Modeling method for communication and detection process of security defense system of Internet of Vehicles

Abstract

The invention discloses a modeling method for a communication and detection process of an Internet of Vehicles security defense system. The modeling method specifically comprises the following steps: step 1, establishing an unloading model of the Internet of Vehicles security defense system; 2, setting dynamic priorities of the tasks; 3, establishing a communication model, and obtaining parameters such as time delay, energy consumption and channel gain of the task in the unloading process; and 4, establishing a detection model to obtain parameters such as detection time, arrival rate, loss probability and the like of the task in the detection process. According to the invention, the important parameters of the vehicle end, the edge server and the task in the unloading process can be obtained, and a basis is provided for optimization.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle networking, and relates to a modeling method for the communication and detection processes of a vehicle networking security defense system. Background Art

[0002] With the rapid development of mobile Internet and industrial intelligence, vehicle networking centered on intelligent connected vehicles brings convenience to travel while also exposing security threats such as remote control and malicious attacks. Traditional network security defense systems, due to the large amount of data processed, are usually built on high-performance hardware platforms. However, vehicle networking belongs to an environment with limited resources (performance, energy consumption, bandwidth), and it is difficult to directly use traditional network security defense systems. Moreover, vehicles are in a moving state, and traditional task offloading methods are only applicable to task offloading at fixed locations. Based on this, the present invention proposes a modeling method for the communication and detection processes of a vehicle networking security defense system. Through modeling, calculation methods for some important parameters in the process from task generation to offloading can be obtained, providing a basis for optimization. Summary of the Invention

[0003] The purpose of the present invention is to obtain important parameters of the vehicle terminal, edge server, and tasks during the offloading process by establishing an offloading model for the vehicle networking security defense system and modeling the communication and detection processes, providing a basis for optimization.

[0004] The technical solution adopted by the present invention is a modeling method for the communication and detection processes of a vehicle networking security defense system, which specifically includes the following steps:

[0005] Step 1, establish an offloading model for the vehicle networking security defense system;

[0006] Step 2, set the dynamic priority of the task;

[0007] Step 3, establish a communication model to obtain parameters such as the delay, energy consumption, and channel gain of the task during the offloading process;

[0008] Step 4, establish a detection model to obtain parameters such as the detection time, arrival rate, and loss probability of the task during the detection process.

[0009] The characteristics of the present invention also lie in:

[0010] The specific process of Step 1 is as follows:

[0011] The structure of the vehicle networking security defense system mainly includes a vehicle terminal, a roadside unit, and an edge terminal. The vehicle terminal is mainly responsible for generating security detection tasks and locally processing the tasks (i.e., performing security detection at the vehicle terminal). The roadside unit is mainly responsible for signal communication between the vehicle terminal and the edge server, while the server at the edge terminal mainly performs task offloading, resource allocation strategy iteration, and task detection processing;

[0012] In the vehicle networking security defense system, all server nodes at the edge are represented as the set E = (e1, e2, …, e M ), where e1, e2, …, e M represent M server nodes. Each server node e m is composed of the triple e m = (f m , S m , p m ). Among them, f m represents the CPU frequency of the vehicle node e m , S m represents the communication range of the server node at the edge, and p m represents the maximum transmission power of the server node at the edge; the security detection task is generated by the vehicle-side security defense system. The vehicle-side security detection nodes (hereinafter referred to as vehicle-side nodes) are represented as the set V = (v1, v2, …, v N ), where v1, v2, …, v N represent N vehicle-side nodes. Each vehicle-side node v n is composed of the triple . Among them is the speed of the vehicle-side node, and f n represents the CPU frequency of the vehicle-side security defense system; represents the complete information of the security detection task k of the vehicle-side node n at time t. The generation time interval of the security detection task k follows a Poisson distribution. The task k at time t is composed of the quadruple . Among them, c t represents the CPU cycles required to calculate one bit of data at time t, represents the data volume size (in bits) of the task k of the vehicle-side node n at time t (l is the priority flag), τ t,k represents the deadline of the task k, and l t,k represents the dynamic priority of the task k at time t.

[0013] The specific process of step 2 is as follows:

[0014] The dynamic priority l t,k of the task k at time t is mainly related to the data volume size of the task, the task processing deadline τ t,k and the security level sl t of the task, and is set according to formula (1):

[0015]

[0016] In formula (1), the weight coefficient τ t,k and sl k Each parameter has two thresholds (α, β); the parameters can be classified according to the thresholds;

[0017] sl k represents the security level of task k, and the security level is divided into three levels: high, medium, and low through α sl and β sl These two security-related thresholds. When sl k ≤α sl the security level is low (can be denoted as 1). When α sl <sl k ≤β sl the security level is medium (can be denoted as 2). When sl k >β sl it is high level (can be denoted as 3);

[0018] The data volume size of task k is also divided into three levels: high, medium, and low according to α d and β d These two data volume size-related thresholds. When the level is low (can be denoted as 1). When the level is medium (can be denoted as 2). When the level is high (can be denoted as 3);

[0019] The processing deadline τ of task k t,k is also divided into three levels: high, medium, and low according to α τ and β τ These two task processing deadline-related thresholds. When (τt, k -t)≤α τ the level is low (can be denoted as 1). When α i,τ <(τ t,k -t)≤β i,τ the level is medium (can be denoted as 2). When (τ t,k -t)>β τ the level is high (can be denoted as 3).

[0020] The specific process of step 3 is as follows:

[0021] Step 3.1, calculate the vehicle superposition signal received by the edge node

[0022] Let p n,m (t) be the vehicle end node v assigned at time t n and the edge node e mThe transmission power between needs to satisfy the constraint represents the maximum transmission power of the vehicle; N0 is the additive white Gaussian noise (AWGN), x n represents the vehicle end node v n The signal sent, x n,m represents the vehicle end node v n to the edge node e m The signal of; represents the vehicle end node v at time t n and the edge node e m The actual channel coefficient between, represents the vehicle end node v at time t n and the edge node e m The estimated channel coefficient between, is the estimated channel error at time t; The received vehicle superimposed signal y at time t at the edge node e m (t) is calculated using the following formula (2): n,m (t), is calculated using the following formula (2):

[0023]

[0024] Step 3.2, calculate the actual channel coefficient of vehicle communication

[0025] Let the small-scale fading of the channel at time t be η n,m (t) which follows a complex Gaussian distribution with zero mean and unit variance, the large-scale path loss exponent of the channel is κ, and the path loss factor is Vehicle user v n and the edge node e m The distance between is Then the actual channel coefficient of vehicle communication at time t is calculated using the following formula (3):

[0026]

[0027] Step 3.3, calculate the estimated channel gain between the vehicle end node and the edge node

[0028] Let the communication bandwidth between the vehicle end node v at time t n and the edge node e m be w n (t), and ∑ n∈N w n (t) ≤ W, where W is the total bandwidth; The estimated channel gain between the vehicle end node v n and the edge node e m is calculated using the following formula (4):

[0029]

[0030] Among them, I(·) is an indicator function, which can be expressed as:

[0031]

[0032] Step 3.4, calculate the signal-to-interference-plus-noise ratio between the vehicle-side node and the edge node

[0033] At time t, the vehicle-side node v n and the edge node e m The signal-to-interference-plus-noise ratio SINR n,m (t) is calculated using the following formula (6):

[0034]

[0035] Step 3.5, calculate the data transmission rate between the vehicle-side node and the edge node

[0036] Let the transmission outage probability of the task be ζ0. At time t, the data transmission rate from the vehicle-side node v n to the edge node e m is (The subscript v indicates it is related to the vehicle-side), σ 2 is the variance of the channel estimation error, and is calculated using the following formula (7):

[0037]

[0038] Step 3.6, calculate the latency of the task transmitted from the vehicle-side node to the edge node

[0039] For task k, the latency from the vehicle-side node v n transmitted to the edge node e m is calculated as follows:

[0040]

[0041] In the above formula, represents the data volume size of task k at time t;

[0042] Step 3.7, calculate the energy consumption of the task transmitted from the vehicle-side node to the edge node

[0043] For task k, the transmission energy consumption from vehicle v n to the edge node e m is calculated as follows:

[0044]

[0045] The specific process of step 4 is as follows:

[0046] Step 4.1 Calculate the detection time of the task locally (i.e., at the vehicle side)

[0047] Let the computing resources at the vehicle side be f n , c t be the number of CPU cycles required to calculate 1 bit of data. Then, the time τ loc required for the vehicle side to detect task k is calculated by formula (10) as follows:

[0048]

[0049] Step 4.2 Calculate the energy consumption of the task during local detection

[0050] Let the energy consumption of the vehicle side for unit data computing resources be ε(f n ) 2 . Then, the energy consumption EC loc of the local detection task at the vehicle side is calculated by formula (11) as follows:

[0051]

[0052] Step 4.3 Calculate the arrival rate of the task at the edge node

[0053] Establish a non-preemptive priority M\M\m task processing queuing model at the edge server node. Only when all high-priority tasks in the task waiting queue for inspection are completed, the edge server starts to process low-priority tasks. Divide all tasks into L types of tasks, each with different priorities, and the arrival rates of each type of task are (λ1, λ2, …, λ l , …, λ L ). Then, the combined arrival rate of the tasks If j ≤ l, then the l-th level task has non-preemptive priority over the j-th level task. The main goal of this queue is to reduce the average waiting time of high-priority tasks while ensuring that the queue length process is stable and ergodic. When tasks enter the server queue, they are sorted according to their priorities, and tasks at the same level follow the first-come-first-served principle;

[0054] Let the number of vehicles in the entire communication range be V, the communication range of the edge node be S m , and the average speed of the vehicles in the communication range be The set of arrival rates of tasks at the edge node is the CPU frequency of the task arriving at the edge server e m is f m . The arrival rate of the task at the edge node Follow the Poisson distribution and calculate using formula (12):

[0055]

[0056] Step 4.4, calculate the service rate of the server

[0057] Let the service rate μ of the edge server be the same as the task arrival rate It has the same Poisson distribution, and each edge server is independent. The service rate μ of a single server is calculated using formula (13):

[0058]

[0059] Since in the above formula The average service rate of multiple servers at the edge side Is calculated using formula (14):

[0060]

[0061] The beneficial effect of the present invention is that by establishing an offloading model for the vehicle network security defense system and modeling the communication and detection processes, important parameters of the vehicle side, edge server, and tasks during the offloading process can be obtained, providing a basis for optimization. Description of the Drawings

[0062] Figure 1 It is through the modeling method and parameter calculation method provided by the present invention that the relationship between the energy consumption and the number of vehicles during the task detection process can be conveniently calculated;

[0063] Figure 2 It is through the modeling method and parameter calculation method provided by the present invention that the relationship between the task offloading rate and the number of vehicles can be calculated when using different offloading algorithms; Specific Embodiments

[0064] The present invention will be described in detail below in conjunction with the drawings and specific embodiments.

[0065] The modeling method for the communication and detection processes of the vehicle network security defense system of the present invention specifically includes the following steps:

[0066] Step 1, establish an offloading model for the vehicle network security defense system;

[0067] Step 2, set the dynamic priority of the task;

[0068] Step 3, establish a communication model to obtain parameters such as the delay, energy consumption, and channel gain of the task during the offloading process;

[0069] Step 4: Establish a detection model to obtain parameters such as the detection time, arrival rate, and loss probability of the task during the detection process.

[0070] The features of the present invention also lie in:

[0071] The specific process of Step 1 is as follows:

[0072] The structure of the vehicle network security defense system mainly includes a vehicle side, a roadside unit, and an edge side. The vehicle side is mainly responsible for generating security detection tasks and locally processing the tasks (i.e., performing security detection on the vehicle side). The roadside unit is mainly responsible for signal communication between the vehicle side and the edge server. The server on the edge side mainly performs task offloading, resource allocation strategy iteration, and task detection processing;

[0073] In the vehicle network security defense system, all server nodes on the edge side are represented as a set E = (e1, e2,..., e M ), where e1, e2,..., e M represent M server nodes. Each server node e m is composed of a triple e m = (f m , S m , p m ). Among them, f m represents the CPU frequency of the vehicle node e m . S m represents the communication range of the server node on the edge side. p m represents the maximum transmission power of the server node on the edge side. The security detection task is generated by the vehicle side security defense system. The vehicle side security detection node (hereinafter referred to as the vehicle side node) is represented as a set V = (v1, v2,..., v N ), where v1, v2,..., v N represent N vehicle side nodes. Each vehicle side node v n is composed of a triple . Among them is the speed of the vehicle side node, and f n represents the CPU frequency of the security defense system on the vehicle side; represents the complete information of the security detection task k of the vehicle side node n at time t. The generation time interval of the security detection task k follows a Poisson distribution. The task k at time t is composed of a quadruple . Among them, c t represents the number of CPU cycles required to calculate one bit of data at time t, represents the data volume size (in bits) of the task k (l is the priority flag) of the vehicle side node n at time t, and τ t,kDenote the deadline of task k as l t,k Denote the dynamic priority of task k at time t as l

[0074] The specific process of step 2 is as follows:

[0075] The dynamic priority l of task k at time t t,k Is mainly related to the data volume of the task The task processing deadline τ t,k And the security level sl of the task k And is set according to formula (1):

[0076]

[0077] In formula (1), the weight coefficient τ t,k And sl k Each parameter has two thresholds (α, β); the parameters can be classified according to the thresholds;

[0078] sl k Denotes the security level of task k. The security level is divided into three levels: high, medium, and low by α sl And β sl These two security-related thresholds. When sl k ≤α sl The security level is low level (can be denoted as 1). When α sl < sl k ≤β sl The security level is medium level (can be denoted as 2). When sl k >β sl It is high level (can be denoted as 3);

[0079] The data volume of task k Is also divided into three levels: high, medium, and low by α d And β d These two data volume-related thresholds. When The level is low level (can be denoted as 1). When The level is medium level (can be denoted as 2). When The level is high level (can be denoted as 3);

[0080] The task processing deadline τ of task k t,k Is also divided into three levels: high, medium, and low by α τ And β τ These two task processing deadline-related thresholds. When (τ t,k - t) ≤ α τ The level is low level (can be denoted as 1). When αi,τ <(τ t,k -t) ≤ β i,τ The level is medium level (which can be denoted as 2) when (τ t,k -t) > β τ The level is high level (which can be denoted as 3).

[0081] The specific process of step 3 is as follows:

[0082] Step 3.1, calculate the vehicle superposition signal received by the edge node

[0083] Let p n,m (t) be the transmission power allocated to the vehicle end node v n and the edge node e m at time t, which needs to satisfy the constraint represents the maximum transmission power of the vehicle; N0 is the additive white Gaussian noise (AWGN), and x n represents the signal sent by the vehicle end node v n ; x n,m represents the signal from the vehicle end node v n to the edge node e m ; represents the actual channel coefficient between the vehicle end node v n and the edge node e m at time t, represents the estimated channel coefficient between the vehicle end node v n and the edge node e m at time t, is the estimated channel error at time t; the vehicle superposition signal y m (t) received at the edge node e n,m at time t is calculated using the following formula (2):

[0084]

[0085] Step 3.2, calculate the actual channel coefficient of vehicle communication

[0086] Let the small-scale fading of the channel at time t be η n,m (t) which follows a complex Gaussian distribution with zero mean and unit variance, the large-scale path loss exponent of the channel is κ, and the path loss factor is The distance between the vehicle user v n and the edge node e m is Then the actual channel coefficient of vehicle communication at time t is calculated using the following formula (3):

[0087]

[0088] Step 3.3, calculate the estimated channel gain between the vehicle end node and the edge node

[0089] Let the communication bandwidth between the vehicle end node v n and the edge node e m at time t be w n (t), and ∑ n∈N w n (t) ≤ W, where W is the total bandwidth; the estimated channel gain between the vehicle end node v n and the edge node e m is calculated using the following formula (4): The calculation is as follows:

[0090]

[0091] where I(·) is the indicator function and can be expressed as:

[0092]

[0093] Step 3.4, calculate the signal-to-interference-plus-noise ratio between the vehicle end node and the edge node

[0094] The signal-to-interference-plus-noise ratio SINR n between the vehicle end node v m and the edge node e n,m (t) at time t is calculated using the following formula (6):

[0095]

[0096] Step 3.5, calculate the data transmission rate between the vehicle end node and the edge node

[0097] Let the transmission outage probability of the task be ζ0, and the data transmission rate n from the vehicle end node v m to the edge node e (the subscript v indicates it is related to the vehicle end) at time t, where σ 2 is the variance of the channel estimation error, is calculated using the following formula (7):

[0098]

[0099] Step 3.6, calculate the delay for the task to be transmitted from the vehicle end node to the edge node

[0100] For task k, the delay n for transmitting from the vehicle end node v m to the edge node e is calculated as follows:

[0101]

[0102] In the above formula, represents the amount of data of task k at time t;

[0103] Step 3.7, calculate the energy consumption of the task transmitted from the vehicle-side node to the edge node

[0104] For task k, from vehicle v n to edge node e m The transmission energy consumption The calculation method is:

[0105]

[0106] The specific process of Step 4 is as follows:

[0107] Step 4.1 Calculate the detection time of the task locally (i.e., at the vehicle side)

[0108] Let the computing resource at the vehicle side be f n , c t be the number of CPU cycles required to calculate 1 bit of data. Then, the time τ loc required for the vehicle side to detect task k is calculated by formula (10) as:

[0109]

[0110] Step 4.2 Calculate the energy consumption of the task detected locally

[0111] Let the computing resource energy consumption per unit data at the vehicle side be ε(f n ) 2 , then the energy consumption EC loc of detecting the task locally at the vehicle side is calculated by formula (11) as:

[0112]

[0113] Step 4.3, calculate the arrival rate of the task at the edge node

[0114] Establish a non-preemptive priority M\M\m task processing queuing model at the edge server node. Only when all high-priority tasks in the task waiting queue are completed, the edge server starts to process low-priority tasks; assume that all tasks are divided into L categories according to their types, each with different priorities, and the arrival rates of each category of tasks are (λ1, λ2,..., λ l ,..., λ L ), then the combined arrival rate of the tasks If \(j \leq l\), the \(l\)-th level task has non-preemptive priority over the \(j\)-th level task; the main goal of this queue is to reduce the average waiting time of high-priority tasks while ensuring that the queue length process is stationary and ergodic; when tasks enter the server queue, they are sorted according to their priorities, and tasks at the same level follow the first-come-first-served principle;

[0115] Let the number of vehicles in the entire communication range be \(V\), and the communication range of the edge node be \(S\) m , and the average speed of vehicles in the communication range is The set of arrival rates of tasks arriving at the edge node is the task arrival at the edge server \(e\) m The CPU frequency of is \(f\) m , and the arrival rate of tasks arriving at the edge node follows a Poisson distribution and is calculated using formula (12):

[0116]

[0117] Step 4.4, calculate the service rate of the server

[0118] Let the service rate \(\mu\) of the edge server be the same as the task arrival rate has the same Poisson distribution, and each edge server is independent. The service rate \(\mu\) of a single server is calculated using formula (13):

[0119]

[0120] Since in the above formula The average service rate of multiple servers at the edge is calculated using formula (14):

[0121]

[0122] Embodiment 1

[0123] Through the modeling method and parameter calculation method provided by the present invention, the relationship between the energy consumption and the number of vehicles during the task detection process can be conveniently calculated, as Figure 1 shown.

[0124] Embodiment 2

[0125] Through the modeling method and parameter calculation method provided by the present invention, the relationship between the task offloading rate and the number of vehicles can be calculated when using different offloading algorithms, as Figure 2 shown.

Claims

1. The technical solution adopted by the present invention is a modeling method for the communication and detection process of the Internet of Vehicles security defense system, which specifically includes the following steps: Step 1: Establish an uninstall model for the Internet of Vehicles security defense system; Step 2: Set the dynamic priority of the task; Step 3: Establish a communication model to obtain the parameters such as delay, energy consumption, channel gain, etc. of the task during the offloading process; Step 4: Establish a detection model to obtain parameters such as detection time, arrival rate, and loss probability of the task during the detection process.

2. The modeling method for the communication and detection process of the Internet of Vehicles security defense system according to claim 1 is characterized by: The specific process of step 1 is as follows: The structure of the Internet of Vehicles security defense system mainly includes the vehicle side, roadside unit and edge side. The vehicle side is mainly responsible for generating security detection tasks and processing them locally (i.e., performing security detection on the vehicle side). The roadside unit mainly connects the signal communication between the vehicle side and the edge server, while the edge server mainly performs task unloading, resource allocation strategy iteration and task detection processing. Assume that in the Internet of Vehicles security defense system, all edge server nodes are represented as a set E = (e1, e2, …, e M ), where e1, e2, …, e M Represents M server nodes, each server node e m By triple e m =(f m ,S m ,p m ), where f m Represents the vehicle node e m CPU frequency, S m represents the communication range of the server node at the edge, p m represents the maximum transmission power of the server node at the edge; the safety detection task is generated by the vehicle-side safety defense system, and the vehicle-side safety detection node (hereinafter referred to as the vehicle-side node) is represented as a set V = (v1, v2, …, v N ), where v1,v2,…,v N Represents N vehicle end nodes, each vehicle end node v n All are composed of triples Composition, among which is the speed of the vehicle end node, f n Represents the CPU frequency of the vehicle's security defense system; represents the complete information of the safety detection task k of the vehicle end node n at time t. The generation time interval of the safety detection task k follows the Poisson distribution. The task k at time t is composed of the four-tuple Composition, among which c t It represents the CPU cycles required to calculate one bit of data at time t. represents the data size (in bits) of task k (l is the priority flag) at vehicle end node n at time t, τ t,k represents the deadline of task k, l t,k represents the dynamic priority of task k at time t.

3. The modeling method for the communication and detection process of the Internet of Vehicles security defense system according to claim 2 is characterized by: The specific process of step 2 is: The dynamic priority l of task k at time t t,k Mainly related to the amount of data in the task Task processing deadline τ t,k and the security level of the task k Related, set according to formula (1): In formula (1), the weight coefficient τ t,k and sl k Each parameter has two thresholds (α, β); the parameters can be graded according to the thresholds; sl k represents the safety level of task k, and the safety level is determined by α sl and β sl These two safety-related thresholds are divided into three levels: high, medium and low. k ≤α sl The security level is low (which can be recorded as 1). sl <sl k ≤β sl The safety level is medium (can be recorded as 2). k >β sl It is a high level (can be recorded as 3); The data size of task k Also according to α d and β d These two thresholds related to the data size are divided into three levels: high, medium and low. When the level is low (can be recorded as 1), The level is medium (can be recorded as 2), when The level is high (can be recorded as 3); The processing deadline τ of task k t,k Also according to α τ and β τ These two thresholds related to the task processing deadline are divided into three levels: high, medium and low. t,k -t)≤α τ When the level is low (can be recorded as 1), when α i,τ <(τ t,k -t)≤β i,τ The level is medium (can be recorded as 2), when (τ t,k -t)>β τ The level is high (can be recorded as 3).

4. The modeling method for the communication and detection process of the Internet of Vehicles security defense system according to claim 3 is characterized by: The specific process of step 3 is as follows: Step 3.1, calculate the vehicle superposition signal received by the edge node Let p n,m (t) is the node v assigned to the vehicle at time t n and edge node e m The transmission power between represents the maximum transmission power of the vehicle; N0 is additive white Gaussian noise (AWGN), x n Represents the vehicle end node v n The signal sent, x n,m Represents the vehicle end node v n To the edge node e m signal; represents the vehicle terminal node v at time t n and edge node e m The actual channel coefficient between represents the vehicle terminal node v at time t n and edge node e m The estimated channel coefficients between is the estimated channel error at time t; at time t, at the edge node e m The vehicle superimposed signal y received at n,m (t) is calculated using the following formula (2): Step 3.2, calculate the actual channel coefficients for vehicle communication Assume that the small-scale fading of the channel at time t is η n,m (t) It obeys a complex Gaussian distribution with zero mean and unit variance. The large-scale path loss index of the channel is κ, and the path loss factor is Vehicle User n and edge node e m The distance between Then the actual channel coefficient of vehicle communication at time t is The following formula (3) is used for calculation: Step 3.3, calculate the estimated channel gain between the vehicle end node and the edge node Assume that at time t the vehicle terminal node v n and edge node e m The communication bandwidth between them is w n (t), and ∑ n∈N w n (t)≤W, W is the total bandwidth; the vehicle end node v n and edge node e m The estimated channel gain between The following formula (4) is used for calculation: Where I(·) is the indicator function, which can be expressed as: Step 3.4, calculate the signal interference noise ratio between the vehicle end node and the edge node The vehicle end node v at time t n and edge node e m SINR n,m (t) is calculated using the following formula (6): Step 3.5, calculate the data transmission rate between the vehicle end node and the edge node Assume that the transmission interruption probability of the task is ζ0, and the vehicle terminal node v at time t n To the edge node e m The data transfer rate is (The symbol v indicates that it is related to the vehicle end), σ 2 is the variance of the channel estimation error, The following formula (7) is used for calculation: Step 3.6, calculate the delay of the task from the vehicle end node to the edge node For task k, from the vehicle end node v n Transmit to edge node e m Delay The calculation method is: In the above formula, Indicates the data size of task k at time t; Step 3.7, calculate the energy consumption of task transmission from vehicle end node to edge node. For task k, from vehicle v n To the edge node e m Transmission energy consumption The calculation method is:

5. The modeling method for the communication and detection process of the Internet of Vehicles security defense system according to claim 4 is characterized by: The specific process of step 4 is as follows: Step 4.1 Calculate the detection time of the task locally (i.e. on the vehicle side) Assume the computing resources of the vehicle side are f n , c t The number of CPU cycles required to calculate 1 bit of data is τ, and the time required for vehicle-side detection task k is loc The calculation formula (10) is: Step 4.2 Calculate the energy consumption of the task detected locally Assume that the computing resource energy consumption of the vehicle end for unit data is ε(f n ) 2 , then the energy consumption of the local detection task on the vehicle side is EC loc The calculation formula (11) is: Step 4.3, calculate the arrival rate of tasks to edge nodes A non-preemptive priority M\M\m task processing queuing model is established at the edge server node. Only when all high-priority tasks in the task waiting queue are completed, the edge server starts to process low-priority tasks. Suppose all tasks are divided into L categories according to their categories, each with different priorities. The arrival rates of each category of tasks are (λ1,λ2,…,λ l ,…,λ L ), then the joint arrival rate of the tasks If j≤l, then the level l task has non-preemptive priority over the level j task; the main goal of this queue is to reduce the average waiting time of high priority tasks while ensuring that the queue length process is smooth and traversable; when tasks enter the server queue, they are sorted according to priority, and tasks of the same level follow the first-come, first-served principle; Assume that the number of vehicles within the entire communication range is V, and the communication range of the edge node is S m , the average speed of vehicles within the communication range is The arrival rate of tasks arriving at edge nodes is the set of task path edge servers e m The CPU frequency is f m , the arrival rate of tasks arriving at edge nodes Following the Poisson distribution, the formula (12) is used to calculate: Step 4.4, calculate the service rate of the server Assume that the service rate μ of the edge server is equal to the task arrival rate With the same Poisson distribution, and each edge server is independent of each other, the service rate μ of a single server is calculated using formula (13): Since in the above formula Average service rate of multiple servers at the edge Calculate using formula (14):

Images