures, preventive maintenance can be made ahead of schedule to make maintenance work more economical and reliable; through the prediction of the time and space of traffic, we should take corresponding control measures in advance so as to make traffic operation safer and more smooth; through the prediction of the damage degree, environmental parameters and time and space variations of traffic accidents, we should take corresponding rescue measures in advance so as to bring less harm to traffic accidents; and so on.

1.4 Optimization ability

Optimization refers to bad results and relative optimum disposal measures determination according to the current situation of things and the situation to be formed and based on prediction ability. When targeting safety, efficiency and economy, we should make corresponding measures to achieve the optimum maintenance and management of road and tunnel structure, so that the ventilation control, traffic control, lighting control, disaster prevention and mitigation can achieve best state.

2 Experience and rules

Experience refers to the skills and knowledge obtained through practice, and rules refer to the principles that are followed during operation. For transportation, experience refers to standards, norms, rules and guidelines. Rules refers to norms such as driving on the right, forbidding overtaking in the tunnel, and driving distance between vehicles should be greater than the dynamic braking distance. Different rules can be formulated around the target of "safety, efficiency and economic".

2.1 Safety rules

The safety of tunnel operation mainly includes three parts, i.e. safety of structures, facilities and traffic operation.

2.1.1 Structural safety

The safety of the tunnel structure depends on materials, construction process, environment and other factors. The time and space changes of stress and strain can be used to reflect the structural safety condition. Currently, material stress, strain variation, the relationship between axial strain and the lateral strain, and the relationship between stress and strain can be determined, which can be used as a rule for structural safety.

2.1.2 Safety of mechanical and electrical facilities

Mechanical and electrical facilities include outdoor facilities and indoor facilities. The safety of outdoor facilities includes the safety of installation and the utility of facility itself, in other words, whether the facility is malfunctioning or working properly. There are detachment of fan impeller, collapse of cable bridges and other accidents in the tunnel. These phenomena are not only related to construction and installation, but also related to the quality of work environment and facilities, such as vibration and corrosion. Therefore, the safety rules of facilities should be an evolutionary mechanism under the action of facilities and operating environment, but the difference is that outdoor facilities are affected by more factors.

2.1.3 Traffic operating safety

The causes of traffic accident in highway tunnel include the design defect of tunnel itself, the unreasonable factor of vehicle itself, the lack of tunnel operation and management, and incorrect behavior of road users. Traffic accidents in highway tunnel include tail collision, wide-swipe accident, collision with tunnel walls and fire disaster. The fire disasters are caused by the collision, the spontaneous combustion of vehicles and the explosion of loading. The rules are as follows:

(1) Tail collision and side collision

The distance between vehicle (including lane-changing behavior) is greater than the dynamic braking distance[9](as shown in Fig. 1), i.e.:

S(i-1,1)=Sr+Szi-Si-1+l0

(1)

whereS(i-1,1): the dynamic braking distance between cari-1 and cari;Sr: driver′s reaction distance;Szi: the braking distance of car i;Si-1: the sliding distance of the car when cari-1 start to brake;l0: the distance between the front of cari-1 and the end of cari.

Fig. 1 Dynamic braking distance diagram

(2) Collision with the tunnel wall

The distance between the vehicle and the tunnel wall is less than the lateral safety clearance.

(3) Self-igniting of loading(or explosion)

The change mechanism of the loading material with the environment and vehicle movement (vibration, velocity, etc.).

(4) Vehicle operation

Vehicle operation includes right driving, signal rules, the rules of marks and traffic line etc.

(5) Vehicle performance

Vehicle performance includes tires, engine, steering wheel and the mechanism with the vehicle movement etc.

(6) Tunnel operating environment

Tunnel operating environment mainly refers to the operating environment index of tunnel, the specified value of CO and Ⅵ under different working conditions (normal, fire, maintenance, etc.)[10].

(7) Driver

This includes the mechanism between driver′s physical condition and vehicle operation.

2.2 Efficient rules

At present, there is no authoritative definition about the efficiency of traffic operation. The section parameters and interval parameters of traffic flow characteristics can be used as indicators to evaluate traffic operation status, including average travel speed, service level, queue length and delay. These indicators are related to each other and not separate, for every index just reflect the traffic running state from a side view. For example, the service level is the ratio of cross-section traffic volume to cross-section capacity, and it is difficult to calculate the ratio of the traffic volume to the capacity of a particular line or road network accurately, while other indicators are the same. The rules cannot be made without goals. As a result, the concept of traffic efficiency is proposed in this paper, which can be expressed as

E=Qαvβ

(2)

whereEis traffic efficiency,Qis traffic volume,vis average travel speed, andαandβare weight coefficients.

The Equation (2) can be used as a rule for high efficiency of transportation. The coefficients ofαandβcan be obtained through learning and training.

2.3 Economic and comfort rules

The economy and the comfort of tunnel operation mainly include the ventilation control system and lighting control system, referring toHighwayTunnelEnergySavingTechnology[11], according to which the corresponding rules can be established.

2.4 Rules between economy and safety

These three goals of safety, efficiency and economy are not at the same level. Safety is at the first level, while efficiency and economy belong to the second level. There is a correlation between safety and economy. Generally speaking, the greater the economic investment is, the better the safety will be. However, the investment is always limited, so there is a balance between the acceptable risk and the investment[12], as shown in Fig. 2.

Fig. 2 Economy-safety relation diagram

This rule can be established based on the relationship between safety and economy.

3 Traffic organization intelligence

The intelligence of traffic organization is based on traffic flow characteristics between its scope of jurisdiction and adjacent sections. The structural form refers to the types of road topology mentioned below. There is no clear definition of the scope of tunnel, but it can be determined according to the location of tunnel facilities. According to the relevant specifications, the camera out of tunnel is located about 100-400 m[10]or 100-250 m[13]from the entrance of tunnel. The sonic and light alarm is located about 50-250 m[14]from the entrance of tunnel . When there is no contact channel, the signal light[10]is set at a parking sight distance from the entrance of tunnel. According to these rules, tunnel′s management scope contains the length range of tunnel itself and the space which is about 50-400 m from the entrance of tunnel.

Tunnel is a part of the road to achieve the goal of "safety, efficiency and economy", and the traffic characteristics of the tunnel′s adjacent road should be considered. In this paper, the highway is considered to consist of toll stations (mainline toll stations, ramp toll stations and service area ramp), highway sections (including bridges) and tunnel sections. According to the road node, traffic condition, and tunnel traffic facilities, the following types of road topology are given.

3.1 Normal traffic

When the traffic is normal, there are four main types of road topology, as shown in Figs.3, 4, 5 and 6.

Fig. 3 Toll station-ramp tunnel topology

Fig. 4 Toll station-toll station tunnel topology

Fig. 5 Ramp-ramp tunnel topology

Fig. 6 Ramp-toll terminal tunnel topology

3.2 Abnormal traffic

Abnormal traffic mainly includes fire, traffic jams and traffic accidents.

3.2.1 Fire

The vehicles′ horizontal channel and pedestrian′s crossing channel will be opened when fire happens. At this time, part of the horizontal channel can only be used for firefighting and rescue vehicles, part of the horizontal channel can be used for people′s escaping. The connecting channel may be open for the vehicles out of the tunnel to turn around, or may not be opened. Its topological structure is shown in Figs. 3, 4, 5, 6 and 7.

3.2.2 Traffic jams

The tunnel accident is very dangerous, so evacuation should be executed as soon as possible to avoid traffic jams. It is a most effective measure to limit the number of vehicles entering the tunnel. At this time, the road topology is shown in Figs. 3, 4, 5, 6 and 8.

3.2.3 Traffic accidents

Traffic accidents in the tunnel can be divided into two types. One is that the accident vehicle needs trailer, while the other one does not need. The latter road topology can refer to Fig. 8. For the former road topology, Fig. 8 can be referred to if the traffic volume is not large, otherwise Fig. 7 should be referred to.

Fig. 7 Tunnel topology of fire

Fig. 8 Tunnel topology of traffic jams

4 Information collection and processing

Intelligent road information is the core of the intelligent highway tunnel, intelligent facility is the precondition of information collection, which produces information that can be processed, interactive and applied. In order to realize the coordinated management of the intelligent highway tunnel, it involves the cooperative control information system of whole sample, whole space-time and whole process ("three wholes"). The whole sample includes: tunnel structural information, vehicle information, driver information, environmental information, facility information, load information, traffic information and management information. The whole space-time refers to the category of time and space spent on the sample collection, namely the trip of vehicle from the starting point to the end point and the time required. The whole process refers to the change of traffic state and traffic information in the category of the whole space and time (including normal operation and risk operation). The details are as follows.

4.1 Tunnel structural information

The tunnel structural information included the load information of the structure and the geometry information of the structure.

4.1.1 Load analysis information of structure

Load analysis information is mainly used for the health diagnosis of tunnel structures, including stress, strain and characteristics of other structural changes, such as fracture location, crack length, crack depth, and leakage etc.

4.1.2 Geometry information of structure

Information of structural geometry can be used to help building tunnel road topology and calculating traffic capacity of tunnel for cooperative management, including tunnel geometry information (information of plane section, longitudinal section and cross section, cross channel for pedestrian and vehicles as well as underground fan room, etc.) and geometry information of roads connected with tunnel.

4.2 Vehicle information

Vehicle information includes the basic information of the vehicle (vehicle type, engine number, license number, the phone number of the owner), the motion information of the vehicle (position and velocity) and the state information of the vehicle (tire pressure, brake conditions, engine conditions, etc.). Vehicle information can be used for traffic controlling, prevention of traffic accident and rescue.

4.3 Driver information

Driver information includes basic information of drivers, such as driving license information, reaction time and physical information (blood pressure, mental state, sudden illness records, contact information of relatives and friends). Driver information can be used for accident prevention and rescue.

4.4 Environmental information

Environmental information includes the weather information, surface friction coefficient, temperature, humidity, concentration of harmful gas and luminance inside/outside the tunnel. Environmental information can be used in traffic control, ventilation and illumination control, accident prevention and early warning.

4.5 Facilities information

Facilities information includes location and running state of fire control facilities, security facilities, monitoring facilities, communication facilities, ventilation and illumination facilities. Facilities information can be used for traffic control and maintenance management.

4.6 Load information

Load information includes type and quantity of cargo, number of passengers as well as their contact information. Load information can be used for accident assessment and rescue.

4.7 Traffic information

Traffic information includes traffic volume, traffic composition, vehicle spacing, velocity and so on. Traffic information can be used for the evaluation of traffic status, promotion, control and warning, etc.

4.8 Management information

Management information includes methods of information transmission and performance, on-guard condition of managers, maintenance of facilities, emergency plans, protocol hospital and firefighting. Management information can be used to support decision-making.

5 Multi-source information fusion

The mining of tunnel operation data is similar to the principle of multi-source information fusion, which is the corresponding method and model according to the actual demand, and this article focuses on the latter. The purpose of multi-source information fusion is to solve the problem of the state evaluation, development deducing and global optimal decision-making in tunnel coordination management.

The key to information fusion is building an appropriate model. At present, information fusion models include functional model, data model and hybrid model. The functional model is based on node sequence, including most typically the UK Intelligence Ring and the Boyd Control Loop (OODA loop). The data model is based on data extraction, including most typically JDL model, waterfall model and Dasarathy model. The hybrid model combines the features of the functional model with that of the data model. All these models can be generally divided into two categories: random model and artificial intelligent model. The random model includes the methods of kalman filtering, bayesian estimation, etc. Artificial intelligent model includes fuzzy logic theory, neural network, rough set theory, Agent, etc.

Hybrid model is suitable for tunnel information fusion. Based on information sources and requirements, the model selection is as follows.

5.1 State evaluation

UK Intelligence Ring model can be used in the status evaluation including the assessment of traffic states, security states and management states.

5.2 Development deducing

JDL model can be used in development deducing to predict the change of traffic states or the change of the traffic accidents such as security early warning, congestion in a certain period of time and space in order to find the optimal decision.

5.3 Optimal decision-making

Optimal decision-making is to find methods to make tunnel collaborative management more safely, efficiently and economically, while making risk, danger and disaster in minimum. Dasarathy model covers all stages of information fusion, including data level fusion, feature selection, feature extraction, feature level fusion, pattern recognition, pattern processing and decision level fusion. In the method, artificial immune system[15]and fuzzy theory[8]have all been proved efficiently.

6 Conclusions

In this paper, five aspects have been studied: (1) Intelligence should have five abilities. (2) The premise of the intelligent tunnel is intelligent traffic organization, intelligent facilities and intelligent vehicles. (3) The topological structures of normal traffic and abnormal traffic are presented. (4) A intelligent tunnel information system with 8 kinds of information is constructed. (5) The management of intelligent tunnel should realize collaborative management of full samples, whole time and whole process.

The intelligent tunnel information system constructed in this paper provides an idea for the construction of the tunnel intelligent platform, which has theoretical significance and application value. However, to realize information interaction between different platforms, the information island problem is still one of the key issues to be solved.

[1]WANG Hong. Intelligent high-speed highway modernization: 2016 National Intelligent Highway Development Forum held in Hangzhou[J]. China ITS Journal, 2017(2): 16.

[2]WANG Jianqiang, LI Shiwei, ZENG Junwei. Research on the development mode of car networking[J]. Computer Technology and Development, 2011, 21(12): 235.

[3]ZENG Lei, WANG Shaofei, HE Xuchun, et al. Concept, architecture and key technology of intelligent highway tunnel[J]. Modern Tunnelling Technology, 2016, 53(4): 1.

[4]SU Jing, WANG Dong, ZHANG Feifei. Application review of vehicle networking technology [J]. Internet of Things Technologies, 2014(6): 69.

[5]CHENG Chonguo, HAN Zhi. Information construction of highway tunnel[C]// National Highway Tunnel Academic Conference. Beijing: China Railway & Transportation Society, 2001: 151.

[6]ZHU Mingyan, GU Yongxing. Tunnel intelligent system[C]// International Symposium on Intelligent and Green Architecture. Beijing: China Architecture & Building Press, 2005: 577.

[7]HAN Zhi. A method of lighting control for highway tunnel: CN 101346025 A[P]. 2008-12-18[2018-03-21].

[8]HAN Zhi. A method of fuzzy control for highway tunnel lighting: CN 101437343 A[P]. 2008-12-18[2018-03-21].

[9]HAN Zhi. Spatial and temporal effects of highway traffic[J]. Journal of Chang′an University(Natural Science Edition), 1988(3): 95.

[10]China Merchants Chongqing Communications Technology Research & Design Institute Co., Ltd. Specifications for design of highway tunnels: JTG D70/2-2014[S]. Beijing: China Communications Press, 2014: 71.

[11]HAN Zhi, FANG Jianqin, HONG Weipeng. Highway tunnel energy-saving technology[M]. Beijing: China Communications Press, 2010: 65.

[12]LUO Yun. Safety economics[M]. Beijing: Chemical Industry Press, 2010: 89.

[13]Ministry of Transportation of the People′s Republic of China. Highway monitoring technology requirements[M]. Beijing: China Communications Press, 2012: 103.

[14]Code for design of automatic fire alarm systems: GB 50116-2013[S]. Beijing: China Planning Press, 2014: 22.

[15]YI Fujun, TIAN Deng, HAN Zhi. Variable speed control method based on artificial immune particle swarm algorithm: CN 101840635 A[P]. 2010-05-06[2018-03-21].