Task unloading and resource allocation device for air-space-ground cooperation Internet of Vehicles

Abstract

The invention discloses a space-air-ground cooperation car networking task unloading and resource allocation device, and belongs to the technical field of calculation, reckoning or counting. According to the device, the problems of space-air-ground cooperation Internet of Vehicles computing task unloading and resource allocation are jointly considered, and a strategy for solving the problem of limited computing resources of the vehicle user terminal is provided, so that the requirement of the Internet of Vehicles for low time delay is met. According to the device, firstly, a space-air-ground cooperation vehicle networking calculation unloading model is constructed, and a calculation task unloading and processing mode of a vehicle user terminal is described; secondly, defining average time delay and energy consumption functions and constraint conditions of all vehicle user terminals for processing calculation tasks; and finally, providing an air-space-ground collaborative car networking mobile edge computing communication and computing resource allocation algorithm to jointly solve a task unloading decision sub-problem and a computing resource allocation sub-problem, and a simulation result shows that the algorithm has relatively good performance.

Description

Technical Field

[0001] The present invention relates to communication technologies, and specifically discloses an air-space-ground cooperative vehicle networking task offloading and resource allocation device, belonging to the technical field of computing, reckoning or counting. Background Art

[0002] With the development of the sixth-generation mobile communication, space-air-ground integration, as an emerging network technology, has been widely applied to various communication fields such as intelligent transportation, aerospace, and emergency disasters. The architecture of the space-air-ground integrated network (SAGIN) mainly consists of a space-based network, an air-based network, and a ground network. Among them, satellites in the space-based network and unmanned aerial vehicles in the air-based network can expand the communication range and achieve wide-area coverage. At the same time, SAGIN can provide diverse remote wireless access services to meet the device requirements of multiple users. Therefore, SAGIN meets the communication needs of modern people, and it is of great significance to study it.

[0003] The Internet of Vehicles (IoV) can collect, share, and process vehicle information, provide road status information, and ensure the safe driving of vehicles. However, the computing resources of vehicles in IoV are limited. If all the computing tasks generated by vehicles are left to be executed locally, it is difficult to meet the low-latency requirements of IoV. The emergence of mobile edge computing (MEC) enables vehicle users to offload local computing-intensive and latency-sensitive tasks to edge servers for processing, thereby alleviating the problem of insufficient local resources of vehicles and improving the service quality of vehicle users. On this basis, reasonably formulating offloading decisions and allocating computing resources can reduce system energy consumption, increase system capacity, reduce task processing latency, and improve resource utilization. At the same time, the addition of SAGIN expands the servers to which vehicle computing tasks in IoV can be offloaded from the ground network to the air-based and space-based networks, further providing vehicle users with multi-level and rich computing resources. Therefore, resource allocation in the space-air-ground integrated vehicular network (SAGVN) has become a current research hotspot. Summary of the Invention

[0004] Aiming at the problem of limited computing power of vehicle user terminals in the space-air-ground cooperative vehicle networking, a task offloading and resource allocation device for space-air-ground cooperative vehicle networking is designed. First, a computing offloading model for space-air-ground cooperative vehicle networking is constructed to describe the computing task offloading and processing methods of vehicle user terminals. Secondly, the average delay and energy consumption functions and constraint conditions for all vehicle user terminals to process computing tasks are defined. Finally, a space-air-ground cooperative vehicle networking mobile edge computing communication and computing resource allocation algorithm is used to jointly solve the task offloading decision sub-problem and the computing resource allocation sub-problem. The simulation results show that the proposed algorithm has good performance.

[0005] The task offloading and resource allocation device for space-air-ground cooperative vehicle networking of the present invention includes the following three steps:

[0006] 1) Construct a computing offloading model for space-air-ground cooperative vehicle networking. The system consists of three layers of networks, namely, the ground-based network, the air-based network, and the space-based network. Among them, the ground-based network contains a Road Side Unit (RSU) installed on one side of the road. The radius of its coverage area is R, which is responsible for communicating with vehicle user terminals. The RSU is equipped with an MEC server connected by optical fiber. The MEC server has certain storage and computing capabilities, denoted as MS. The ground-based network also contains N vehicle user terminals, denoted as the set VUE = {VUE1,..., VUE i ,..., VUE N}, where i ∈ [1, N]. Among them, VUE i represents the i-th vehicle user terminal. All vehicles are traveling in the same direction on the road, and vehicles can communicate with each other. The radius of the coverage area of each vehicle is r. When VUE i is within the communication range of the RSU, the task can be forwarded to the MEC server through the RSU for execution. The air-based network contains a hovering unmanned aerial vehicle (UAV) equipped with a server denoted as UAS, which has certain storage and computing capabilities. The communication range of the UAV network is the same as that of the RSU. The space-based network contains a Low Earth Orbit (LEO) satellite and a ground station. The LEO is equipped with a server denoted as SAS, which has strong computing capabilities. The ground station is responsible for communicating with the LEO, and the communication range of the satellite can cover the entire ground network.

[0007] The present invention only considers the uplink of signal transmission, and there are K sub-channels in the system, denoted as the set C = {C1,..., C j ,..., C K}, where \(j\in[1, K]\), allowing channel multiplexing for V2V offloading links, V2I offloading links, UAV offloading links, and ground station offloading links within the communication range of the RSU, and each channel is allowed to be multiplexed by at most two VUEs. In the terrestrial network, the V2I offloading link and the V2V offloading link use flat fading channels. In the aerial network, the UAV offloading link uses a line-of-sight link. In the space-based network, the ground station offloading link uses a flat fading channel, and the satellite offloading link uses a Weibull channel.

[0008] Each vehicle user terminal has a certain computing ability, and the task of the VUE i can be executed locally. However, when the local computing ability of the VUE i is insufficient, it can be offloaded to other vehicles or servers for processing. In the space-based network, the VUE i can offload the task to the SAS for execution through the ground station. The VUE i also has different time delays and energy consumptions when offloading tasks to different servers. It is assumed in the present invention that the task of the vehicle user terminal VUE i is an indivisible unit, that is, the task can only be offloaded in one offloading method. In the present invention, only the upload delay of the task and the computing delay of the task are considered, and the total delay of the task consists of the upload delay and the computing delay of the task.

[0009] The delay of the VUE i task calculated locally is expressed as:

[0010]

[0011] where represents the binary offloading decision. When , the task of the VUE i is executed locally. When , the task of the VUE i selects other methods for offloading. \(d\) i represents the data volume size of the VUE i task, and \(\omega\) i represents the number of CPU cycles per bit of data of the VUE i . represents the CPU frequency of the VUE i .

[0012] The energy consumption of the VUE i task calculated locally is expressed as:

[0013]

[0014] where \(\kappa\) v represents the VUE iThe effective open capacitance coefficient of the chip structure.

[0015] VUE i Using C j The rate at which the task is offloaded to the MS can be expressed by the Shannon formula as:

[0016]

[0017] Where B j represents the sub-channel bandwidth of C j P i , P m represent the transmit powers of VUE i and VUE m respectively, and σ 2 represents the noise power of the channel. D represents the number of vehicles within the communication range of the RSU, and λ i，j represents the indicator variable for allocating C j to VUE i . represents the interference of the V2I channel where VUE m reuses VUE i , where λ m，j represents the indicator variable for allocating C j to VUE m . represents the channel gain between the MS and VUE m . represents the channel gain between the MS and VUE i , and its expression is where β i is the shadow fading, is the distance between VUE i and the MS, α is the path loss exponent, and h i is the small-scale fading.

[0018] VUE i The delay for VUE

[0019]

[0020] where represents the binary offloading decision, and when the task of VUE i is executed at the MS.

[0021] VUE i The delay when the task of VUE

[0022]

[0023] where Indicates the CPU frequency of the MS.

[0024] Therefore, the total latency of the VUE i task executed on the MS is:

[0025]

[0026] VUE i The offloading energy consumption of the task offloaded to the MS is expressed as:

[0027]

[0028] VUE i The energy consumption of the task of VUE

[0029]

[0030] where p r Indicates the computing power of the MS.

[0031] Therefore, the total energy consumption of the VUE i task executed on the MS is:

[0032]

[0033] VUE i Using C j The rate at which the task is offloaded to the surrounding vehicle VUE p can be expressed by the Shannon formula as:

[0034]

[0035] where, Indicates the interference of the V2V channel reused by the VUE a where λ i Indicates the indication variable assigned to the VUE a，j when C j is assigned to the VUE a P a Indicates the transmission power of the VUE a . Indicates the channel gain between the VUE p and the VUE a . Indicates the channel gain between the VUE p and the VUE i , and its expression is where is the distance between the VUE i and the VUE p .

[0036] VUEi The latency for offloading a task to a surrounding vehicle VUE p is:

[0037]

[0038] where, represents the binary offloading decision. When the task of VUE i is executed on VUE p .

[0039] VUE i The latency for the task of VUE p when calculating on a surrounding vehicle VUE

[0040]

[0041] where, represents the CPU frequency of VUE p .

[0042] Therefore, the total latency for the task of VUE i when executed on a surrounding vehicle VUE p is:

[0043]

[0044] where, O i represents the number of vehicles within the communication range of VUE i .

[0045] VUE i The energy consumption for offloading a task to a surrounding vehicle VUE p is:

[0046]

[0047] VUE i The energy consumption for the task of VUE p when calculating on a surrounding vehicle VUE

[0048]

[0049] Therefore, the total energy consumption for the task of VUE i when executed on a surrounding vehicle VUE p is:

[0050]

[0051] VUE i The rate at which VUE j uses C to offload a task to the UAS can be expressed by the Shannon formula as:

[0052]

[0053] Among them, represents VUE q Reusing the interference of the VUE i in the UAV channel, where λ q，j represents assigning C j to the VUE q indicator variable, P q represents the transmission power of the VUE q . represents the channel gain between the UAS and the VUE q . represents the channel gain between the UAS and the VUE i , and its expression is where θ0 represents the channel power gain at the reference distance l0 = 1m, h t and h r respectively represent the antenna power gains of the VUE i and the UAV, w represents the wavelength, where is the distance between the VUE i and the UAS.

[0054] The latency for the VUE i to offload the task to the UAS is:

[0055]

[0056] Among them, represents the binary offloading decision, when the task of the VUE i is executed on the UAS.

[0057] The latency for the task of the VUE i to be computed on the UAS is:

[0058]

[0059] Among them, represents the CPU frequency of the UAS.

[0060] Therefore, the total latency for the task of the VUE i to be executed on the UAS is:

[0061]

[0062] The energy consumption for the VUE i to offload the task to the UAS is:

[0063]

[0064] VUE i The energy consumption of the task calculated by the UAS is expressed as:

[0065]

[0066] where p u represents the computing power of the UAS.

[0067] Therefore, the total energy consumption of the VUE i task executed by the UAS is:

[0068]

[0069] VUE i The rate at which VUE uses C j to offload the task to the ground station can be expressed by the Shannon formula:

[0070]

[0071] where represents the interference of the ground station channel reused by VUE z reusing VUE i Among them, λ z，j represents the C j assigned to VUE z indicator variable, P z respectively represent the transmission power of VUE z . represents the channel gain between the ground station and VUE i . represents the channel gain between the ground station and VUE i , and its expression is where the distance between the ground station and VUE i .

[0072] The task offloading rate from the ground station to the SAS can be expressed by the Shannon formula as:

[0073]

[0074] where P e represents the transmission power of the ground station, g e，s represents the channel gain between the ground station and the SAS, which can be expressed as

[0075]

[0076] where l e，s represents the distance between the ground station and the SAS, λ represents the wavelength of the signal, G ea and G sarepresent the gains at the ground station and SAS antenna respectively, F rain Indicates rainfall decline.

[0077] VUE i The latency of offloading the task to SAS is expressed as:

[0078]

[0079] in, Indicates a binary uninstall decision when VUE i The task is performed in SAS.

[0080] VUE i The delay of the task in SAS calculation is expressed as:

[0081]

[0082] in, Indicates the CPU frequency of SAS.

[0083] Therefore, VUE i The total delay of the task execution in SAS is:

[0084]

[0085] VUE i The energy consumption of offloading tasks to SAS is expressed as:

[0086]

[0087] VUE i The energy consumption of the task in SAS is expressed as:

[0088]

[0089] Among them, p s Indicates the computing power of SAS.

[0090] Therefore, VUE i The total energy consumption of the task executed in SAS is:

[0091]

[0092] 2) Define the average delay and energy consumption function and constraints for all vehicle user terminals to process computing tasks. The present invention aims to minimize the average delay and energy consumption of all vehicle user terminals to process computing tasks, which can be expressed as Wherein, μ represents the delay weight factor.

[0093] The constraints on the unloading variables are The CPU cycle frequency constraint of VUE is where f v，max represents the maximum CPU frequency of VUE; the CPU cycle frequency constraint of MS is where f r，max represents the maximum CPU frequency of MS; the CPU cycle frequency constraint of UAS is where f u，max represents the maximum CPU frequency of UAS; the CPU cycle frequency constraint of SAS is where f s，max represents the maximum CPU frequency of SAS; processing VUE i The total delay constraint of tasks is where represents the maximum tolerable delay for executing VUE i tasks; the total energy consumption constraint of processing VUE i tasks is where represents the maximum energy consumption of VUE i task execution; the constraint of the channel multiplexing indication variable is λ m，j ∈ {0, 1}, m ∈ D, j ∈ K.

[0094] 3) Sequentially select the vehicle farthest from the RSU communication range for channel multiplexing; set the CPU cycle frequency to a fixed value, and use an improved genetic algorithm to solve the task offloading decision problem; through the obtained task offloading decision, use an improved particle swarm optimization algorithm to solve the computing resource allocation problem to meet the requirements of low latency in IoV. Finally, compare the proposed algorithm with the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing - particle swarm optimization algorithm through simulation experiments. It can be seen from the simulation results that the proposed algorithm can effectively reduce the total delay and total energy consumption of VUE task execution.

[0095] The present invention adopts the above technical solutions and has the following beneficial effects: The present invention comprehensively considers the computing task offloading and processing methods of vehicle user terminals in the space - air - ground collaborative vehicle - to - everything network, aims at the optimization problem of minimizing the average delay and energy consumption of all vehicle user terminals for processing computing tasks, and takes task offloading decision, CPU cycle frequency, delay, energy consumption, and channel multiplexing as constraints to reduce the processing delay and energy consumption of vehicle user terminal tasks, meet the requirements of low latency in the vehicle - to - everything network, and achieve high - efficiency utilization of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings, where:

[0097] Figure 1It is an air-ground-space collaborative vehicle networking computing offloading model;

[0098] Figure 2 It is a schematic diagram of the comparison between the data volume and the total task processing delay;

[0099] Figure 3 It is a schematic diagram of the comparison between the data volume and the total task processing energy consumption. Specific implementation manners

[0100] The following further describes in detail the specific implementation manners of the present invention with reference to the accompanying drawings.

[0101] Figure 1 An air-ground-space collaborative vehicle networking computing offloading model. The system consists of three layers of networks, namely a ground-based network, an air-based network, and a space-based network. Among them, the ground-based network contains an RSU, which is set on one side of the road, and the radius of its coverage is R. It is responsible for communicating with vehicle user terminals. The RSU is equipped with an MEC server, which is connected by optical fiber. The MEC server has certain storage and computing capabilities, denoted as MS. The ground-based network also contains N vehicle user terminals, denoted as the set VUE = {VUE1,..., VUE i ,..., VUE N}, i ∈ [1, N], where VUE i represents the i-th vehicle user terminal. All vehicles drive in the same direction on the road, and vehicles can communicate with each other. The radius of the coverage of each vehicle is r. When VUE i is within the communication range of the RSU, the task can be forwarded to the MEC server through the RSU for execution. The air-based network contains a hovering unmanned aerial vehicle (UAV), which is equipped with a server denoted as UAS and has certain storage and computing capabilities. The communication range of the UAV network is the same as that of the RSU. The space-based network contains a low-earth orbit (LEO) satellite and a ground station. The LEO satellite is equipped with a server denoted as SAS and has strong computing capabilities; the ground station is responsible for communicating with the LEO satellite, and the communication range of the satellite can cover the entire ground network.

[0102] To optimize the performance of the air-ground-space collaborative vehicle network, the present invention aims to minimize the average delay and energy consumption of vehicle user terminals, expressed as

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112] λ m，j ∈ {0, 1}, m ∈ D, j ∈ K (34j)

[0113] Constraints (34b)-(34c) indicate that the task can only be executed locally or offloaded to a server for execution; constraints (34d)-(34g) indicate that the CPU frequency during task execution and the total CPU cycle frequency for executing all tasks cannot exceed the maximum CPU cycle frequency; constraint (34h) indicates that the total time delay during task execution cannot exceed the maximum tolerable time delay; constraint (34i) indicates that the total energy consumption during task execution cannot exceed the maximum energy consumption; constraint (34j) indicates that each offloading link is only allowed to multiplex one subchannel.

[0114] This optimization problem can be divided into an offloading decision sub-problem and a computing resource allocation sub-problem, where the offloading decision sub-problem is expressed as

[0115]

[0116]

[0117]

[0118]

[0119]

[0120] The computing resource allocation sub-problem is expressed as

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127]

[0128] Through the communication and computing resource allocation algorithm for the vehicle Internet of Things with space-air-ground cooperation, the above two sub-problems can be jointly solved, specifically including the following steps:

[0129] (1) Calculate the channel gain according to parameters such as vehicle location, task data volume, shadow fading, and small-scale fading;

[0130] (2) Define the set of unallocated vehicles and the number of unallocated vehicles UV in the set, the set of unallocated channels and the number of unallocated channels UC in the set;

[0131] (3) Select the two VUEs with the farthest distance within the communication range of the RSU in the set to reuse the same channel, move the two VUEs that reuse the channel out of the set and move the channel they reuse out of the set

[0132] (4) Judge whether UV > UC is satisfied. If it is satisfied, return to (3). If it is not satisfied, enter (5);

[0133] (5) Allocate independent channels for the remaining VUEs in the set ;

[0134] (6) Let the initial CPU cycle frequency f0 be the optimal CPU cycle frequency;

[0135] (7) Use the improved genetic algorithm and the current optimal CPU cycle frequency to solve the optimal offloading decision;

[0136] (8) Use the improved particle swarm optimization algorithm and the current optimal offloading decision to solve the optimal CPU cycle frequency.

[0137] Figure 2 It is a schematic diagram of the comparison between the data volume and the total task processing delay. This schematic diagram shows the variation of the total delay of the proposed algorithm, the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm for processing tasks with the data volume. The abscissa in the figure represents the data volume, and the ordinate represents the total task processing delay. The red line, blue line, yellow line, and purple line represent the proposed algorithm, the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm respectively. It can be seen from the figure that the total task processing delay of the four algorithms will gradually increase with the increase of the data volume. Compared with the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm, the total task processing delay of the proposed algorithm is reduced by 7.13%, 3.04%, and 21.08% respectively.

[0138] Figure 3It is a schematic diagram of the comparison between the data volume and the total energy consumption for task processing. This schematic diagram shows the variation of the total energy consumption for task processing of the proposed algorithm, the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm with the data volume. The abscissa in the figure represents the data volume, and the ordinate represents the total task processing delay. The red line, blue line, yellow line, and purple line respectively represent the proposed algorithm, the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm. It can be seen from the figure that the total energy consumption for task processing of the four algorithms gradually increases with the increase of the data volume. Compared with the random channel allocation algorithm, the average resource allocation algorithm, and the simulated annealing-particle swarm optimization algorithm, the total energy consumption for task processing of the proposed algorithm is reduced by 3.87%, 55.21%, and 15.10% respectively.

[0139] The above specific embodiments further elaborate on the invention purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above specific embodiments are only used as exemplary descriptions and do not limit the protection scope of the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. Establish an empty space-air-ground collaborative vehicle networking task offloading and resource allocation device, characterized in that, It includes the following steps: 1) Construct an air-ground-space collaborative vehicle Internet of Things computing offloading model to describe the computing task offloading and processing methods of vehicle user terminals; 2) Define the average delay and energy consumption functions and constraint conditions for all vehicle user terminals to process computing tasks; 3) Adopt an air-ground-space collaborative vehicle Internet of Things mobile edge computing communication and computing resource allocation algorithm to jointly solve the task offloading decision sub-problem and the computing resource allocation sub-problem. The simulation results show that the proposed algorithm has good performance.

2. The device for establishing the task offloading and resource allocation of the space-air-ground collaborative vehicle networking according to claim 1, characterized in that Step 1) Construct an empty space-ground collaborative vehicle network computing offloading model. The model consists of three layers of networks, namely the ground-based network, the air-based network, and the space-based network. Among them, the ground-based network contains a roadside unit, which is set on one side of the road. The radius of its coverage area is R, and it is responsible for communicating with vehicle user terminals. The roadside unit is equipped with a mobile edge computing server, which is connected by optical fiber. The mobile edge computing server has certain storage and computing capabilities, denoted as MS. The ground-based network also contains N vehicle user terminals, denoted as the set VUE = {VUE1,..., VUE i ,..., VUE N}, i ∈ [1, N], where VUE i represents the i-th vehicle user terminal. All vehicles are driving in the same direction on the road, and vehicles can communicate with each other. The radius of the coverage area of each vehicle is r. When VUE i is within the communication range of the roadside unit, the task can be forwarded to MS through the roadside unit for execution. In the air-based network, there is a hovering unmanned aerial vehicle equipped with a server denoted as UAS, which has certain storage and computing capabilities. The communication range of the unmanned aerial vehicle network is the same as that of the roadside unit. The space-based network contains a LEO and a ground station. The LEO is equipped with a server denoted as SAS, which has strong computing capabilities. The ground station is responsible for communicating with the LEO, and the communication range of the satellite can cover the entire ground network; VUE i The latency for the task to be computed locally is expressed as Among them, represents the binary offloading decision. When the VUE i task is executed locally, when the VUE i task selects other methods for offloading, d i represents the data volume size of the VUE i task, ω i represents the number of CPU cycles per bit of data of the VUE i , represents the CPU frequency of the VUE i ; VUE i The energy consumption of the task calculated locally is expressed as where κ v represents the effective open capacitance coefficient of the chip structure of VUE i ; VUE i Using C j The rate at which the task is offloaded to the MS can be expressed by the Shannon formula as Among them, B j represents the sub-channel bandwidth of C j , P i , P m respectively represent the transmission powers of VUE i and VUE m , σ 2 represents the noise power of the channel, D represents the number of vehicles within the communication range of the roadside unit, λ i，j represents the indication variable for allocating C j to VUE i , represents the interference of the V2I channel where VUE m reuses VUE i , among which, λ m，j represents the indication variable for allocating C j to VUE m , represents the channel gain between the MS and VUE m , represents the channel gain between the MS and VUE i , and its expression is where β i is the shadow fading, is the distance between VUE i and the MS, α is the path loss exponent, h i is the small-scale fading; VUE i The latency of offloading tasks to the MS is expressed as Among them, represents the binary offloading decision, when the tasks of the VUE i are executed on the MS; VUE i The latency of the task during MS calculation is expressed as where f i r represents the CPU frequency of the MS; Therefore, VUE i The total latency of the task executed at MS is T i r，all = T i r，o + T i r，c (6) VUE i The energy consumption of offloading tasks to the MS is expressed as VUE i The energy consumption for the task to be calculated in MS is expressed as where p r represents the calculated power of the MS; Therefore, VUE i The total energy consumption of the task executed by MS is VUE i Use C j Offload tasks to the surrounding vehicle VUE p The rate of... can be expressed by the Shannon formula as Among them, represents VUE a reuses the interference of the V2V channel of VUE, where λ i the V2V channel of VUE, where λ a，j represents allocating C j to VUE a the indication variable assigned to VUE, P a represents the transmission power of VUE a the transmission power of VUE represents the channel gain between VUE p and VUE a the channel gain between them represents the channel gain between VUE p and VUE i the channel gain between them, and its expression is where is the distance between VUE i and VUE p the distance between them; VUE i Offload tasks to the surrounding vehicle VUE p The time delay of Among them, represents a binary offloading decision, when the task of the VUE i is executed at the VUE p ; VUE i The task of the surrounding vehicle VUE p The time delay during calculation is Among them, represents the CPU frequency of VUE p ; Therefore, VUE i The total delay for the task executed in the surrounding vehicle VUE p is Among them, O i represents the number of vehicles within the VUE i communication range; VUE i Offload the task to the surrounding vehicle VUE p The energy consumption of VUE i The task of the surrounding vehicle VUE p The energy consumption for calculation is expressed as Therefore, VUE i The total energy consumption for the task executed by the surrounding vehicle VUE p is VUE i Using C j The rate at which tasks are offloaded to the UAS can be expressed by the Shannon formula as Among them, represents VUE q reusing the interference of the VUE i unmanned aerial vehicle channel, where λ q，j represents allocating C j to VUE q indicator variable, P q represents the transmit power of VUE q ; represents the channel gain between the UAS and VUE q ; represents the channel gain between the UAS and VUE i ; its expression is where θ0 represents the channel power gain at the reference distance l0 = 1m, h t and h r respectively represent the antenna power gains of VUE i and the unmanned aerial vehicle, w represents the wavelength, where is the distance between VUE i and the UAS; VUE i The latency of offloading tasks to the UAS is Among them, represents a binary offloading decision, when the task of the VUE i is executed at the UAS; VUE i The latency of the task during UAS calculation is Among them, f i u represents the CPU frequency of the UAS; Therefore, VUE i The total delay of the task executed by the UAS is T i u，all = T i u，o + T i u，c (20) VUE i The energy consumption of offloading tasks to the UAS is VUE i The energy consumption of the task calculated by the UAS is expressed as where p u represents the calculated power of the UAS; Therefore, VUE i The total energy consumption of the mission executed by the UAS is VUE i Using C j The rate of offloading tasks to the ground station can be expressed by the Shannon formula as Among them, represents VUE z reuse of VUE i interference of the ground station channel, where λ z，j represents assigning C j to VUE z indicator variable, P z respectively represent VUE z transmission power, represents the channel gain between the ground station and VUE i ; represents the channel gain between the ground station and VUE i The expression is where distance between the ground station and VUE i ; The task offloading rate from the ground station to the SAS can be expressed by the Shannon formula as Among them, P e represents the transmission power of the ground station, and g e，s represents the channel gain between the ground station and the SAS, which can be expressed as Among them, l e，s represents the distance between the ground station and the SAS, λ represents the wavelength of the signal, G ea and G sa respectively represent the gains at the ground station and the SAS antenna, and F rain represents rainfall fading; VUE i The latency of offloading tasks to the SAS is expressed as Among them, represents a binary offloading decision, and when the task of the VUE i is executed on the SAS; VUE i The latency of the task during SAS calculation is expressed as: Among them, f i s represents the CPU frequency of the SAS; Therefore, VUE i The total latency of the task executed in SAS is T i s，all = T i s，o + T i s，c (29) VUE i The energy consumption of offloading tasks to the SAS is expressed as VUE i The energy consumption for the task calculated in SAS is expressed as where p s represents the computing power of the SAS; Therefore, VUE i The total energy consumption of the task executed in SAS is 3. The device for establishing the task offloading and resource allocation of the space-air-ground collaborative vehicle network according to claim 1, wherein In step 2), the average delay and energy consumption functions for all vehicle user terminals to process computing tasks are defined and expressed as where μ represents the delay weight factor; The constraints of the offloading variables are The CPU cycle frequency constraint of the VUE is Among them, f v，max represents the maximum CPU frequency of VUE; The CPU cycle frequency constraint of the MS is Among them, f r，max represents the maximum CPU frequency of the MS; The CPU cycle frequency constraint of the UAS is where f u，max represents the maximum CPU frequency of the UAS; The CPU cycle frequency constraint of the SAS is Where f s，max represents the maximum CPU frequency of the SAS; Process VUE i The total delay constraint of the task is Among them, T i max represents the maximum tolerable delay for executing the VUE i task; Process VUE i The total energy consumption constraint of the task is Among them, represents VUE i the maximum energy consumption for task execution; The constraints of the channel reuse indicator variables are λ m，j ∈ {0, 1}, m ∈ D, j ∈ K (42) 4. The device for establishing task offloading and resource allocation in the space-air-ground collaborative vehicle network according to claim 1, wherein In step 3), an air-ground-space collaborative vehicle Internet of Things mobile edge computing communication and computing resource allocation algorithm is adopted to jointly solve the task offloading decision sub-problem and the computing resource allocation sub-problem; the optimization problem is respectively transformed into an offloading decision sub-problem and a computing resource allocation sub-problem, where the offloading decision sub-problem is expressed as: The computing resource allocation sub-problem is expressed as The air-ground-space collaborative vehicle Internet of Things mobile edge computing communication and computing resource allocation algorithm specifically includes the following steps: (1) Calculate the channel gain according to parameters such as vehicle location, task data volume, shadow fading, and small-scale fading; (2) Define the set of unassigned vehicles and the number of unassigned vehicles UV in the set, the set of unassigned channels and the number of unassigned channels UC in the set; (3) Select two VUEs with the farthest distance within the communication range of the RSU from the set to reuse the same channel, and remove the two VUEs that reuse the channel from the set and remove the channel they reuse from the set (4) Determine whether UV > UC is satisfied. If it is satisfied, return to (3); if not, enter (5); (5) is for the set Allocate independent channels for the remaining VUEs in (6) Let the initial CPU cycle frequency f0 be the optimal CPU cycle frequency; (7) Use the improved genetic algorithm and the current optimal CPU cycle frequency to solve the optimal offloading decision; (8) Use the improved particle swarm optimization algorithm and the current optimal offloading decision to solve the optimal CPU cycle frequency.

Images