Highway ramp confluence area safety warning and cooperative management and control method based on intelligent control algorithm

Abstract

The invention discloses a safety warning and cooperative management and control method for a road ramp confluence area based on an intelligent control algorithm. The method comprises the following steps: carrying out data acquisition on bidirectional traffic flows on a main line and ramps through a thundersight all-in-one machine; a second-order type-2 fuzzy neural network intelligent control algorithm is operated through an edge intelligent controller; inputting different traffic flow data on the main line and the ramp collected by the all-in-one machine into the edge intelligent controller, analyzing the data through an intelligent control algorithm, and generating a control strategy in which the main line speed control and the ramp guide mode cooperate with each other; the edge intelligent controller directly controls the information guide screen according to a control strategy obtained by an intelligent control algorithm, and the roadside guide sign respectively issues active control and safety guide information to main line and ramp vehicles. The traffic flow conditions of the main line and the ramp are considered at the same time, and the vehicles on the main line and the ramp are guided in a targeted mode.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a method for safety warning and coordinated management of a highway ramp merging area based on an intelligent control algorithm. Background Art

[0002] The merging area of ​​a highway includes ramps, acceleration lanes, and the main line connected to them. As the ramps and main line vehicles merge in this area, the traffic volume suddenly increases in time and space, which can easily lead to problems such as reduced traffic flow speed and even chaos. It is a common accident-prone area on highways.

[0003] When vehicles merge at the ramp entrance, it is necessary to observe the driving status of vehicles on the main road and make corresponding decisions. When the number of vehicles waiting to merge is large, congestion is likely to occur at the ramp entrance, making it difficult to provide a fast, efficient and safe travel experience on the highway. With the continuous increase in the number of cars per capita, the traffic volume on highways has increased sharply, making the probability of traffic congestion and traffic accidents at the ramp merge increasingly high.

[0004] Existing ramp control schemes are mainly divided into two categories. One is an intelligent scheme based on vehicle-road collaboration. Through vehicle-road data interaction, the control parameters that need to be executed by each vehicle to avoid accidents are derived and sent to the vehicles for execution. This type of scheme can only be implemented under the condition of vehicle-road collaboration, and is currently difficult to apply in practice; the other type simply collects the traffic flow characteristics of the main line, analyzes possible accidents, and issues simple avoidance warnings to vehicles, but the implementation effect is limited. Therefore, seeking a front-end intelligent control algorithm that takes into account the traffic conditions of the main line and ramp at the same time and provides targeted guidance to the main line and ramp vehicles respectively is of positive significance for alleviating traffic problems at ramp confluences. Summary of the invention

[0005] The purpose of the present invention is to provide a method for safety warning and coordinated management of highway ramp merging areas based on intelligent control algorithms, so as to solve the problems of low level of highway safety management and high frequency of traffic accidents in ramp merging areas caused by conflicts between main line and ramp traffic.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for safety warning and coordinated control of highway ramp merging areas based on an intelligent control algorithm comprises the following steps:

[0008] Data collection of two-way traffic flow on the main line and ramps is carried out through the integrated radar and vision device;

[0009] Run the second-order type-2 fuzzy neural network intelligent control algorithm through the edge intelligent controller;

[0010] The different traffic flow data on the main line and ramp collected by the integrated radar vision device are input into the edge intelligent controller, and the control strategy of the main line speed control and ramp guidance mode coordinating with each other is generated through intelligent control algorithm analysis;

[0011] The control strategy derived from the intelligent control algorithm is directly used by the edge intelligent controller to control the information guidance screen and roadside guidance signs to issue active control and safety guidance information to mainline and ramp vehicles respectively.

[0012] Furthermore, the edge intelligent controller directly interacts with the traffic flow data collection equipment and the warning guidance information release equipment through the front-end networking mode, and realizes data processing, analysis, decision-making and security control at the front end, thereby improving the real-time performance of the system operation and reducing the potential network delay risk of centralized decision-making.

[0013] Furthermore, the second-order type-2 fuzzy neural network intelligent control algorithm reduces the dimension and amount of calculation of fuzzy rules by introducing new intermediate variables to adapt to the fuzzy decision-making of multiple input variables of the main line and ramps by edge computing.

[0014] Furthermore, the main line and ramp multiple input variables include main line average speed, main line vehicle spacing, ramp average vehicle speed, and ramp fleet length.

[0015] Furthermore, the second-order type-2 fuzzy neural network intelligent control algorithm includes a 7-layer network architecture, and introduces the main line priority and ramp priority intermediate layers. The fuzzy input variables are weighted through the priority variables to achieve further fuzzy feature extraction of the input variables. By adding the intermediate fuzzy analysis layer, the control problem is decomposed and integrated in stages, reducing the complexity of multi-input variable fuzzy rule design and network calculation.

[0016] Furthermore, the 7-layer network architecture includes the following 7 layers:

[0017] The first layer: input layer. The input is divided into two groups, namely the average speed of the main line, the distance between main line vehicles and the average speed of ramps, and the length of ramp fleets. The first layer nodes directly pass the input variables without transformation;

[0018] The second layer: the membership function layer. The nodes in this layer correspond to the type II fuzzy membership functions of the input variables. The type II Gaussian membership function with a certain mean and uncertain mean square error is selected. The input variables of the second layer are the output variables of the first layer.

[0019] The third layer: the first-order fuzzy rule layer, which formulates fuzzy rules for mainline variables and ramp variables respectively;

[0020] The fourth layer: the first-order fuzzy reduction conclusion layer, generates intermediate variables for the mainline variables and ramp variables, namely the mainline priority and ramp priority, which are divided into three fuzzy subsets {L, M, H} representing {low, medium, high} respectively;

[0021] The fifth layer: the second-order fuzzy rule layer, which formulates the system fuzzy control rules based on the intermediate variables;

[0022] The sixth layer: the second-order fuzzy reduction conclusion layer, which reduces the second-order fuzzy set to a first-order fuzzy set and generates the output interval of the control variable;

[0023] The seventh layer: control system output layer, generates actual control parameters.

[0024] Furthermore, the second-order type-2 fuzzy neural network intelligent control algorithm is an integrated network, which realizes the coordinated optimization of the main line and ramp control output variables according to the main line and ramp traffic flow parameters.

[0025] Beneficial effects:

[0026] The present invention takes into account the traffic conditions of the main line and ramp at the same time, and provides targeted guidance for the main line and ramp vehicles respectively. The designed type-2 second-order fuzzy neural network controller processes input variables through type-2 fuzzy functions, which can better handle the complex fluctuations of the main line and ramp traffic flows and improve the robustness of the system. In addition, the type-2 second-order fuzzy neural network controller introduces main line priority and ramp priority, increases the level of type-2 fuzzy neural network, and reduces the complexity of multi-input variable fuzzy rule design and network calculation. The controller achieves optimized coordination of main line speed control and ramp safety guidance through fuzzy rule decision-making, and can achieve better ramp safety control effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0028] Figure 1 It is a flow chart of a method for safety warning and coordinated control of highway ramp merging area based on intelligent control algorithm according to an embodiment of the present application;

[0029] Figure 2 is another flow chart of a method for safety warning and coordinated control of a highway ramp merging area based on an intelligent control algorithm according to an embodiment of the present application;

[0030] Figure 3 It is a system architecture diagram of a method for safety warning and coordinated control of highway ramp merging area based on intelligent control algorithm according to one embodiment of the present application;

[0031] Figure 4It is a schematic diagram of front-end equipment deployment of a method for safety warning and coordinated control of a highway ramp merging area based on an intelligent control algorithm according to an embodiment of the present application;

[0032] Figure 5 It is an algorithm architecture diagram of a method for safety warning and coordinated management of highway ramp merging areas based on an intelligent control algorithm provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0034] Example

[0035] The embodiment of the present application discloses a method for safety warning and coordinated management of highway ramp merging areas based on an intelligent control algorithm. The method is applied to the field of road traffic safety management. By collecting main line and ramp traffic flow data before the highway merging area, the potential traffic conflicts at the intersection are predicted and analyzed through an intelligent algorithm, and a main line and ramp safety guidance strategy group is generated. Warning and guidance are implemented by directly controlling the roadside equipment through the front end, thereby improving the real-time performance of safety management and reducing the probability of traffic accidents and congestion in the ramp merging area.

[0036] Specifically, the present application provides a method for safety warning and coordinated control of highway ramp merging areas based on an intelligent control algorithm, such as Figure 1 and Figure 2 As shown, the following steps are included:

[0037] S1, collect data on the two-way traffic flow on the main line and ramps through the integrated radar and vision machine. Specifically, by deploying integrated radar and vision machines in the main line direction and ramp direction, the main line traffic flow and the traffic flow that will merge into the main line are collected in a refined manner. In this embodiment, the average speed of the two lanes outside the main line and the distance between the vehicles in the lanes outside the main line, as well as the average speed of the ramp and the length of the fleet are obtained to intelligently analyze the potential risks of the merging area, and generate the main line speed control and ramp safety guidance mode parameters.

[0038] S2, running the second-order type-2 fuzzy neural network intelligent control algorithm through the edge intelligent controller. The intelligent control algorithm uses type-2 fuzzy functions to process input variables to improve the robustness of the algorithm and cope with the complex fluctuations in mainline and ramp traffic flow. Among them, the second-order type-2 fuzzy neural network intelligent control algorithm reduces the fuzzy rule dimension and calculation amount by introducing new intermediate variables to adapt to the fuzzy decision-making of edge computing processing of multiple input variables of the mainline and ramp.

[0039] Specifically, in order to achieve intelligent decision-making and autonomous optimization of the control method, this application constructs a new type-II fuzzy neural network as the system architecture of the control method. The current fuzzy system and fuzzy neural network are limited in practical applications by the difficulty of accurately defining fuzzy rules and the complexity of calculation. They are generally limited to having a good effect in processing scenarios with two input variables, and this situation is more prominent in type-II fuzzy systems and type-II neural networks.

[0040] In addition, in order to comprehensively consider the impact of mainline and ramp traffic flow and realize refined control, this control system selects different variables of mainline and ramp confluence to realize ramp warning analysis and control decision. In response to the requirements of multiple input variables, this application proposes and designs a type II second-order fuzzy neural network controller, introduces mainline priority and ramp priority, and further extracts fuzzy features of input variables through weighted processing of priority variables. By adding this intermediate fuzzy analysis layer, the control problem is decomposed and integrated in stages, reducing the complexity of multi-input variable fuzzy rule design and network calculation, meeting the needs of real-time control and online optimization adjustment of the roadside terminal of the control system.

[0041] S3, inputting different traffic flow data on the main line and ramps collected by the integrated radar and vision device into the edge intelligent controller, and generating a control strategy for the coordination of main line speed control and ramp guidance mode through intelligent control algorithm analysis.

[0042] Specifically, the intelligent control algorithm selects the mainline average speed, the mainline vehicle spacing, the ramp average vehicle speed, and the ramp fleet length as the control system input parameters to comprehensively consider the operation of the mainline and ramp traffic flows, while taking into account the different characteristics of the mainline and ramp traffic flows.

[0043] S4, the control strategy derived from the intelligent control algorithm is directly used by the edge intelligent controller to control the information guidance screen and the roadside guidance signs to release active control and safety guidance information to the main line and ramp vehicles respectively. Specifically, after the edge intelligent controller formulates an optimized safety guidance strategy, it directly controls the main line information guidance screen connected to the front edge side to release the main line speed control information, and controls the ramp edge guidance signs to guide with different light colors and frequencies according to the control mode, so as to achieve targeted safety collaborative guidance for main line vehicles and ramp vehicles.

[0044] In this embodiment, the edge intelligent controller realizes data interaction by directly connecting with roadside sensors and information guidance equipment through a local area network at the front end. The edge intelligent controller directly obtains the main line and ramp traffic flow data, and uses the proposed intelligent algorithm to generate a main line and ramp safety guidance strategy combination under the potential traffic conflict prediction mechanism, and directly controls information release equipment such as information guidance screens and roadside guidance signs to issue warning guidance information to vehicles.

[0045] In actual scenarios, information guidance screens, edge guidance signs, horns and other release equipment can be reasonably selected on the main line and ramps according to the on-site traffic operation environment. Multiple devices can be combined to achieve information composite release capabilities and achieve the best warning and guidance effects.

[0046] Preferably, Figure 3 As shown, the edge intelligent controller directly interacts with the traffic flow data collection equipment and the warning guidance information release equipment through the front-end networking mode, and realizes data processing, analysis, decision-making and security control at the front end.

[0047] Preferably, the second-order type-2 fuzzy neural network intelligent control algorithm includes a 7-layer network architecture, and introduces the main line priority and ramp priority intermediate layers. The fuzzy input variables are weighted through the priority variables to achieve further fuzzy feature extraction of the input variables. By adding an intermediate fuzzy analysis layer, the control problem is decomposed and integrated in stages, reducing the complexity of multi-input variable fuzzy rule design and network calculation.

[0048] like Figure 4 As shown, the specific functions of the 7-layer network architecture are:

[0049] The first layer: input layer. The input is divided into two groups, namely the average speed of the main line, the main line vehicle spacing and the average speed of the ramp, and the length of the ramp fleet. The first layer nodes directly pass the input variables without transformation.

[0050]

[0051] Where v1 represents the average speed of the main line, hw1 represents the distance between vehicles on the main line, v2 represents the average speed of the ramp, and ql1 represents the length of the ramp fleet. is the first layer mainline type II fuzzy input and output, It is the first layer of ramp type II fuzzy input and output, and the connection weights of this layer are all 1.

[0052] The second layer: the membership function layer. The nodes of this layer correspond to the type II fuzzy membership function of the input variables. The type II Gaussian membership function with a determined mean and an uncertain mean square error is selected. The input variables of the second layer are the output variables of the first layer. In this implementation example, the average speed of the main line is divided into 5 fuzzy subsets {VS, S, M, F, VF} according to the speed interval [0, 120], which respectively represent "very slow, slow, normal, fast, and very fast"; the main line vehicle spacing is divided into 3 fuzzy subsets {S, M, L} according to the distance interval [0, 10], which respectively represent {short, moderate, and long}; the average speed of the ramp is divided into 3 fuzzy subsets {S, M, F} according to the speed interval [0, 60], which respectively represent "slow, normal, and fast"; the length of the ramp fleet is divided into 3 fuzzy subsets {S, M, L} according to the vehicle number interval [0, 7], which respectively represent {small, moderate, and many}.

[0053]

[0054] in is the input and output of the second layer main line membership function, Output variable The upper and lower limits, m mij , is the center and mean of the main line membership function, is the input and output of the second-layer ramp membership function, Output variable The upper and lower limits of m rij , is the center and mean of the ramp membership function. This layer defines a set of membership functions for each output variable of the first layer. The input of each set of membership functions is the output of the first layer, and the connection weights of this layer are all 1.

[0055] The third layer: the first-order fuzzy rule layer. This control system formulates fuzzy rules for the main line variables and ramp variables respectively. Each fuzzy neuron in this layer performs an "AND" operation and uses the "product" operator to calculate the premise activation area of ​​the fuzzy rule. The product of the upper bound output variables of the second layer is the upper bound input variable of the third layer, and the product of the lower bound output variables of the second layer is the lower bound input variable of the third layer. The fuzzy neuron transformation is expressed as:

[0056]

[0057] in It is the input and output of the third-layer mainline fuzzy rule reasoning. For variables The upper and lower limits of For variables The upper and lower limits of is the input and output of the third-level ramp fuzzy rule reasoning, For variables The upper and lower limits of For variables The upper and lower limits of the connection weights of this layer are all 1.

[0058] The fourth layer: the first-order fuzzy reduction conclusion layer. This control system generates intermediate variables for the mainline variables and ramp variables, namely the mainline priority and ramp priority, which are divided into three fuzzy subsets {L, M, H}, which represent {low, medium, high} respectively. Each rule fuzzy neuron in the third layer has a corresponding conclusion fuzzy neuron. The fuzzy neuron transformation is expressed as:

[0059]

[0060]

[0061] in Input and output of the fuzzy conclusion of the fourth main line, For variables The upper and lower limits of For variables The upper and lower limits of is the lower limit ratio of the main line variable, The upper limit ratio of the main line variable, The fourth-level ramp fuzzy conclusion input and output, For variables The upper and lower limits of For variables The upper and lower limits of is the lower limit ratio of ramp variable neurons, The upper limit ratio of ramp variable neurons, the network connection weights of the main line and ramp conclusion of this layer are and

[0062] The fifth layer is the second-order fuzzy rule layer. This control system formulates the system fuzzy control rules based on the intermediate variables. Each fuzzy neuron in this layer performs the "and" operation and uses the "product" operator to calculate the premise activation area of ​​the fuzzy rule. The upper bound input variable of the fifth layer is the product of the sum of the weighted value of the upper bound output variable of the second layer and the upper bound output variable of the fourth layer. The lower bound input variable of the fifth layer is the product of the sum of the weighted value of the lower bound output variable of the second layer and the lower bound output variable of the fourth layer. The upper bound input variable of the fifth layer is the cross product of the upper bound input variable of the fifth layer, and the cross product of the lower bound input variable of the fifth layer. The fuzzy neuron transformation is expressed as:

[0063]

[0064]

[0065] in The fifth layer fuzzy rule reasoning input and output, For variables The upper and lower limits of For variables The upper and lower limits of the connection weights of this layer are all 1.

[0066] The sixth layer: the second-order fuzzy reduction conclusion layer, which reduces the second-order fuzzy set to a first-order fuzzy set and generates the output interval of the control variable. Each rule fuzzy neuron in the fifth layer has a corresponding conclusion fuzzy neuron, and the fuzzy neuron transformation is expressed as:

[0067]

[0068] in The fifth layer fuzzy rule reasoning input and output, For variables The upper and lower limits of For variables The upper and lower limits of is the lower limit ratio of neurons, is the upper limit ratio of neurons, and the connection weight of this layer is

[0069] The seventh layer: the control system output layer generates actual control parameters. This control system defines two control variables u1 and u2, which control the main line speed and ramp warning light mode respectively.

[0070]

[0071] Among them, u1∈[0,120] is the main line safe driving control speed, u2={1,2,3,4,5,6} represents one of the six modes, u z , The upper and lower limits of the output variable.

[0072] In this implementation example, 1000 sets of input parameters are collected, and the corresponding output control parameters are configured according to expert experience. The proposed second-order type-II fuzzy controller is trained to generate an intelligent controller with optimized parameters. The input is the input vector composed of the collected mainline average speed, mainline vehicle spacing, ramp average speed, and ramp fleet length, x i =(v1 i ,hw1 i ,v2 i ,ql2 i ), i = {1, 2, ..., 1000}, the training error is located as:

[0073]

[0074] Where E is the total error, E1 is the error between the actual output and the expected output of the main line control variable, is the design expected output of the main line control variable and the corresponding output of the system, E2 is the error between the actual output of the ramp control variable and the expected output, For the design expected output of ramp control variables and the corresponding system output, the following variables are defined for parameter learning and correction:

[0075]

[0076]

[0077] The optimization and correction calculation of the second-order fuzzy reduced connection weights of the sixth layer is as follows:

[0078]

[0079] in, The calculation is as follows:

[0080] calculate

[0081] 1. Through the transformation matrix Q z right w z Sort in descending order

[0082] 2. Through the transformation matrix Q z right I (5) , Just transform it, record it as

[0083] 3. Calculate the transformation point L

[0084] 1) Initialization, definition calculate make

[0085] 2) Search satisfy

[0086] 3) Calculation And define

[0087] 4) If Stop the search and return otherwise make And return to step 2) to continue searching.

[0088] 4. Definition but δ k (k=1,2,…,9) can be calculated by the following formula:

[0089] ( δ 1, δ 2,…, δ 9) T = Q z r z

[0090] calculate

[0091] 1. Through the transformation matrix right Sort in descending order

[0092] 2. Through the transformation matrix to I (5) , Just transform it, record it as

[0093] 3. Calculate the transformation point Y

[0094] 1) Initialization, definition calculate make

[0095] 2) Search satisfy

[0096] 3) Calculation And define

[0097] 4) If Stop the search and return otherwise make And return to step 2) to continue searching.

[0098] 4. Definition but It can be calculated by the following formula:

[0099]

[0100] z={1,2}

[0101] The optimization and correction calculation of the first-order fuzzy reduced-type connection weights of the main line of the fourth layer is as follows:

[0102]

[0103] in The calculation is as follows:

[0104]

[0105] in The calculation is as follows:

[0106] calculate

[0107] 1. Through the transformation matrix Q h right w h Sort in descending order

[0108] 2. Through the transformation matrix Q h right Just transform it, record it as

[0109] 3. Calculate the transformation point L r

[0110] 1) Initialization, definition calculate make

[0111] 2) Search satisfy

[0112] 3) Calculation And define

[0113] 4) If Stop the search and return otherwise make And return to step 2) to continue searching.

[0114] 4. Definition but It can be calculated by the following formula:

[0115]

[0116] calculate

[0117] 1. Through the transformation matrix right Sort in descending order

[0118] 2. Through the transformation matrix right Just transform it, record it as

[0119] 3. Calculate the transformation point Y r

[0120] 1) Initialization, definition calculate make

[0121] 2) Search satisfy

[0122] 3) Calculation And define

[0123] 4) If Stop the search and return otherwise make And return to step 2) to continue searching.

[0124] 4. Definition but It can be calculated by the following formula:

[0125]

[0126] Preferably, the second-order type-2 fuzzy neural network intelligent control algorithm is an integrated network, which realizes the coordinated optimization of the main line and ramp control output variables according to the main line and ramp traffic flow parameters. Figure 5 As shown, the specific intelligent control system control and online optimization strategy in this implementation example is as follows:

[0127] After the second-order type-2 fuzzy neural network is trained and optimized, the edge intelligent controller directly connects to the integrated radar and vision devices deployed on the main line and ramps to collect the main line and ramp traffic flow parameters, including the average speed of the main line, the main line vehicle spacing, the average ramp speed, and the length of the ramp fleet, and brings them into the second-order type-2 fuzzy neural network to generate a safety guidance strategy group for the main line and ramp. At the same time, the edge intelligent controller stores the collected data and periodically trains the second-order type-2 fuzzy neural network to optimize the network parameters.

[0128] The highway merging area safety warning guidance method and the proposed second-order type-2 fuzzy neural network controller provided in the embodiments of the present application can be applied to the highway ramp merging area safety guidance method provided in the above embodiments. For relevant details, refer to the above method embodiments. The implementation principles and technical effects are similar and will not be repeated here.

[0129] It should be noted that the merging area safety warning guidance method and the proposed second-order type-2 fuzzy neural network controller provided in the embodiments of the present application only use the division of the above-mentioned functional modules / functional units as examples when performing vehicle safety following distance detection and warning. In actual applications, the above-mentioned functions can be assigned to different functional modules / functional units as needed, that is, the internal structure of the driving safety warning device is divided into different functional modules / functional units to complete all or part of the functions described above.

[0130] Although the embodiments of the present invention are described in the specification, these embodiments are only for reference and should not limit the protection scope of the present invention. Various omissions, substitutions and changes within the scope of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for safety warning and coordinated control of highway ramp merging areas based on intelligent control algorithms, characterized in that: The following steps are involved: Data collection of two-way traffic flow on the main line and ramps is carried out through the integrated radar and vision device; Run the second-order type-2 fuzzy neural network intelligent control algorithm through the edge intelligent controller; The different traffic flow data on the main line and ramp collected by the integrated radar vision device are input into the edge intelligent controller, and the control strategy of the main line speed control and ramp guidance mode coordinating with each other is generated through intelligent control algorithm analysis; The control strategy derived from the intelligent control algorithm is directly used by the edge intelligent controller to control the information guidance screen and roadside guidance signs to issue active control and safety guidance information to mainline and ramp vehicles respectively.

2. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 1 is characterized in that: The edge intelligent controller directly interacts with the traffic flow data collection equipment and the warning guidance information release equipment through the front-end networking mode, and realizes data processing, analysis, decision-making and safety control at the front end.

3. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 1 is characterized in that: The second-order type-2 fuzzy neural network intelligent control algorithm reduces the fuzzy rule dimension and calculation amount by introducing new intermediate variables to adapt to the fuzzy decision-making of multiple input variables of the main line and ramps processed by edge computing.

4. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 3 is characterized in that: The main line and ramp multiple input variables include main line average speed, main line vehicle spacing, ramp average vehicle speed, and ramp fleet length.

5. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 1 is characterized in that: The second-order type-2 fuzzy neural network intelligent control algorithm includes a 7-layer network architecture, and introduces the main line priority and ramp priority intermediate layers. The fuzzy input variables are weighted through the priority variables to achieve further fuzzy feature extraction of the input variables. By adding the intermediate fuzzy analysis layer, the control problem is decomposed and integrated in stages, reducing the complexity of multi-input variable fuzzy rule design and network calculation.

6. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 5 is characterized in that: The 7-layer network architecture includes the following 7 layers: The first layer: input layer. The input is divided into two groups, namely the average speed of the main line, the distance between main line vehicles and the average speed of ramps, and the length of ramp fleets. The first layer nodes directly pass the input variables without transformation; The second layer: the membership function layer. The nodes in this layer correspond to the type II fuzzy membership functions of the input variables. The type II Gaussian membership function with a certain mean and uncertain mean square error is selected. The input variables of the second layer are the output variables of the first layer. The third layer: the first-order fuzzy rule layer, which formulates fuzzy rules for mainline variables and ramp variables respectively; The fourth layer: the first-order fuzzy reduction conclusion layer, generates intermediate variables for the mainline variables and ramp variables, namely the mainline priority and ramp priority, which are divided into three fuzzy subsets {L, M, H} representing {low, medium, high} respectively; The fifth layer: the second-order fuzzy rule layer, which formulates the system fuzzy control rules based on the intermediate variables; The sixth layer: the second-order fuzzy reduction conclusion layer, which reduces the second-order fuzzy set to a first-order fuzzy set and generates the output interval of the control variable; The seventh layer: control system output layer, generates actual control parameters.

7. The method for safety warning and coordinated management of highway ramp merging area based on intelligent control algorithm according to claim 1 is characterized in that: The second-order type-2 fuzzy neural network intelligent control algorithm is an integrated network that realizes the coordinated optimization of the mainline and ramp control output variables according to the mainline and ramp traffic flow parameters.

Images