Internet of vehicles task unloading and forwarding algorithm

Abstract

The invention relates to the technical field of data processing, in particular to an Internet of Vehicles task unloading and forwarding algorithm based on an MEC and an SDN controller. The real-time states of all vehicles within the coverage range of a roadside unit (RSU) are obtained through an SDN controller, available optimal relay vehicle information is fed back to a current MEC server, the MEC server calculates a global optimal forwarding path by using a bat algorithm, forwarding vehicle node information is sent to a task unloading vehicle, and the task unloading vehicle performs task unloading. And the vehicle unloads and forwards a task which cannot obtain a timely calculation service from the current MEC server to an idle MEC server in a future arrival road section through a forwarding vehicle node for calculation, and when the vehicle with the task unloaded arrives at the server, a calculation result is obtained from the server. The calculation cost and the calculation time delay can be effectively reduced, and the calculation performance and the user experience of the vehicle are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and to a task offloading and forwarding strategy based on MEC and software defined network (SDN) controllers, specifically, a task offloading and forwarding algorithm for Internet of Vehicles based on MEC and SDN controllers. Background Art

[0002] Vehicular Ad Hoc Networks (VANETs) have emerged as a promising technology for providing safety and comfort services to vehicle users. The three commonly discussed communication modes in VANETs are vehicle-to-infrastructure (V2I), vehicle-to-vehicle (V2V), and V2I2V. In VANETs, ​​vehicles are equipped with onboard units (OBUS) based on dedicated short-range communications (DSRC). Through OBUS, vehicles can communicate between themselves and with fixed roadside units (RSUs) using IEEE802.11p communication technology, and RSUs can provide information support and connectivity to passing vehicles, including traffic information and safety warnings. The dynamic topology changes caused by the mobility of vehicles make offloading transmission complicated. In MEC cloud-based vehicular networks, in order to improve the efficiency of task completion, an optimal task offloading scheme with MEC server selection and communication management must be provided. VANET reduces the amount of data transmitted through the cellular network by offloading data to the MEC server, increasing the available bandwidth allowed for other users. The main advantages of data offloading are reduced network burden on LTE networks, increased bandwidth provided, or saved costs for using cellular networks.

[0003] Software Defined Networking (SDN) is a centralized approach to network management where the underlying network infrastructure is abstracted from the application. SDN provides automation and programmability options for data centers, campuses, and wide area networks, helping to make intelligent decisions. Due to the separation of data and control planes, SDN is an approach that helps in dynamic optimization processes. Summary of the invention

[0004] In recent years, many studies have introduced SDN controllers into VANET task offloading, such as: proposing a V2V task offloading algorithm for software-defined vehicle cooperative networks (SVCN). In a multi-access edge computing server architecture, a predictive utility-centered path switching (PUPS) method for k-hop V2V data offloading is proposed based on an SDN controller; mobile edge computing nodes and fixed edge computing nodes are integrated to collaboratively provide low-latency computing services. In order to better utilize these heterogeneous edge computing resources, the concepts of software-defined networking and edge-computing-aided IOV (EC-SDIOV) are proposed.

[0005] However, the above studies all assume that the system is quasi-static and do not take into account the problem of connection interruption during vehicle-to-vehicle or vehicle-to-infrastructure communication due to the different future driving directions of the requesting vehicle and other vehicles. Once the connection between the requesting vehicle and the MEC server is interrupted during the offloading of computing tasks, the network latency and computing cost will increase.

[0006] Therefore, an offloading task forwarding calculation scheme is proposed in which the SDN controller is deployed in the cloud to forward the offloading task of the requesting vehicle to the idle MEC server of its future arrival section by the auxiliary relay vehicle.

[0007] Compared with other schemes, the computational cost of forwarding the unloading task of the requesting vehicle to the idle MEC server through the relay vehicle is minimized within its delay tolerance while satisfying the unloading task delay constraint.

[0008] A task offloading and forwarding algorithm based on mobile edge computing (MEC) and software defined network (SDN) controllers (Deploying SDN controllers in the cloud for assisted MEC, SAM) includes the following steps: (1) Obtain the real-time status of all vehicles within the coverage area of ​​the roadside unit (RSU) through the SDN controller deployed in the cloud; (2) Feedback the available optimal relay vehicle information to the current MEC server; (3) The MEC server uses the bat algorithm to calculate the global optimal forwarding path and sends the forwarding vehicle node information to the task unloading vehicle; (4) The task offloading vehicle will forward the tasks that cannot obtain timely computing services from the current MEC server through the forwarding vehicle node to the idle MEC server in the future arrival section for computing; (5) When the task unloading vehicle arrives at this server, the calculation results are obtained from the server.

[0009] The SDN controller is deployed in the cloud and can dynamically adjust the coverage of the RSU and the resource allocation of the MEC server according to the vehicle's location, speed, direction, task type and other information. The bat algorithm is an optimization algorithm based on swarm intelligence, which can find the global optimal solution in a multi-dimensional space and has the ability to converge quickly and avoid falling into local optimality. The forwarding vehicle node refers to a vehicle with wireless communication and storage capabilities, which can establish a relay link between the task offloading vehicle and the target MEC server to achieve task offloading and forwarding. The task offloading vehicle refers to a vehicle that needs to perform computing-intensive or data-intensive tasks. The task can be offloaded to the MEC server for calculation to save local resources and improve computing efficiency and user experience.

[0010] The purpose of the present invention is to solve the problem that the current MEC server cannot provide timely service for the unloading tasks of vehicles in the coverage area and the problem caused by the inconsistency of the future driving direction of the vehicles in the high-density vehicle scenario. The technical solution of the present invention is: the real-time status of all vehicles within the coverage area of ​​the roadside unit (RSU) is obtained through the SDN controller, and the available optimal relay vehicle information is fed back to the current MEC server. The MEC server uses the bat algorithm to calculate the global optimal forwarding path and sends the forwarding vehicle node information to the task unloading vehicle. The vehicle will unload the tasks that the current MEC server cannot obtain timely computing services through the forwarding vehicle node to the idle MEC server in the future arrival section for calculation. When the task unloading vehicle arrives at this server, the calculation result is obtained from the server.

[0011] The purpose of the present invention is to provide a vehicle networking task offloading and forwarding algorithm based on MEC and SDN controller to solve the shortcomings of the existing technology. The technical solution includes: (1) By deploying the SDN controller in the cloud, the vehicle information within the coverage area of ​​each RSU can be globally perceived, and the idle computing resources of vehicles on the road can be fully utilized to improve the efficiency of establishing multi-hop links while reducing the resource waste caused by randomly selecting relay vehicles.

[0012] (2) A detailed analysis is performed on the time it takes for available relay vehicles to forward offloaded tasks, and the optimal multi-hop chain for offloading computing tasks to idle MEC servers on the road section where the requesting vehicle will arrive in the future is obtained.

[0013] (3) A cost function is established that is mainly based on the delay in the calculation process of task offloading forwarding. The bat algorithm is used to solve the problem of minimizing the total cost in the process of relay vehicles participating in the task offloading of request vehicles. The amount of resources consumed in the relay vehicle forwarding process is analyzed to minimize the overall cost while completing the task forwarding.

[0014] To sum up, it can effectively reduce computing costs and computing delays, and improve the vehicle's computing performance and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] Figure 1 The SDN controller is deployed in the cloud to assist the MEC server in offloading and forwarding the requested vehicle computing tasks. Figure 2 A network model diagram of the unloading and forwarding system for requesting vehicle computing tasks provided by the present invention; Figure 3 The figure is a comparison chart of the task offloading cost using the method of the present invention and the other two methods. DETAILED DESCRIPTION

[0017] Step 1: The task offloading vehicle is within the coverage of the roadside unit (RSU) and sends a task offloading request to the current MEC server, along with its real-time status, such as location, speed, direction, etc.: The requesting vehicle that needs to offload the computing task but needs to forward the computing task due to the current MEC server computing resources shortage is marked as i. , the computational task requesting vehicle i can be described as , requesting vehicle i to complete a total of N computing tasks at a certain time, represented as a set , and each computation task ,in Indicates the size of the task, represents the amount of computing resources required to complete the task, represents the sensitivity of requesting vehicle i to the increase of task delay, is the expected delay to complete the task, and the delay cost of completing the task within the expected delay is 0; the relay vehicle that is consistent with the future driving direction of the requesting vehicle and provides unloading task forwarding service for the requesting vehicle is marked as .

[0018] Step 2: After receiving the task offloading request, the current MEC server cannot provide services for the task offloading vehicle due to insufficient computing resources. It obtains the available vehicle information within the coverage area of ​​other RSUs fed back by the SDN controller, that is, the vehicles that are consistent with the future driving direction of the task offloading vehicle: The requesting vehicle can shorten its task execution time by offloading the computing task to the MEC server. Since the connected vehicles within the coverage of the MEC server upload their mobility information to the MEC server, the MEC server can know whether the relay vehicle is available to the requesting vehicle i. Relay vehicle Availability of requested vehicle i. If the requested vehicle i can access ,Right now If it is consistent with the future driving direction of the requested vehicle, .otherwise, Therefore, in Under the condition of Forwarding Tasks .

[0019] Step 3: The current MEC server uses the bat algorithm to calculate the global optimal forwarding path based on the information of the relay vehicle, thereby obtaining the optimal multi-hop chain of the idle MEC server. The necessary condition for the optimal multi-hop chain from the requesting vehicle to the idle MEC server is that the idle MEC server in the future arrival section has the shortest and most stable relay link from the task unloading vehicle to the task unloading vehicle, and sends the relay vehicle node information of the forwarding path to the task unloading vehicle. Obtaining the optimal multi-hop chain depends on the following factors: Assuming that there are at least one or more relay vehicles between the requesting vehicle and the idle MEC server of the road section it will reach in the future, whether these relay vehicles will be selected by the SDN controller as relay vehicles to provide unloading task forwarding services for the requesting vehicle depends on factors such as the number of relay vehicles and the shortest single-hop forwarding time between the requesting vehicle and the relay vehicle. Therefore, the optimal multi-hop chain can be obtained by selecting the optimal relay vehicle calculation time through the SDN controller.

[0020] The SDN controller selects the optimal relay vehicle calculation time analysis process as follows: There are one or more relay vehicles between the requesting vehicle and the idle MEC server at the future destination. Whether these relay vehicles will be selected by the MEC server as the requesting vehicle service is related to their own reward value (Rv, Reward value). The reward value is related to the matching factor (Mf, Matching factor) and correlation factor (Cf, Correlation factor) of the relay vehicle to the requesting vehicle.

[0021] The reward value of each relay vehicle is calculated as follows: (1) in, Represents the matching factor between the relay vehicle and the request vehicle. The relay vehicle chain that provides the request vehicle with a shorter forwarding time has a higher , which is calculated as follows: (2) The total forwarding time for a relay vehicle to complete the unloading task is as follows: (3) in, , , The computing task that requests the vehicle to unload is passed through the relay vehicle The forwarding times are: (1) : The time for requesting vehicle i to access the relay vehicle will be affected by the vehicle's speed and position; (2) :The task volume input data of the task is forwarded to the relay vehicle through the V2V connection The time is given by: (4) (5) in, is the channel bandwidth, is the transmission power between vehicles, is the uplink channel gain in V2V communication mode, is Gaussian white noise, is the channel fading exponent; (3) : Offloaded computing tasks In relay vehicle The waiting time for a relay vehicle Only the unloading tasks of the same requesting vehicle can be transmitted at the same time. Therefore, when other requesting vehicles have unloading tasks to be forwarded, they need to wait for the unloading task forwarding of the previous requesting vehicle to be completed.

[0022] Due to the high density of road vehicles, the requesting vehicle cannot reach the idle MEC server before the expected delay of unloading the task. Represents the remaining time for the requested vehicle to travel to an idle MEC server, which is calculated as follows: (6) in, Represents the distance from the current location of the requesting vehicle to the idle MEC server. Represents the distance the requesting vehicle has traveled. Represents the time it takes the requested vehicle to travel these distances.

[0023] In formula (1) Represents the matching factor between the relay vehicle and the request vehicle, which is the number of vehicles in the multi-hop chain between the relay vehicle and the request vehicle. The fewer the number of vehicles in the multi-hop chain, the higher the relay vehicle has. .

[0024] Therefore, the total time for a relay vehicle to forward the unloading task of the requesting vehicle is: (8) In summary, the higher the reward value of the relay vehicle, the less time it spends providing forwarding services for the unloading tasks of the requesting vehicle.

[0025] In addition, the detailed analysis process of the task forwarding time between vehicles and the MEC server calculation time is as follows: At the same time, the analysis process of the MEC server calculation time is: When the last relay vehicle The task When the requesting vehicle is about to arrive at the coverage area of ​​an idle MEC server in the road section, Establish a V2I connection with an idle MEC server to offload tasks Upload to idle MEC service for processing, task Time to upload to an idle MEC server: (8) in, Represents idle MEC servers and relay vehicles The uplink transmission rate between is expressed as: (9) in, Idle MEC servers are allocated to relay vehicles Bandwidth, and They represent the transmission power from the relay vehicle to the MEC server and the uplink channel gain of the V2I communication mode respectively.

[0026] When the task After uploading to the idle MEC server, the idle MEC server will The calculation time is: (10) in, Is the idle MEC server assigned to the task computing resources.

[0027] According to formulas (8) and (10), we can get the task The overall computation time after forwarding to the idle MEC server coverage is: (11) Combining formulas (7) and (11), we can get the task of requesting vehicle i to unload its The overall delay calculated by forwarding to the idle MEC server in the future arrival section through the multi-hop chain is: (12) Step 4: After the task offloading vehicle receives the relay vehicle node information of the forwarding path, the tasks that cannot be calculated in time on the current MEC server will be unloaded and forwarded to the idle MEC server in the future arrival section through the relay vehicle node according to the forwarding path for calculation.

[0028] Step 5: After receiving the task from the task offloading vehicle, the idle MEC server performs the task offloading calculation and saves the result on the server. When waiting for the task offloading vehicle to arrive, it returns the result to the task offloading vehicle.

[0029] Vehicular self-organizing networks (VANETs) have become a promising technology for providing safety and comfort services to vehicle users. The three commonly discussed communication modes in VANETs are vehicle-to-infrastructure (V2I), vehicle-to-vehicle (V2V), and V2I2V. In VANETs, ​​vehicles are equipped with onboard units (OBUS) based on dedicated short-range communications (DSRC). Through OBUS, vehicles can use IEEE802.11p communication technology to communicate between them and with fixed roadside units (RSUs), which can provide information support and connections for passing vehicles, including traffic information and safety warnings. The dynamic topology changes caused by the mobility of vehicles make unloading transmission complicated. In the MEC cloud-based vehicle network, in order to improve the efficiency of task completion, an optimal task offloading solution with MEC server selection and communication management must be provided. The existing technology does not take into account the problem of connection interruption during vehicle-to-vehicle or vehicle-to-infrastructure communication due to the different future driving directions of the requesting vehicle from other vehicles. Once the connection between the requesting vehicle and the MEC server is interrupted during the computing task offloading process, the network delay and computing cost will increase.

[0030] Therefore, the present invention proposes a solution for forwarding unloading tasks by deploying an SDN controller in the cloud to forward the unloading tasks of the auxiliary relay vehicle to the requesting vehicle to the idle MEC server of the road section where the auxiliary relay vehicle will arrive in the future. The calculation time for forwarding unloading tasks is effectively reduced. The present invention can be applied to the existing field of vehicle-mounted communications, and the unloading tasks of the auxiliary relay vehicle to the requesting vehicle can be forwarded through the flexibility of SDN software deployed in the cloud. The present invention can effectively reduce the forwarding calculation cost of vehicle-mounted communication tasks, thereby improving vehicle-mounted communication efficiency and driving safety.

[0031] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0032] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

[0033] Glossary MEC: Multi-access edge computing is a technology that places computing and storage resources at the edge of the network. By deploying computing resources at the edge of the network, MEC can achieve lower latency and higher bandwidth, thereby supporting real-time data processing, application execution, and service provision.

[0034] Software Defined Networking (SDN): A new type of network architecture that separates the network control plane and the data plane to achieve centralized control and management of the network, thereby improving the programmability, flexibility, and automation of the network.

Claims

1. A vehicle networking task offloading and forwarding algorithm, characterized in that: The following steps are involved: (1) Obtain the real-time status of all vehicles within the coverage area of ​​the roadside unit (RSU) through the SDN controller deployed in the cloud; (2) Feedback the available optimal relay vehicle information to the current MEC server; (3) The MEC server uses the bat algorithm to calculate the global optimal forwarding path and sends the forwarding vehicle node information to the task unloading vehicle; (4) The task offloading vehicle will forward the tasks that cannot obtain timely computing services from the current MEC server through the forwarding vehicle node to the idle MEC server in the future arrival section for computing; (5) When the task unloading vehicle arrives at this server, the calculation results are obtained from the server.

2. According to claim 1, a vehicle networking task offloading and forwarding algorithm is characterized in that Said Step (1): The task offloading vehicle is within the coverage of the roadside unit RSU and sends a task offloading request to the current MEC server, along with its real-time status, such as location, speed, direction, etc.: The requesting vehicle that needs to offload the computing task but needs to forward the computing task due to the current MEC server computing resources shortage is marked as i. , the computational task of requesting vehicle i can be described as , requesting vehicle i to complete a total of N computing tasks at a certain time, represented as a set , and each computation task ,in Indicates the size of the task, represents the amount of computing resources required to complete the task, represents the sensitivity of requesting vehicle i to the increase of task delay, is the expected delay to complete the task, and the delay cost of completing the task within the expected delay is 0; the relay vehicle that is consistent with the future driving direction of the requesting vehicle and provides unloading task forwarding service for the requesting vehicle is marked as ; Step 2: After receiving the task offloading request, the current MEC server cannot provide services for the task offloading vehicle due to insufficient computing resources. It obtains the available vehicle information within the coverage area of ​​other RSUs fed back by the SDN controller, that is, the vehicle with the same future driving direction as the task offloading vehicle: The requesting vehicle can shorten its task execution time by offloading the computing task to the MEC server. Since the connected vehicles within the coverage of the MEC server will upload their mobility information to the MEC server, the MEC server can know whether the relay vehicle is available to the requesting vehicle i. Relay vehicle Availability of the requesting vehicle i; if the requesting vehicle i can access ,Right now If it is consistent with the future driving direction of the requested vehicle, ; otherwise, Therefore, in Under the condition of Forwarding Tasks ; Step 3: The current MEC server uses the bat algorithm to calculate the global optimal forwarding path based on the information of the relay vehicle, thereby obtaining the optimal multi-hop chain of the idle MEC server. The necessary condition for the optimal multi-hop chain from the requesting vehicle to the idle MEC server is that the idle MEC server in the future arrival section has the shortest and most stable relay link from the task unloading vehicle to the task unloading vehicle, and sends the relay vehicle node information of the forwarding path to the task unloading vehicle. To obtain the optimal multi-hop chain depends on the following factors: Assuming that there are at least one or more relay vehicles between the requesting vehicle and the idle MEC server of the road section it will reach in the future, whether these relay vehicles will be selected by the SDN controller as relay vehicles to provide unloading task forwarding services for the requesting vehicle depends on factors such as the number of relay vehicles and the shortest single-hop forwarding time between the requesting vehicle and the relay vehicle. Therefore, the optimal multi-hop chain is obtained by selecting the optimal relay vehicle and calculating the time through the SDN controller; The SDN controller selects the optimal relay vehicle calculation time analysis process as follows: There are one or more relay vehicles between the requesting vehicle and the idle MEC server at the future destination. Whether these relay vehicles will be selected by the MEC server as the requesting vehicle service is related to their own reward value, which is related to the matching factor and correlation factor of the relay vehicle to the requesting vehicle. The reward value of each relay vehicle is calculated as follows: （1） in, Represents the matching factor between the relay vehicle and the request vehicle. The relay vehicle chain that provides the request vehicle with a shorter forwarding time has a higher , which is calculated as follows: （2） The total forwarding time for a relay vehicle to complete the unloading task is as follows: （3） in, , , The computing task that requests the vehicle to unload is passed through the relay vehicle The forwarding times are: : Request vehicle i to access the relay vehicle The time will be affected by the vehicle speed and position; :The task volume input data of the task is forwarded to the relay vehicle through the V2V connection The time is given by: （4） （5） in, is the channel bandwidth, is the transmission power between vehicles, is the uplink channel gain in V2V communication mode, is Gaussian white noise, is the channel fading exponent; : The waiting time of the offloaded computing task on the relay vehicle; a relay vehicle Only the unloading tasks of the same requesting vehicle can be transmitted at the same time. Therefore, when other requesting vehicles have unloading tasks to be forwarded, they need to wait for the unloading tasks of the previous requesting vehicle to be forwarded. Due to the high density of road vehicles, the requesting vehicle cannot reach the idle MEC server before the expected delay of unloading the task. Represents the remaining time for the requested vehicle to travel to an idle MEC server, which is calculated as follows: （6） in, Represents the distance from the current location of the requesting vehicle to the idle MEC server. Represents the distance the requesting vehicle has traveled. represents the time it takes the requested vehicle to travel those distances; In formula (5) Represents the matching factor between the relay vehicle and the request vehicle, which is the number of vehicles in the multi-hop chain between the relay vehicle and the request vehicle. The fewer the number of vehicles in the multi-hop chain, the higher the relay vehicle has. ; Therefore, the total time for a relay vehicle to forward the unloading task of the requesting vehicle is: （7） In summary, the higher the reward value of the relay vehicle, the less time it spends providing forwarding services for the unloading tasks of the requesting vehicle; In addition, the detailed analysis process of the task forwarding time between vehicles and the MEC server calculation time is as follows: At the same time, the analysis process of the MEC server calculation time is: When the last relay vehicle The task When the requesting vehicle is about to arrive at the coverage area of ​​an idle MEC server in the road section, Establish a V2I connection with an idle MEC server to offload tasks Upload to idle MEC service for processing, task Time to upload to an idle MEC server: （8） in, Represents idle MEC servers and relay vehicles The uplink transmission rate between is expressed as: （9） in, Idle MEC servers are allocated to relay vehicles bandwidth, and They represent the transmission power from the relay vehicle to the MEC server and the uplink channel gain of the V2I communication mode respectively; When the task After uploading to the idle MEC server, the idle MEC server will The calculation time is: （10） in, Is the idle MEC server assigned to the task computing resources; According to formulas (8) and (10), we can get the task The overall computation time after forwarding to the idle MEC server coverage is: （11） Combining formulas (7) and (11), we can get the task of requesting vehicle i to unload its The overall delay calculated by forwarding to the idle MEC server in the future arrival section through the multi-hop chain is: （12） Step 4: After the task offloading vehicle receives the relay vehicle node information of the forwarding path, the task that cannot be calculated on the current MEC server in time is offloaded and forwarded to the idle MEC server in the future arrival section through the relay vehicle node according to the forwarding path for calculation; Step 5: After receiving the task from the task offloading vehicle, the idle MEC server performs the task offloading calculation and saves the result on the server. When waiting for the task offloading vehicle to arrive, it returns the result to the task offloading vehicle.

Images