An Optimization Method for Task Offloading in Vehicular Ad Hoc Networks Based on Mobile Sensing

Abstract

The present invention discloses an optimization method for task offloading in a vehicle-to-everything (V2X) network based on mobile sensing. The method includes: (1) initializing system parameters, including parameters related to roadside units, vehicles, tasks, and initial algorithm parameters; (2) generating an offloading ratio allocation vector for computing tasks, where each element in the vector represents the ratio of tasks received by the corresponding vehicle; (3) optimizing the energy consumption of the system by optimizing the task offloading ratio, and finding the optimal solution for task offloading based on the proposed algorithm; (4) performing task offloading according to the allocation strategy corresponding to the optimal solution. The present invention can quickly and effectively optimize the task offloading calculation, effectively reduce the latency, and lower the cost of offloading calculation.

Description

Technical Field

[0001] The present invention relates to the field of 5G communication, and particularly to an optimization method for vehicle network task offloading based on mobile perception. Background Art

[0002] With the rapid development of in-vehicle technologies and the continuous growth of computing demands for in-vehicle data transmission services, in-vehicle terminals with limited computing capabilities face huge challenges, especially in terms of their computing resources. The big data technology in the 5G era requires higher efficiency and intelligence in data processing. To cope with the explosive computing demands of automotive terminals and the limited processing capabilities of vehicle terminals, cloud-based vehicle networks are widely regarded as a new paradigm for improving service efficiency. However, the remote transmission of task files between vehicle terminals and cloud servers and the fluctuations of wireless channels may cause considerable delays, affecting communication quality and reducing user experience. In addition, the dynamic and uncertain vehicle environment brings additional challenges to maintain a long-term satisfactory user experience.

[0003] To solve the above problems, mobile edge computing (MEC) is an alternative in vehicle networks, providing high-reliability and low-latency services for vehicles and users. MEC can make up for the deficiency of low communication latency in cloud computing by deploying servers near the user edge. On the one hand, the MEC server has computing resources to meet the offloading demands of vehicles within the coverage of related roadside units (RSUs) or adjacent RSUs. On the other hand, once the MEC server receives a computing request from a vehicle terminal, it can provide a quick interactive response. In addition, by offloading intensive computing workloads to the nearby MEC server, the latency for vehicle-to-vehicle (V2V) or vehicle-to-infrastructure (V2I) execution can be significantly reduced. The communication service quality can be improved, and the burden on the task processor of the vehicle terminal can also be greatly reduced.

[0004] However, in traditional offloading schemes, the mobility of vehicles is rarely considered, which will result in the situation that vehicles cannot connect to the base station, leading to service discontinuity and instability of computing resources, which pose new challenges to offloading decisions. In view of this situation, the present invention proposes an optimization algorithm for vehicle network task offloading based on mobile perception, enabling vehicles with computing capabilities to provide VEC services in the vehicle fog network, thus solving the problem of high vehicle mobility. Summary of the Invention

[0005] Object of the Invention. The object of the present invention is to provide an optimization algorithm for vehicle network task offloading based on mobile perception for the resource allocation problem of edge computing in the vehicle network environment, quickly and effectively optimizing the computing resource allocation to minimize the overall energy consumption of the system.

[0006] Technical solution: To achieve the above-mentioned invention objective, the present invention adopts the following technical solution:

[0007] An optimization method for task offloading in a vehicle-to-everything (V2X) network based on mobile sensing, comprising the following steps:

[0008] (1) Initialize system parameters, where the parameters include the coverage radius r of the roadside unit, the vertical distance m from the unit center to the road, the number k of vehicles within the roadside unit, the transmission bandwidth between the roadside unit and the vehicle, the length d of each time slot, the input data volume I of the task, the computing resources C required for the task, and the location information of the vehicle.

[0009] (2) Randomly generate an offloading ratio allocation vector β = [β1, β2,..., β k ,...β K , where each element satisfies 0 ≤ β k ≤ 1 and k represents the number of vehicles within the roadside unit.

[0010] (3) List the Lagrangian equation for the objective function of system energy consumption and write the corresponding Lagrangian dual problem.

[0011] (4) Obtain the Lagrange multipliers λ k , μ k , and θ according to the subgradient method.

[0012] (5) Divide each element β k in the vector by the sum of all elements in the row to satisfy the constraint conditions:

[0013] 0 ≤ β k ≤ 1 and

[0014] (6) According to the Lagrange multipliers, obtain the task allocation and offloading vector as:

[0015]

[0016] (7) Iterate steps (3)-(6) n times, and finally allocate the corresponding task offloading ratio according to the vector .

[0017] Beneficial effects: Compared with the prior art, the present invention proposes an optimization algorithm for task offloading in a vehicle-to-everything (V2X) network based on mobile sensing based on a novel offloading scheme, which can effectively improve the utilization rate of vehicle computing resources and minimize the energy consumption of the system. Description of the drawings

[0018] Figure 1 is the scenario model diagram of the method of the present invention:

[0019] Figure 2 This is a specific flow chart of the optimization method for task offloading computing energy consumption in the Internet of Vehicles:

[0020] Figure 3 It is a comparison diagram of energy consumption functions using the algorithm of the present invention and other algorithms. DETAILED DESCRIPTION

[0021] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, modifications of various equivalent forms of the present invention made by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0022] The embodiment of the present invention analyzes in detail the optimization method of task offloading computing energy consumption in the Internet of Vehicles in combination with specific scenarios. The following detailed analysis is made of the settings and parameter settings of the following scenarios.

[0023] 1. Road and vehicle parameters

[0024] We consider a one-way straight road with RSUs deployed along one side of the road. Assume that the coverage radius of the roadside base station is about 200 meters. The radius of the microcell is represented by r, and the vertical distance from the cell center to the road is represented by m. The total distance s traveled by vehicles within the coverage area can be calculated by knowing the RSU coverage radius and the vertical distance from the cell center to the road m. represents the set of vehicles. For convenience, we introduce a discrete time system and divide the time T of vehicle k into i equal time periods. d represents the length of each time slot. The speed of vehicle k can be written as v k , k={1,2,...,k,...,K} obeys uniform distribution, so the distance traveled by vehicle k in each time slot is v k d.

[0025] 2. Establishment of system model

[0026] In the patent, we proposed a new offloading computing scheme. First, we divided the vehicle's operating time into multiple identical time slots. Within each time slot, we assumed that the distance between the vehicle and the base station is constant. The RSU distributes a task to multiple vehicles according to a ratio. By optimizing the task offloading ratio, the system's energy consumption is optimized. The optimization of system energy consumption is converted into a convex optimization problem for solution. Then, a Lagrangian and its dual method based on a convex algorithm is used to solve this problem. A low-complexity algorithm is proposed to optimize the task offloading ratio by continuously iterating the Lagrangian multiplier until it converges. Finally, simulation results verify the high efficiency of the proposed scheme. Compared with other algorithms, the algorithm proposed in this article not only improves the utilization of vehicle computing resources, but also minimizes the overall energy consumption of the system.

[0027] 3. System energy consumption

[0028]

[0029]

[0030]

[0031] in, represents the task transmission time assigned to vehicle k, β = [β1, β2, ..., β k , ...β K ] represents the offload ratio allocation vector of the computing task, and C refers to the computing resources required for the computing task. The first half of the expression is the system's transmission energy consumption, and the second half is the system's local computing energy consumption.

[0032] like Figure 2 As shown, a novel offloading solution disclosed in an embodiment of the present invention proposes a vehicle network task offloading optimization method based on mobile perception, including the following steps:

[0033] (1) Initialize system parameters, including the coverage radius r of the roadside unit, the vertical distance m from the unit center to the road, the number of vehicles in the roadside unit k, the transmission bandwidth between the roadside unit and the vehicle, the length of each time slot d, the input data volume I of the task, the computing resources C required for the task, and the location information of the vehicle.

[0034] (2) Randomly generate a computing task offloading ratio distribution vector β = [β1, β2, ..., β k , ...β K ], each element satisfies 0≤β k ≤1 and k represents the number of vehicles in the roadside unit.

[0035] (3) List the Lagrangian equation for the objective function of the system energy consumption and write the corresponding Lagrangian dual problem.

[0036] (4) Obtain the Lagrange multipliers λ k , μ k , and θ according to the subgradient method.

[0037] (5) Divide each element β in the vector k by the sum of all elements in the row to satisfy the constraint conditions:

[0038] 0 ≤ β k ≤ 1 and

[0039] (6) Obtain the task allocation and offloading vector as:

[0040]

[0041] (7) Iterate steps (3)-(6) for n times, and finally allocate the corresponding task offloading ratio according to the vector .

[0042] Figure 3 A detailed comparison diagram of the consumption function using the algorithm of the present invention and other algorithms is provided. To verify the advantages of the method of the present invention over the prior art, the following simulation parameters are set in the present invention: the coverage radius of the cell is 200m, the vertical distance from the center of the coverage unit to the road is 80m, the RSU bandwidth allocated to each vehicle is 1Mhz, the transmission powers of the RSU and the vehicle are both 0.1w, the input data volume of the task is 100Mb, the length of each time slot is 0.1s, the number of vehicles is 8, and the driving speed is between (0, 70) km / h. It can be seen from the figure that the genetic algorithm converges around the 35th generation, and its optimization result is likely to fall into a local optimal solution. The beetle antennae search algorithm converges around the 30th generation. Since its initial value and convergence value are both large, more system functions need to be consumed. The algorithm proposed in this patent has a small computational amount and short time consumption.

Claims

1. A method for optimizing task offloading in a vehicle-to-everything network based on mobile sensing, characterized in that The method includes the following steps: (1) Initialize system parameters, where the parameters include the coverage radius r of the roadside unit, the vertical distance m from the unit center to the road, the number k of vehicles within the roadside unit, the transmission bandwidth between the roadside unit and the vehicles, the length d of each time slot, the input data volume I of the task, the computing resources C required for the task, and the location information of the vehicles; (2) Randomly generate a computing task offloading ratio distribution vector β = [β1, β2, ..., β k ,...β K ], each element satisfies 0≤β k ≤1 and k represents the number of vehicles in the roadside unit; (3) List the Lagrangian equation for the objective function of the system energy consumption and write the corresponding Lagrangian dual problem; (4) Obtain the Lagrange multiplier λ using the subgradient method k , μ k ,θ; (5) Each element β in the vector k Divide by the sum of all elements in the row to satisfy the constraint: 0≤β k ≤1 and (6) Obtain the task allocation and offloading vector according to the Lagrange multiplier which is (7) Steps (3)-(6) are iterated n times, and finally, according to the vector The corresponding task offloading ratio is allocated.

2. The method for optimizing task offloading in a vehicle networking based on mobile sensing according to claim 1, wherein In step (3), the objective function is expressed as: where β k represents the offloading ratio distribution vector of computing tasks, Represents the time required to offload the task.

Images