Real-time prioritization method for Internet of Vehicles messages based on machine learning and decision theory

Abstract

This invention discloses a real-time prioritization method for IoV messages based on machine learning and decision theory. This method, applied in a network environment consisting of several roadside units (ROUs) and vehicle-mounted units (VMUs), includes the following steps: 1. Vehicle message generation and broadcast; 2. RMU generation and forwarding of utility functions; and 3. Vehicles receive and prioritize the utility functions and attribute weights. This method utilizes machine learning to improve the accuracy of the utility functions and weights generated by RMUs applying decision theory to IoV scenarios. This method helps IoV vehicles prioritize cached messages based on message content, ensuring real-time determinism in the resulting ranking.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle networking applications, and specifically relates to a real-time priority sorting method for vehicle networking messages based on machine learning and decision theory. Background Art

[0002] The introduction of intelligent connected vehicles and autonomous vehicles into road networks has placed multiple demands on the Internet of Vehicles (IoV) to ensure the safety and efficiency of intelligent transportation systems. IoV vehicles need to process messages received from neighboring vehicles and roadside units (ROUs) based on different applications. For example, digital signatures of messages are verified to ensure the authenticity of the message sender, but this incurs significant computational overhead. Due to hardware limitations, the onboard units (OBUs) on vehicles cannot process such a large number of received messages within a given timeframe, especially in high-density road networks. This means that only specific information can be computed in a timely manner, making responding to emergency messages in IoV networks a challenging task. Current approaches primarily prioritize messages based on a single attribute or manual settings, neglecting the use of an effective scheme to determine how to prioritize messages based on multiple attributes in a dynamic vehicle environment.

[0003] The existing methods for processing messages in the Internet of Vehicles have the following problems:

[0004] 1. Failure to ensure that high-priority information is evaluated and processed in a timely manner;

[0005] 2. It cannot ensure that the priority of information changes dynamically as the Internet of Vehicles environment changes;

[0006] 3. The priority of a message cannot be quantified based on the content of the message. Summary of the Invention

[0007] In order to address the shortcomings of the above-mentioned existing methods, the present invention proposes a real-time priority sorting method for vehicle network messages based on machine learning and decision theory, in order to utilize machine learning algorithms to optimize the process of vehicle message priority sorting, so as to improve the efficiency and accuracy of vehicle message sorting, thereby solving the problem that traditional decision-making methods cannot perform high-reliability real-time sorting of vehicle messages when the vehicle environment changes dynamically.

[0008] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0009] The present invention is a method for real-time prioritization of vehicle network messages based on machine learning and decision theory. The method is applied to a local area network consisting of a trusted third party, a roadside unit, and a vehicle's onboard unit, and is performed according to the following steps:

[0010] Step 1: Vehicle message generation and broadcasting;

[0011] Step 1.1: Vehicle i traveling in the local area network senses and collects information about itself and the environment in real time as various attributes, including: the vehicle's current speed v i , the instantaneous acceleration of the vehicle at present i 、The vehicle's current heading angle h i , the Euclidean distance d between the vehicle's current position and the roadside unit's position i , thus forming the attribute vector of vehicle i in, represents the j-th attribute of vehicle i;

[0012] Step 1.2: Vehicle i will r i Place in own message m i The message is then cached in the message queue to wait for forwarding;

[0013] Step 1.3: Vehicle i broadcasts its own message m to other vehicles in the local area network and the roadside unit i , other vehicles and roadside units receive the message m from vehicle i i Then, add it to its own message library in chronological order;

[0014] Step 2: The roadside unit generates and forwards the utility function and attribute weights;

[0015] Step 2.1: Let the current k-th cycle be T k ; Let the current k-th cycle T k The starting time is s k ; Let the current k-th cycle T k End time e k ; Let the kth period T k The corresponding time period t k for e k -s k ;

[0016] The roadside unit selects the message set M received in the kth period Tk from its own message database k ;

[0017] Step 2.2: Let the message set M k The jth attribute of all messages in is denoted as P k，j ;

[0018] The roadside unit statistics j-th attribute P k，j The minimum value min(P k，j ) and the maximum value max(P k，j), and thus use formula (1) to get the j-th attribute P k，j The value range of ΔP k，j ;

[0019] ΔP k，j =max(P k，j )-min(P k，j ) (1)

[0020] Step 2.3: Define the current number of iterations as z, initialize z = 1, and define the number of data intervals for the zth iteration as μ z , and initialize Among them, ceil() means rounding up;

[0021] Use formula (2) to construct μ z Constraints:

[0022]

[0023] In formula (2), α and β represent two predefined parameters; and α≤ΔP k，j , β≥1;

[0024] Step 2.4: The roadside unit will ΔP k，j Divide into μ z intervals, thereby obtaining the attribute value interval set of the zth iteration in, Indicates the yth partition of the zth iteration z attribute value ranges, and has:

[0025]

[0026] Step 2.5: The roadside unit takes the jth attribute P k，j In M k All values ​​of μ are divided into z In the interval, the statistical result vector of the zth iteration is obtained and in, Indicates the yth partition of the zth iteration z The number of j-th attributes contained in the interval;

[0027] Step 2.6: Roadside units use machine learning algorithms to After fitting, the fitting curve of the zth iteration is obtained And calculate the fitting loss function value of the zth iteration

[0028] Step 2.7: Set μ z +1 is assigned to μ z After that, judge Is it true? If so, execute step 2.8; otherwise, get the fitting loss function value of the previous z iterations and execute step 2.9; where floor() means rounding down;

[0029] Step 2.8: After assigning z+1 to z, return to step 2.4 and execute sequentially;

[0030] Step 2.9: Obtain the number of data intervals corresponding to the minimum fitting loss function value from the fitting loss function value of the previous z iterations as the optimal data interval number μ * ;

[0031] Step 2.10: The roadside unit uses machine learning algorithms to calculate μ * The corresponding statistical result vector Perform fitting to obtain the best fitting curve

[0032] Step 2.11: Use formula (3) to get the j-th attribute P in the k-th period Tk k，j The utility function

[0033]

[0034] In formula (3), Norm(·) represents the normalization function;

[0035] Step 2.12: Use equation (4) to get the kth period T k The jth attribute P k，j Weight

[0036]

[0037] In formula (4), express the number of nonzero elements in ;

[0038] Step 3: The vehicle receives the utility function and attribute weights and prioritizes them;

[0039] Step 3.1: When vehicle i receives the utility function sent by the roadside unit in the current kth period Tk and weights When the message base composed of the messages received by vehicle i in the kth cycle is recorded as make The xth message in make express The number of messages in The attribute vector in is make The jth attribute in is

[0040] Step 3.2: Initialize x=1;

[0041] Step 3.3: Vehicle i calculates the xth message using equation (5) The utility value of

[0042]

[0043] In formula (5), Indicates the xth message in the kth period Tk Attribute vector The jth attribute of Utility function of

[0044] Step 3.4: If Then assign x+1 to x and return to step 3.3 to execute sequentially. Otherwise, it means that The utility values ​​of all messages in the , and form a utility value vector

[0045] Step 3.5: U k，i in The utility values ​​of the messages are sorted in descending order to obtain the real-time priority sorting results of the Internet of Vehicles messages.

[0046] The method for real-time prioritization of vehicle network messages based on machine learning and decision theory according to the present invention is also characterized in that the roadside unit in step 2.5 is to convert M into k All values ​​of μ are divided into z In the interval:

[0047] Step 2.5.1: Traverse the roadside unit message set M k The jth attribute P in k，j For each value of , let the nth value be n∈{1,2,3,......,|M k |},|M k | represents the roadside unit message set M k The number of messages;

[0048] Step 2.5.2: Initialize n = 1;

[0049] Step 2.5.3: Initialize y z =1;

[0050] Step 2.5.4: Make express Middle y z The left boundary value of the interval, express Middle y z The right boundary value of the interval is used to determine whether formula (6) is valid. If it is valid, then Divide into , execute step 2.5.5 in sequence, otherwise, z +1 is assigned to y z , repeat step 2.5.4 until equation (6) is established;

[0051]

[0052] Step 2.5.5: If n<|M k |, assign n+1 to n, and return to step 2.5.3 and execute sequentially until n=|M k |Until now.

[0053] The roadside unit calculates the jth attribute P in the kth period Tk after a certain time interval φ. k，j The resulting utility function and attribute weights To update:

[0054] Step 4.1: Preset a threshold δ; define and initialize variable b = 1;

[0055] Step 4.2: Let s k +b*φ as the k+bth period T k+b The start time s k+b , let e k +b*φ as the k+bth period T k+b End time e k+b , let the k+bth period T k+b The corresponding time period is e k+b -s k+b ;

[0056] Step 4.3: The roadside unit is selected in the k+bth period T k+b Message received within M k+b , and according to M k+b Get the k+bth period T k+b The jth attribute P k+b，j The utility function Thus, we can get With the jth attribute in the kth period T k The utility function JS Divergence

[0057]

[0058] In formula (7), KL() represents the KL divergence calculation function;

[0059] Step 4.4: If Then assign b+1 to b and return to step 4.2 to execute sequentially. Otherwise, execute step 4.5.

[0060] Step 4.5: Calculate k+b periods T using equation (4) k+b Attribute weight

[0061] Step 4.6: The roadside unit will calculate the k+bth cycle T k+b The jth attribute P obtained k+b，j The utility function and its weight And broadcast to vehicles in the local area network.

[0062] An electronic device of the present invention includes a memory and a processor, and is characterized in that the memory is used to store a program that supports the processor to execute any of the real-time priority sorting methods for Internet of Vehicles messages, and the processor is configured to execute the program stored in the memory.

[0063] The present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, which is characterized in that when the computer program is run by a processor, it executes any step of the method for real-time priority sorting of vehicle network messages.

[0064] Compared with the existing technology, the beneficial effects of the present invention are embodied in:

[0065] 1. The present invention utilizes the computing and storage resources of roadside units (RSUs) and uses decision theory and machine learning techniques to help vehicles in a local area network generate utility functions and attribute weights, thereby improving the efficiency of message processing while ensuring the authenticity of message processing.

[0066] 2. The message processing method reported in this invention is based on the characteristics of machine learning that can predict and simulate real events. A message processing method based on machine learning curve fitting is designed. The curve prediction model based on machine learning simulates the current traffic environment to ensure the authenticity of message processing.

[0067] 3. The order in which messages are processed depends on the attribute priority of the messages. The present invention filters out some messages of lower importance by the importance of several attributes contained in the messages, thereby improving the quality of the processed messages.

[0068] 4. This paper proposes a new two-step message priority framework: (1) message analysis at the roadside unit (RSU) and (2) message priority sorting at the onboard unit (OBU), thereby optimizing the multi-layered and complex structure of traditional Internet of Vehicles message processing while ensuring the orderliness of message priority decision-making.

[0069] 5. When generating the utility function, the present invention uses machine learning technology to perform function curve fitting to solve the traditional problem in traditional decision theory that the utility function and weight coefficient are arbitrarily given, thereby enhancing the simulation robustness of the utility function and weight coefficient to a certain extent in the Internet of Vehicles scenario.

[0070] 6. The present invention allows RSU to dynamically generate functions and attribute weights according to real-time traffic conditions, ensuring the high timeliness of the method. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Schematic diagram of the computing framework of the present invention;

[0072] Figure 2 Schematic diagram of calculation on the vehicle side of the present invention;

[0073] Figure 3 Generates a time series graph of utility functions and attribute weights for roadside units. DETAILED DESCRIPTION

[0074] In this embodiment, a real-time priority sorting method for Internet of Vehicles messages based on machine learning and decision theory is applied to Figure 1 In the local area network shown, which consists of a trusted third party, a roadside unit, and the vehicle's onboard unit, the following steps are performed:

[0075] Step 1: Vehicle message generation and broadcasting, Figure 1 The overall framework diagram for message transmission between vehicles and RSU is as follows:

[0076] Step 1.1: Vehicle i traveling in the local area network senses and collects information about itself and the environment in real time as various attributes, including: the vehicle's current speed v i , the instantaneous acceleration of the vehicle at present i 、The vehicle's current heading angle h i , the Euclidean distance d between the vehicle's current position and the roadside unit position i , thus forming the attribute vector of vehicle i in, represents the j-th attribute of vehicle i;

[0077] Step 1.2: Vehicle i will r i Place in own message mi The message is then cached in the message queue to wait for forwarding;

[0078] Step 1.3: Vehicle i broadcasts its message m to other vehicles and roadside units in the local area network i , other vehicles and roadside units receive the message m from vehicle i i Then, add it to its own message library in chronological order;

[0079] like Figure 1 As shown in the figure, after the roadside unit receives the message broadcast by the vehicle, it generates a utility function and attribute weights for the collected messages, and then broadcasts the results to the vehicle. The specific steps are:

[0080] Step 2: The roadside unit generates and forwards the utility function and attribute weights;

[0081] Step 2.1: Let the current k-th cycle be T k ; Let the current k-th cycle T k The starting time is s k ; Let the current k-th cycle T k End time e k ; Let the kth period T k The corresponding time period t k for e k -s k ;

[0082] The roadside unit selects the kth period T from its own message database k The message set M received k ;

[0083] Step 2.2: Let the message set M k The jth attribute of all messages in is denoted as P k，j ;

[0084] Roadside unit statistics j-th attribute P k，j The minimum value min(P k，j ) and the maximum value max(P k，j ), and thus use formula (1) to get the j-th attribute P k，j The value range of ΔP k，j ;

[0085] ΔP k，j =max(P k，j )-min(P k，j ) (1)

[0086] Step 2.3: Define the current number of iterations as z, initialize z = 1, and define the number of data intervals for the zth iteration as μ z , and initialize Among them, ceil() means rounding up;

[0087] Use formula (2) to construct μ z Constraints:

[0088]

[0089] In formula (2), α and β represent two predefined parameters; and α≤ΔP k，j , β≥1;

[0090] Step 2.4: The roadside unit will ΔP k，j Divide into μ z intervals, thereby obtaining the attribute value interval set of the zth iteration in, Indicates the yth partition of the zth iteration z attribute value ranges, and has:

[0091]

[0092] Step 2.5: The roadside unit takes the jth attribute P k，j In M k All values ​​of μ are divided into z In the interval, the statistical result vector of the zth iteration is obtained and in, Indicates the yth partition of the zth iteration z The larger the number of j-th attributes contained in the interval, the represents the attribute value distribution of most vehicles in the kth cycle under the jth attribute;

[0093] Step 2.5.1: Traverse the roadside unit message set M k The jth attribute P in k，j For each value of , let the nth value be n∈{1,2,3,......,|M k |},|M k | represents the roadside unit message set M k The number of messages;

[0094] Step 2.5.2: Initialize n = 1;

[0095] Step 2.5.3: Initialize y z =1;

[0096] Step 2.5.4: Make express Middle y z The left boundary value of the interval, express Middle y z The right boundary value of the interval is used to determine whether formula (6) is valid. If it is valid, then Divide into , execute step 2.5.5 in sequence, otherwise, z +1 is assigned to y z , repeat step 2.5.4 until equation (6) is established;

[0097]

[0098] Step 2.5.5: If n<|M k |, assign n+1 to n, and return to step 2.5.3 and execute sequentially until n=|M k Step 2.6: The roadside unit uses machine learning algorithms to After fitting, the fitting curve of the zth iteration is obtained And calculate the fitting loss function value of the zth iteration

[0099] Step 2.7: Set μ z +1 is assigned to μ z After that, judge Is it true? If so, execute step 2.8; otherwise, get the fitting loss function value of the previous z iterations and execute step 2.9; where floor() means rounding down;

[0100] Step 2.8: After assigning z+1 to z, return to step 2.4 and execute sequentially.

[0101] Step 2.9: Obtain the number of data intervals corresponding to the minimum fitting loss function value from the fitting loss function value of the previous z iterations as the optimal data interval number μ * , μ * Can better reflect the current traffic conditions;

[0102] Step 2.10: The roadside unit uses machine learning algorithms to calculate μ * The corresponding statistical result vector Perform fitting to obtain the best fitting curve

[0103] Step 2.11: Use equation (3) to get the kth period T k The jth attribute P k，j The utility function

[0104]

[0105] In formula (3), Norm(·) represents the normalization function;

[0106] Step 2.12: Use equation (4) to get the kth period T k The jth attribute P k，j Weight Attribute weight reflects the importance of an attribute in the current scenario:

[0107]

[0108] In formula (4), express the number of nonzero elements in ;

[0109] Step 3: The vehicle receives the utility function and attribute weights and prioritizes them;

[0110] Step 3.1: When vehicle i receives the roadside unit in the current kth period T k Utility function sent and weights When the message base composed of the messages received by vehicle i in the kth cycle is recorded as make The xth message in make express The number of messages in The attribute vector in is make The jth attribute in is

[0111] Step 3.2: Initialize x=1;

[0112] Step 3.3: Vehicle i calculates the xth message using equation (5) The utility value of

[0113]

[0114] In formula (5), represents the kth period T k Message x in Attribute vector The jth attribute of Utility function of

[0115] Step 3.4: If Then assign x+1 to x and return to step 3.3 to execute sequentially. Otherwise, Figure 2 As shown, we get The utility values ​​of all messages in the , and form a utility value vector

[0116] Step 3.5: U k，i in The utility values ​​of the messages are sorted in descending order to obtain the real-time priority sorting results of the Internet of Vehicles messages, and the messages with high utility values ​​are given priority processing.

[0117] In this embodiment, the roadside unit performs a check on the kth cycle T after a certain time interval φ. k The jth attribute P k，j The resulting utility function and attribute weights To update, such as Figure 3 As shown:

[0118] Step 4.1: Preset a threshold δ; define and initialize variable b = 1;

[0119] Step 4.2: Let s k +b*φ as the k+bth period T k+b The start time s k+b , let e k +b*φ as the k+bth period T k+b End time e k+b , let the k+bth period T k+b The corresponding time period is e k+b -s k+b ;

[0120] Step 4.3: The roadside unit is selected in the k+bth period T k+b Message received within M k+b , and according to M k+b Get the k+bth period T k+b The jth attribute P k+b，j The utility function Thus, we can get With the jth attribute in the kth period T k The utility function JS Divergence

[0121]

[0122] In formula (7), KL() represents the KL divergence calculation function;

[0123] Step 4.4: If Then assign b+1 to b and return to step 4.2 for execution. Otherwise, execute step 4.5.

[0124] Step 4.5: Calculate k+b periods T using equation (4) k+bAttribute weight

[0125] Step 4.6: The roadside unit will calculate the k+bth cycle T k+b The jth attribute P obtained k+b，j The utility function and its weight And broadcast to vehicles in the local area network.

[0126] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0127] In this embodiment, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are executed.

Claims

1. A real-time priority sorting method for Internet of Vehicles messages based on machine learning and decision theory, characterized by It is applied in a local area network consisting of a trusted third party, a roadside unit, and the vehicle's onboard unit and proceeds as follows: Step 1: Vehicle message generation and broadcast; Step 1.1: Vehicle i traveling in the local area network senses and collects information about itself and the environment in real time as various attributes, including: the vehicle's current speed , the instantaneous acceleration of the vehicle , the vehicle's current heading angle , the Euclidean distance between the vehicle's current position and the roadside unit's position , thus forming the attribute vector of vehicle i ,in, represents the j-th attribute of vehicle i; Step 1.2: Vehicle i will Place in own message The message is then cached in the message queue to wait for forwarding; Step 1.3: Vehicle i broadcasts its own message to other vehicles in the local area network and the roadside unit , other vehicles and roadside units will receive the message from vehicle i Then, add it to its own message library in chronological order; Step 2: The roadside unit generates and forwards the utility function and attribute weights; Step 2.1: Let the current k-th cycle be T k ; Let the current k-th cycle T k The starting time is s k ; Let the current k-th cycle T k End time e k ; Let the kth period T k The corresponding time period t k for e k s k ; The roadside unit selects the kth period T from its own message database k The message set M received k ; Step 2.2: Let the message set M k The j-th attribute of all messages in is recorded as ; The roadside unit statistics jth attribute The minimum value min( ) and the maximum value max( ), and thus use formula (1) to get the jth attribute The value range of ∆ ; min( ) (1) Step 2.3: Define the current number of iterations as z, initialize z=1, and define the number of data intervals for the zth iteration as µ z , and initialize µ z ceil( ), where ceil() means rounding up; Use formula (2) to construct µ z Constraints: ， (2) In formula (2), and represents two predefined parameters; and , ; Step 2.4: Roadside unit will ∆ Divide into µ z intervals, thereby obtaining the attribute value interval set of the zth iteration ,in, Indicates the zth iteration division attribute value ranges, and has: ( my( ) + ( -1) * ,my( ) + * ) Step 2.5: The roadside unit takes the jth attribute In M k All values ​​in In the interval, the statistical result vector of the zth iteration is obtained ,and = { },in, Indicates the zth iteration division The number of j-th attributes contained in the interval; Step 2.6: Roadside units use machine learning algorithms to After fitting, the fitting curve of the zth iteration is obtained , and calculate the fitting loss function value L( ); Step 2.7: Set µ z +1 is assigned to µ z Then, judge µ z ≤floor( ) is true, if so, execute step 2.8; otherwise, get the fitting loss function value of the first z iterations and execute step 2.9; where floor() means rounding down; Step 2.8: After assigning z+1 to z, return to step 2.4 and execute sequentially; Step 2.9: Obtain the number of data intervals corresponding to the minimum fitting loss function value from the fitting loss function value of the previous z iterations as the optimal number of data intervals ; Step 2.10: The roadside unit uses machine learning algorithms to The corresponding statistical result vector Perform fitting to obtain the best fitting curve ; Step 2.11: Use equation (3) to get the kth period T k The jth attribute The utility function : (3) In formula (3), represents the normalization function; Step 2.12: Use equation (4) to get the kth period T k The jth attribute Weight : (4) In formula (4), express the number of nonzero elements in ; Step 3: The vehicle receives the utility function and attribute weights and prioritizes them; Step 3.1: When vehicle i receives the roadside unit in the current kth period T k Utility function sent and weights When the message base composed of the messages received by vehicle i in the kth cycle is recorded as ,make The xth message in , let | | indicates The number of messages in The attribute vector in is ( ) ,make ( ) is the jth attribute in ; Step 3.2: Initialize x=1; Step 3.3: Vehicle i calculates the xth message using equation (5) The utility value of : (5) In formula (5), represents the kth period T k Message x in Attribute vector ( The jth attribute of Utility function of Step 3.4: If x | | , then assign x+1 to x and return to step 3.3 to execute sequentially. Otherwise, it means that The utility values ​​of all messages in the , and form a utility value vector ={ , ,… , ,…, }; Step 3.5: in| The utility values ​​of the messages are sorted in descending order to obtain the real-time priority sorting results of the Internet of Vehicles messages.

2. The method for real-time prioritization of Internet of Vehicles messages based on machine learning and decision theory according to claim 1, characterized in that: The roadside unit in step 2.5 is as follows at the zth iteration: M k All values ​​in In the interval: Step 2.5.1: Iterate over the roadside unit message collection The jth attribute in For each value of , let the nth value be , n , Represents a roadside unit message collection The number of messages; Step 2.5.2: Initialize n=1; Step 2.5.3: Initialization =1; Step 2.5.4: Make express Middle The left boundary value of the interval, express Middle The right boundary value of the interval is used to determine whether formula (6) is valid. If it is valid, then Divide into , execute step 2.5.5 in sequence, otherwise, +1 assigned to , repeat step 2.5.4 until equation (6) is established; (6) Step 2.5.5: If n< , assign n+1 to n, return to step 2.5.3 and execute sequentially until n until.

3. The method for real-time prioritization of Internet of Vehicles messages based on machine learning and decision theory according to claim 1, characterized in that: The roadside unit at a certain time interval Then for the kth period T k The jth attribute The resulting utility function and attribute weights To update: Step 4.1: Set a threshold ; Define and initialize variables ; Step 4.2: Let + As the k+bth period T k+b Start time ,make As the k+bth period T k+b End time , let the k+bth period T k+b The corresponding time period is e k+b s k+b ; Step 4.3: The roadside unit is selected in the k+bth period T k+b Message received within M k+b , and according to M k+b Get the k+bth period T k+b The jth attribute The utility function , and thus we can get With the jth attribute in the kth period T k The utility function JS Divergence : (7) In formula (7), Represents the KL divergence calculation function; Step 4.4: , then assign b+1 to b and return to step 4.2 to execute sequentially. Otherwise, execute step 4.5; Step 4.5: Calculate k+b periods T using equation (4) k+b Attribute weight ; Step 4.6: The roadside unit will calculate the k+bth cycle T k+b The jth attribute obtained The utility function and its weight And broadcast to vehicles in the local area network.

4. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the real-time priority sorting method for Internet of Vehicles messages as described in any one of claims 1-3, and the processor is configured to execute the program stored in the memory.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for real-time prioritization of vehicle network messages described in any one of claims 1 to 3 are executed.

Images