An edge computing task offloading evaluation method in a vehicle-road cooperation environment

Abstract

The application discloses a kind of edge computing task unloading evaluation methods under vehicle-road cooperation environment, comprising:1, position deployment and equipment number are carried out to roadside unit and network edge server in vehicle-road cooperation environment, and it is composed of network containing multiple edge computing units, realize that computing resource is distributed on demand;2, intelligent network connection car computing task is divided into three types of processing tasks according to processing demand;3, delay and computing resource are considered comprehensively, and the behavior of the third type of computing task of intelligent network connection vehicle is unloaded to network edge server on road is established model evaluation;4, in combination with the evaluation value obtained by evaluation model and the threshold value set, the unloading behavior of the third type of computing task of intelligent network connection vehicle is constrained.The application guarantees the efficiency, rationality of road roadside network edge server computing and storage resource use under the premise of meeting the basic operation demand of intelligent network connection car in road.

Description

Technical Field

[0001] The present invention relates to the field of traffic information communication technology, and specifically to a method for evaluating the offloading of edge computing tasks in a vehicle-road collaborative environment. Background Art

[0002] With the rapid growth of IoT devices and smart terminals, traditional cloud computing models are facing latency and bandwidth bottlenecks. Edge computing effectively reduces data transmission time and network congestion by offloading computing tasks to edge devices close to the data source, thereby improving response speed and system reliability. However, in a vehicle-infrastructure collaborative environment, the massive amount of real-time data generated by intelligent connected vehicles and the high demand for low latency put relatively tight computing resources on network edge servers. Furthermore, the uneven distribution of traffic data makes it difficult to efficiently utilize computing resources on network edge servers. Therefore, there is an urgent need to adopt more efficient computing offloading strategies to optimize resource utilization and improve system performance.

[0003] Effective offloading strategies must not only consider the task's computational requirements and network conditions, but also comprehensively evaluate factors such as device energy consumption, real-time performance, and task priority. Current offloading strategies often consider the shared computational resources of local devices and network edge servers when allocating computational resources. However, from the perspective of local devices, computational resource allocation primarily focuses on task data transmission, resulting in existing offloading strategies being relatively strict in restricting offloading behavior. This restriction is detrimental to the efficient utilization of network edge server resources. Established offloading strategies will further curb offloading behavior, hindering the efficient utilization of network edge server resources. Summary of the Invention

[0004] In order to address the shortcomings of the above-mentioned existing technologies, the present invention proposes an offloading evaluation method for edge computing tasks in a vehicle-road collaborative environment, so as to ensure the efficiency and rationality of the use of computing and storage resources of network edge servers while meeting the basic operating requirements of intelligent connected vehicles on the road.

[0005] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0006] The characteristic of the offloading evaluation method of edge computing tasks in a vehicle-road cooperative environment of the present invention is that on the road side of the vehicle-road cooperative environment, a communication radius of A roadside unit and a network edge server based on mobile edge computing are configured, so that n roadside units RSU_1, RSU_2, ..., RSU_j, ..., RSU_n and n network edge servers MEC_1, MEC_2, ..., MEC_j, ..., MEC_n are sequentially arranged according to the vehicle's forward direction, and a network including multiple edge computing units is formed; wherein RSU_j represents the jth roadside unit; MEC_j represents the jth network edge server; the jth within the range of the road served by RSU_j is set as the Intelligent connected vehicles are recorded as , will smart connected vehicles The CPU cycle frequency of the built-in server is The uninstallation evaluation method is carried out in the following steps:

[0007] Step 1: Use formula (1) to obtain the computing resources allocated by MEC_j :

[0008] (1)

[0009] In formula (1), is the sum of unused computing resources of n network edge servers within the road, is the sum of the number of intelligent connected vehicles within the service road range of RSU_j, is the sum of the number of intelligent connected vehicles within the service range of all roadside units and is obtained by formula (2);

[0010] (2)

[0011] Step 2: According to the processing requirements of intelligent connected vehicles for computing tasks, Computational tasks Divided into three categories, including: the first category of computing tasks To allow only smart connected vehicles The built-in server performs local computing tasks, including computing tasks from network edge servers and computing tasks from roadside units;

[0012] The second type of computing tasks It is an information task shared with the roadside unit without the need for data feedback;

[0013] The third type of computing tasks This allows both built-in servers to perform local computing tasks and offload computing tasks to network edge servers.

[0014] Step 3: If Computational tasks for or , then end the process, if Computational tasks for , then execute step 4;

[0015] Step 4: Comprehensive consideration When computing locally The latency and computational resources required, and When offloading to MEC_j for calculation Required latency and computing resources, establish Evaluation model when unloading to MEC_j, used to calculate Evaluation value when offloading to MEC_j calculation ;

[0016] Step 5: If , it means Do not offload to MEC_j for calculation; otherwise, it means Unload to MEC_j for calculation, where Indicates the set threshold.

[0017] The method for evaluating offloading of edge computing tasks in a vehicle-road collaborative environment according to the present invention is also characterized in that step 4 includes:

[0018] Step 4.1: Use formula (3) to get exist Computation time on the built-in server ;

[0019] (3)

[0020] In formula (3), for The complexity of for The data size;

[0021] Step 4.2: Use formula (4) to get exist Computing resources required for calculations on the built-in server ;

[0022] (4)

[0023] In formula (4), for Energy consumption parameters of the built-in server;

[0024] Step 4.3: Using equations (5) and (6), we can obtain Data transmission time with RSU_j in uplink , data transmission time in downlink , and thus using formula (7) we can get Computation time on MEC_j , and then use formula (8) to get Will Computation time offloaded to MEC_j ;

[0025] (5)

[0026] (6)

[0027] (7)

[0028] (8)

[0029] In formula (5) to formula (6), for The data transmission overhead of RSU_j in the uplink is for The data transmission overhead of RSU_j in the downlink is is the data transmission rate;

[0030] In formula (7), is the CPU cycle frequency of MEC_j;

[0031] Step 4.4: Use formula (9) to get When calculating MEC_j, Computing resources allocated by the built-in server ;

[0032] (9)

[0033] In formula (9), for for the allocated transmit power;

[0034] Step 4.5: Use formula (10) to get The total remaining computing resources of the network edge servers , and using formula (11) we get Pricing factor ; Thus, using formula (12) we can get Evaluation value when offloading to MEC_j calculation ;

[0035] (10)

[0036] (11)

[0037] (12)

[0038] In formula (10), is the kth network edge server MEC_ Remaining computing resources;

[0039] In formula (12), for exist The scaling factor of the computing resources required for calculation on the built-in server, for exist The scaling factor for calculating the delay on the built-in server, and 0 , , ;

[0040] The 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 the uninstall evaluation method, and the processor is configured to execute the program stored in the memory.

[0041] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program executes the steps of the uninstallation evaluation method when the computer program is executed by a processor.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. This paper proposes an innovative computation offloading method in a vehicle-infrastructure collaborative environment, specifically for applications connecting intelligent connected vehicles. This method optimizes the allocation of computational tasks to intelligent connected vehicles by establishing a detailed evaluation model, thereby effectively balancing local and edge computing.

[0044] 2. In traditional edge computing models, computational offloading usually relies on static strategies, which may not always be optimal in terms of processing real-time performance and energy efficiency. The method of the present invention, from the perspective of intelligent connected vehicles, establishes a dynamic evaluation model that is based on the comparison between local computing and network edge computing and can flexibly adjust the offloading strategy. The evaluation model contains proportional factors related to computing resources and latency, which can be externally adjusted according to specific needs. If the current environmental conditions or task priorities require priority saving of computing resources, the system can adjust the proportional factors and increase the weight of computing resources, so that computing tasks are more likely to be offloaded to the network edge server to reduce the consumption of computing resources of the built-in server of the intelligent connected vehicle; if the task is less sensitive to time, the time proportional factor can be increased so that the task is preferentially calculated locally to reduce delays. Such dynamic adjustment capabilities enable this method to adapt to various complex and changing road scenarios and task requirements.

[0045] 3. Through this flexible adjustment mechanism, the present invention enables intelligent connected vehicles to decide whether to offload computations based on the evaluation value obtained from the evaluation model and a preset threshold. This not only improves the utilization efficiency of network edge server computing and storage resources, but also optimizes the overall performance of task processing. The present invention's method significantly improves the efficiency of computing resource allocation in vehicle-infrastructure collaborative environments, especially in dynamic and changing autonomous driving environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of the execution steps of the present invention;

[0047] Figure 2 This is a diagram showing the layout of the roadside units and network edge servers of the present invention;

[0048] Figure 3 This is a flowchart of the evaluation model of the present invention. DETAILED DESCRIPTION

[0049] In this embodiment, Figure 2 As shown, on the road side of the vehicle-road cooperative environment, each interval of a distance m respectively set a communication radius of m roadside units and a network edge server based on mobile edge computing, so that n roadside units RSU_1, RSU_2, ..., RSU_j, ..., RSU_n and n network edge servers MEC_1, MEC_2, ..., MEC_j, ..., MEC_n are set in sequence according to the vehicle's forward direction, and form a network containing multiple edge computing units; among them, RSU_j represents the jth roadside unit, and supports intelligent connected vehicles to perform computing offload within the scope of their service road, transferring part of the computing tasks to the network edge server for computing; MEC_j represents the jth network edge server; within the road Each roadside unit exchanges information every 3 seconds and reallocates computing resources on the network edge server. The exchanged information includes the number of intelligent connected vehicles within the service area of ​​each roadside unit and other information that needs to be shared.

[0050] The first Intelligent connected vehicles are recorded as , in smart connected vehicles There is a built-in server in the system, which supports local computing of computing tasks. The CPU cycle frequency of the built-in server is ;like Figure 1 As shown in the figure, a method for evaluating the offloading of edge computing tasks in a vehicle-road cooperative environment is performed in the following steps:

[0051] Step 1: Use formula (1) to obtain the computing resources allocated by MEC_j , used to support computing vehicles Computational tasks offloaded to MEC_j;

[0052] (1)

[0053] In formula (1), is the sum of unused computing resources of n network edge servers within the road, is the sum of the number of intelligent connected vehicles within the service range of the j-th roadside unit RSU_j, is the sum of the number of intelligent connected vehicles within the service range of all roadside units and is obtained by formula (2);

[0054] (2)

[0055] Step 2: Calculation task Local computing takes a certain amount of time and computing resources, but the computing power of its built-in server is limited and cannot meet the latency and computing resource requirements of all computing tasks. Some computing tasks need to be offloaded to MEC_j. According to the processing requirements of intelligent connected vehicles for computing tasks, Computational tasks are divided into three categories; including: the first category of computing tasks To allow only smart connected vehicles The built-in server performs local computing tasks, including computing tasks from network edge servers and computing tasks from roadside units;

[0056] The second type of computing tasks It is an information task shared with the roadside unit without the need for data feedback;

[0057] The third type of computing tasks This allows both built-in servers to perform local computing tasks and offload computing tasks to network edge servers.

[0058] Step 3: If Computational tasks for or , then end the process, if Computational tasks for , then execute step 4;

[0059] Step 4: Comprehensive consideration When computing locally The latency and computational resources required, and When offloading to MEC_j for calculation Required latency and computing resources, establish The evaluation model when unloading to MEC_j is as follows Figure 3 As shown;

[0060] Step 4.1: Use formula (3) to get exist Computation time on the built-in server ;

[0061] (3)

[0062] In formula (3), for The complexity of for The data size.

[0063] Step 4.2: When When computing locally, the required computing resources are similar to The CPU cycle frequency of the built-in server is related to the principle of dynamic voltage frequency adjustment technology, and the formula (4) is used to obtain exist Computing resources required for calculations on the built-in server ;

[0064] (4)

[0065] In formula (4), for The energy consumption parameters of the built-in server are related to the hardware structure of the local device.

[0066] Step 4.3: Using equations (5) and (6), we can obtain Data transmission time with RSU_j in uplink , data transmission time in downlink , and thus using formula (7) we can get Computation time on MEC_j , and then use formula (8) to get Will Computation time offloaded to MEC_j ;

[0067] (5)

[0068] (6)

[0069] (7)

[0070] (8)

[0071] In formula (5) to formula (6), for The data transmission overhead of RSU_j in the uplink is for The data transmission overhead of RSU_j in the downlink is is the data transmission rate;

[0072] In formula (7), is the CPU cycle frequency of MEC_j.

[0073] Step 4.4: According to the Shannon-Hartley law, use formula (9) to obtain When calculating MEC_j, Computing resources allocated by the built-in server ;

[0074] (9)

[0075] In formula (9), for for the allocated transmit power;

[0076] Since the size of the MEC_j calculation result is usually much smaller than the input data, the computing resources allocated for the transmission of the calculation result and the reprocessing of the transmitted data can be ignored.

[0077] Step 4.5: Use formula (10) to get the road The total remaining computing resources of the network edge servers , and using formula (11) we get Pricing factor ; Thus, using formula (12) we can get Evaluation value when offloading to MEC_j calculation ;

[0078] (10)

[0079] (11)

[0080] (12)

[0081] In formula (10), is the kth network edge server MEC_ Remaining computing resources;

[0082] In formula (12), for exist The scaling factor of the computing resources required for calculation on the built-in server, for exist The scaling factor for calculating the delay on the built-in server, and 0 , , .

[0083] Step 5: If , it means Do not offload to MEC_j for calculation; otherwise, it means Unload to MEC_j for calculation; where, Indicates the set threshold value. In this embodiment, The value is 、 、 hour Evaluation value when offloading to MEC_j calculation.

[0084] 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.

[0085] 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 method for evaluating offloading edge computing tasks in a vehicle-road collaborative environment, characterized in that: Let RSU_ j Indicates the j Roadside units; let MEC_ j Indicates the j Network edge servers; RSU_ j Within the service road Intelligent connected vehicles are recorded as , will smart connected vehicles The CPU cycle frequency of the built-in server is The uninstallation evaluation method is carried out in the following steps: Step 1: Use formula (1) to get MEC_ j Allocated computing resources : (1) In formula (1), is the sum of unused computing resources of n network edge servers within the road, RSU_ j The total number of intelligent connected vehicles within the service road range, is the sum of the number of intelligent connected vehicles within the service range of all roadside units and is obtained by formula (2); (2) Step 2: According to the processing requirements of intelligent connected vehicles for computing tasks, Computational tasks include: The third type of computing tasks This allows both built-in servers to perform local computing tasks and offload computing tasks to network edge servers. Step 3: If Computational tasks for , then execute step 4; Step 4: Comprehensive consideration When computing locally The latency and computational resources required, and Offload to MEC_ j When performing calculations Required latency and computing resources, establish Offload to MEC_ j The evaluation model for calculating Offload to MEC_ j Evaluation value during calculation ; Step 4.1: Use formula (3) to get exist Computation time on the built-in server ; (3) In formula (3), for The complexity of for The data size; Step 4.2: Use formula (4) to get exist Computing resources required for calculations on the built-in server ; (4) In formula (4), for Energy consumption parameters of the built-in server; Step 4.3: Using equations (5) and (6), we can obtain with RSU_ j Data transmission time in uplink , data transmission time in downlink , and thus using formula (7) we can get In MEC_ j Computation time on , and then use formula (8) to get Will Offload to MEC_ j Computation time ; (5) (6) (7) (8) In formula (5) to formula (6), for with RSU_ j Data transmission overhead in the uplink, for with RSU_ j Data transmission overhead in the downlink, is the data transmission rate; In formula (7), For MEC_ j CPU cycle frequency; Step 4.4: Use formula (9) to get In MEC_ j When calculating, Computing resources allocated by the built-in server ; (9) In formula (9), for for the allocated transmit power; Step 4.5: Use formula (10) to get The total remaining computing resources of the network edge servers , and using formula (11) we get Pricing factor ; Thus, using formula (12) we can get Offload to MEC_ j Evaluation value during calculation ; (10) (11) (12) In formula (10), is the kth network edge server MEC_ Remaining computing resources; In formula (12), for exist The scaling factor of the computing resources required for calculation on the built-in server, for exist The scaling factor for calculating the delay on the built-in server, and 0 , , ; Step 5: If , it means Do not offload to MEC_ j Otherwise, it means Offload to MEC_ j Calculate on, where Indicates the set threshold.

2. The offloading evaluation method for edge computing tasks in a vehicle-road collaborative environment according to claim 1 is characterized in that: Computational tasks Also includes: the first type of computing tasks To allow only smart connected vehicles The built-in server performs local computing tasks, including computing tasks from network edge servers and computing tasks from roadside units; The second type of computing tasks It is an information task shared with the roadside unit and does not require data feedback.

3. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports a processor to execute the uninstallation evaluation method according to claim 1 or 2, and the processor is configured to execute the program stored in the memory.

4. 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 uninstallation evaluation method according to claim 1 or 2 are executed.

Images