Task collaborative unloading method and device for space-air-ground integrated scene, and simulation system

Abstract

The invention discloses a task collaborative unloading method and device for a space-air-ground integrated scene and a simulation system, and relates to the technical field of edge computing collaborative unloading. According to the task collaborative unloading method for the space-air-ground integrated scene, unloading of a target task is achieved through two stages; in the first stage, a matched execution end is selected according to task information of a target task; when the execution end is a satellite cluster end, second-stage unloading is carried out, and in the second-stage unloading, the matched execution satellite is selected to execute the target task according to the satellite information of the satellites. According to the method, the effectiveness of task collaborative unloading in a space-air-ground integrated system composed of the user side, the satellite cluster side and the cloud server side is improved.

Description

Technical Field

[0001] The present invention relates to the field of edge computing collaborative offloading technology, and in particular to a task collaborative offloading method and device, and a simulation system for an air-ground-space integrated scenario. Background Art

[0002] With the development of the integrated space-air-ground information network, edge computing has been widely used in the integrated space-air-ground network. It uses satellites to realize the communication between IoT terminals and cloud centers and to perform satellite-borne edge computing, that is, to deploy computing and storage hardware devices on satellites, migrate the capabilities of cloud centers to satellites, and realize local processing of data storage and task computing by satellites. With the continuous enrichment of the application scenarios of satellite-borne edge computing, a satellite cluster collaborative working mode based on clusters has emerged. In the integrated space-air-ground system composed of the user end, satellite cluster end and cloud server end, how to improve the effectiveness of task collaborative offloading in the integrated space-air-ground system composed of the user end, satellite cluster end and cloud server end has become a problem that needs to be solved. Summary of the invention

[0003] The present invention provides a task collaborative unloading method and device, and a simulation system for an air-space-ground integrated scenario, which are used to improve the effectiveness of task collaborative unloading in an air-space-ground integrated system composed of a user end, a satellite cluster end, and a cloud server end.

[0004] In order to solve the above technical problems, the first aspect of the present invention discloses a task collaborative offloading method for an air-ground integrated scenario, the method being used in an air-ground integrated system consisting of a user end, a satellite cluster end and a cloud server end, the method comprising:

[0005] Acquire task information of a target task generated by a user end, wherein the task information includes at least a task type and a task data volume;

[0006] According to the task information, an execution end matching the task information is selected from the user end, the satellite cluster end and the cloud server end, and the target task is unloaded to the execution end;

[0007] Wherein, if the execution end is a satellite cluster end, the step of unloading the target task to the execution end includes:

[0008] Acquire satellite information of each satellite in the satellite cluster, wherein the satellite information includes at least satellite status information and satellite attribute information;

[0009] An execution satellite matching the mission information is selected from the satellite cluster according to the satellite information, and the target mission is unloaded to the execution satellite.

[0010] As an optional implementation, in the first aspect of the present invention, the satellite status information includes the remaining time of unfinished tasks, the remaining data volume of unfinished tasks and the number of tasks; the satellite attribute information includes data transmission capacity parameters and computing capacity parameters;

[0011] And, selecting an execution satellite matching the mission information from the satellite cluster according to the satellite information includes:

[0012] According to the satellite attribute information and the mission information, the adaptation value corresponding to each satellite in the satellite cluster is calculated, and the satellite with the largest corresponding adaptation value is determined as the execution satellite.

[0013] As an optional implementation manner, in the first aspect of the present invention, the adaptation value is calculated according to the following formula:

[0014]

[0015] Among them, T i represents the remaining execution time of the unfinished task i in the satellite, D i represents the remaining data of the unfinished task i in the satellite, k is the preset mixing parameter, N m is the number of satellite missions, F m is the satellite’s computing power parameter, B m It is the data transmission capability parameter of the satellite.

[0016] As an optional embodiment, in the first aspect of the present invention, the method further comprises:

[0017] Acquire the position information of the execution satellite, and calculate the remaining time for the execution satellite to reach the communication edge according to the position information, wherein before the execution satellite reaches the communication edge, the communication between the execution satellite and the user terminal is maintained; after the execution satellite reaches the communication edge, the communication between the execution satellite and the user terminal is disconnected;

[0018] Determine whether the remaining time is less than or equal to a preset time threshold. If the remaining time is less than or equal to the preset time threshold, suspend the processing of the target task by the executing satellite, and control the executing satellite to migrate the task data corresponding to the target in the task suspension state to the backup satellite, wherein the backup satellite is a satellite that has not reached the communication edge.

[0019] As an optional implementation manner, in the first aspect of the present invention, selecting an execution end matching the task information from the user end, the satellite cluster end, and the cloud server end according to the task information includes:

[0020] If the task data volume of the target task is less than a first threshold, determining the user end as the execution end;

[0021] If the task amount of the target task is greater than or equal to the first threshold and less than the second threshold, the satellite cluster end is determined as the execution end;

[0022] If the task amount of the target task is greater than or equal to the second threshold and less than the third threshold, the cloud server end is determined to be the execution end.

[0023] As an optional implementation manner, in the first aspect of the present invention, if the task data amount of the target task is less than the first threshold, determining the user end as the execution end includes:

[0024] Input the target task into a preset computing simulation model, and obtain a user-side execution time prediction value and a transmission time prediction value corresponding to the target task output by the computing simulation model, wherein the user-side execution time prediction value is used to predict the time required for the user-side to execute the target task, and the transmission time prediction value is used to predict the time required for the target task to be transmitted between the user-side and the satellite cluster side;

[0025] If the task data volume of the target task is less than a first threshold, and the predicted value of the execution time of the user end is greater than the predicted value of the transmission time, the user end is determined to be the execution end.

[0026] As an optional implementation, in the first aspect of the present invention, the calculation simulation model calculates the user-side execution time prediction value by the following formula:

[0027]

[0028] in, Predict the execution time for the user side; is the computing power parameter of the user end; β is the preset computing power scaling factor; T wait The waiting delay of the target task determined according to the serial processing queue of the user end; is the data volume of the target task;

[0029] And, the calculation simulation model calculates the transmission time prediction value by the following formula:

[0030]

[0031] In the above formula, is the transmission time prediction value, is the upload time for uploading the target task to the satellite cluster end, The time it takes for the satellite cluster to transmit the processing results corresponding to the target task back to the user end;

[0032] Among them, upload time Determined by the following formula:

[0033]

[0034] In the above formula, is the data transmission rate between the user end and the satellite cluster end; n is the total number of tasks in the sending and receiving state between the user end and the satellite cluster end;

[0035] Return time Determined by the following formula:

[0036]

[0037] In the above formula, μ is the ratio of the preset processing result data volume to the target task data volume.

[0038] As an optional implementation, in the first aspect of the present invention, the data transmission rate between the user terminal and the satellite cluster terminal is Determined by the following formula:

[0039]

[0040] In the above formula, B is the communication bandwidth between the user end and the satellite cluster end, is the data upload capability parameter of the user end, N0 is the Gaussian white noise power, is the channel gain between the user end and the satellite cluster end;

[0041] Among them, the channel gain between the user end and the satellite cluster end is Determined by the following formula:

[0042]

[0043] In the above formula, d m,n is the geographical distance between the user terminal and the satellite cluster terminal, and s is the preset path loss index.

[0044] The second aspect of the present invention discloses a task collaborative offloading simulation system for an air-ground-integrated scenario, the simulation system is used to simulate the task collaborative offloading method for the air-ground-integrated scenario in the first aspect of the present invention, the system comprising:

[0045] Computing platform model, used to model the user side, satellite cluster side and cloud server side;

[0046] A target task model, used to model the target task generated by the user end;

[0047] Computational model, used to model the transmission and computation process of the target task;

[0048] A collaborative offloading model is used to model the offloading process of the target task, and obtain the execution time and energy consumption of the target task when the task collaborative offloading method of the air-ground integrated scenario in the first aspect of the present invention is executed;

[0049] The non-cooperative offloading model is used to model the process of target task execution by the user end or target task transmission to the cloud server end through the satellite cluster end, and obtain the execution time and energy consumption of the target task when it is executed;

[0050] A comparison model is used to compare the execution time and energy consumption output by the collaborative offloading model and the non-collaborative offloading model, so as to determine the effectiveness of the task collaborative offloading method for the air-ground integrated scenario in the first aspect of the present invention.

[0051] The third aspect of the present invention discloses a task collaborative offloading device for an air-ground integrated scenario, the device comprising:

[0052] A memory storing executable program code;

[0053] a processor coupled to the memory;

[0054] The processor calls the executable program code stored in the memory to execute the task collaborative unloading method for the air-ground integrated scenario disclosed in the first aspect of the present invention.

[0055] The fourth aspect of the present invention discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute the task collaborative unloading method for the air-ground integrated scenario disclosed in the second aspect of the present invention.

[0056] Compared with the prior art, the task collaborative unloading method for the air-space-ground integrated scenario in the embodiment of the present invention realizes the unloading of the target task through two stages. In the first stage, a matching execution end is selected according to the task information of the target task; when the execution end is a satellite cluster end, the second stage of unloading is performed. In the second stage of unloading, a matching execution satellite is selected to execute the target task according to the satellite information of the satellite. This method improves the effectiveness of task collaborative unloading in the air-space-ground integrated system composed of the user end, the satellite cluster end and the cloud server end. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0058] Figure 1 Schematic diagram of an air-ground integrated system consisting of a user end, a satellite cluster end and a cloud server end in an embodiment of the present invention;

[0059] Figure 2 It is a flowchart of a task collaborative offloading method for an air-ground integrated scenario disclosed in an embodiment of the present invention;

[0060] Figure 3 It is a schematic diagram of a process of selecting an execution satellite matching mission information from a satellite cluster disclosed in an embodiment of the present invention;

[0061] Figure 4 It is a pseudo code diagram for realizing the process of selecting the execution satellite matching the mission information from the satellite cluster;

[0062] Figure 5 is a schematic diagram of the task data migration process disclosed in an embodiment of the present invention;

[0063] Figure 6 It is a pseudo code diagram for implementing task data migration;

[0064] Figure 7 It is a flowchart of constructing a task collaborative offloading simulation system for an air-ground integrated scenario disclosed in an embodiment of the present invention;

[0065] Figure 8 It is a task state machine diagram of a task model in a task collaborative offloading simulation system for an air-ground integrated scenario;

[0066] Fig. 9 It is a structural schematic diagram of a task collaborative unloading device for an air-ground integrated scenario disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or end including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or ends.

[0069] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0070] Embodiment 1

[0071] See also Figure 2 , Figure 2 A task collaborative offloading method for an air-space-ground integrated scenario disclosed in an embodiment of the present invention is used in an air-space-ground integrated system composed of a user terminal, a satellite cluster terminal and a cloud server terminal. The method can be integrated in a local server or a cloud server, or in any of the user terminal, the satellite cluster terminal and the cloud server terminal. Figure 2 As shown, the method may include:

[0072] Step 101: Obtain task information of a target task generated by a user terminal.

[0073] In the embodiment of the present invention, the task information may include at least a task type and a task data volume. The task type may be used to determine whether the target task can be executed in the user end, the satellite cluster end, or the cloud server end, and the task data volume may be the amount of data required by the task during the transmission process.

[0074] Step 102: Filter out an execution end that matches the task information from the user end, satellite cluster end, and cloud server end according to the task information, and unload the target task to the execution end.

[0075] In the embodiment of the present invention, the user end, the satellite cluster end and the cloud server end each have different task processing characteristics. For example, in general, the cloud computing center of the cloud server end can perform parallel computing without loss of computing power, the satellite edge computing platform of the satellite cluster end can perform parallel computing with loss of computing power, and the terminal device of the user end cannot perform parallel computing. Therefore, the target task can be assigned to the execution end with threshold matching for execution according to the task information of the target task. For example, tasks with small data volume and short processing time are offloaded to the user end for execution, tasks with medium data volume and medium processing time are offloaded to the satellite cluster end for execution, and tasks with large data volume and long processing time are offloaded to the cloud server end for execution.

[0076] In the embodiment of the present invention, further, if the execution end is a satellite cluster end, unloading the target task to the execution end may include:

[0077] Acquire satellite information of each satellite in the satellite cluster, where the satellite information may at least include satellite status information and satellite attribute information;

[0078] According to the satellite information, the executing satellite matching the mission information is selected from the satellite cluster, and the target mission is unloaded to the executing satellite.

[0079] As the application scenarios of satellite-borne edge computing continue to expand, a cluster-based collaborative working mode of satellite clusters has emerged. Therefore, it is necessary to further offload the target tasks and determine the executing satellites.

[0080] In an embodiment of the present invention, the determination of the execution satellite is based on the load balancing of each satellite in the star cluster, with the purpose of improving the efficiency of task execution and resource utilization. In an example case, due to the difference in the distribution of surface users, the services received by the satellite nodes in the low-orbit satellite network are uneven, and the satellites with limited resources cannot complete the processing independently, resulting in satellite congestion. For example, there are M satellites in the low-orbit satellite network, and each satellite can build an intersatellite link with the adjacent satellite to form a star cluster. The cluster head satellite will collect all satellite information of each satellite in its cluster, and the satellite information may at least include satellite status information and satellite attribute information; according to the satellite information, the execution satellite matching the task information is selected from the satellite cluster, and the target task is unloaded to the execution satellite. Among them, the cluster head satellite requests the execution satellite with idle resources, unloads the task data to the execution satellite through the intersatellite link, and returns the result to the original satellite after the execution satellite processes.

[0081] It can be seen that the task collaborative unloading method for the air-space-ground integrated scenario in the embodiment of the present invention realizes the unloading of the target task through two stages. In the first stage, a matching execution end is selected according to the task information of the target task; when the execution end is a satellite cluster end, the second stage of unloading is performed. In the second stage of unloading, a matching execution satellite is selected to execute the target task according to the satellite information of the satellite. This method improves the effectiveness of task collaborative unloading in the air-space-ground integrated system composed of the user end, the satellite cluster end and the cloud server end.

[0082] In an optional embodiment, the satellite status information may include the remaining time of the unfinished task, the remaining data volume of the unfinished task and the number of tasks; the satellite attribute information may include data transmission capacity parameters and computing capacity parameters;

[0083] And, selecting an execution satellite matching the mission information from the satellite cluster according to the satellite information may include:

[0084] According to the satellite attribute information and mission information, the adaptation value corresponding to each satellite in the satellite cluster is calculated, and the satellite with the largest corresponding adaptation value is determined as the execution satellite.

[0085] Optional, such as Figure 3 As shown in the figure, after the task is offloaded to the edge for execution in the second stage, the cluster head satellite that receives the task scores the satellites in the cluster according to the satellite status information and satellite attribute information, calculates the adaptation value, selects the execution node according to the node score in the cluster, and sends the task to the execution node. After receiving the task, the execution node allocates the corresponding computing resources for the task, starts to execute the task, and transmits the task back after the execution is completed.

[0086] Optionally, the adaptation value can be calculated according to the following formula:

[0087]

[0088] Among them, T i represents the remaining execution time of the unfinished task i in the satellite, D i represents the remaining data of the unfinished task i in the satellite, k is the preset mixing parameter, N m is the number of satellite missions, F m is the satellite’s computing power parameter, B m It is the data transmission capability parameter of the satellite.

[0089] Among them, the calculation process of the above adaptation value can be achieved through Figure 4 Pseudocode implementation in .

[0090] In another optional embodiment, the method may further include:

[0091] Acquire the position information of the executing satellite, and calculate the remaining time for the executing satellite to reach the communication edge according to the position information, wherein before the executing satellite reaches the communication edge, the communication between the executing satellite and the user end is maintained; after the executing satellite reaches the communication edge, the communication between the executing satellite and the user end is disconnected;

[0092] Determine whether the remaining time is less than or equal to a preset time threshold. If the remaining time is less than or equal to the preset time threshold, suspend the processing of the target task by the executing satellite, and control the executing satellite to migrate the task data corresponding to the target in the task suspension state to the backup satellite, where the backup satellite is a satellite that has not reached the communication edge.

[0093] Since satellites are in a high-speed moving state in the inter-satellite link, the terminal equipment may lose connection with the satellite performing the mission, resulting in the interruption of computing tasks or even data loss. Therefore, a task migration model considering the dynamic and time-varying characteristics of the satellite network is proposed.

[0094] The constant movement of satellites in orbit causes changes in the satellite network. After a satellite has been operating in a star cluster for a period of time, it may be out of the communication range with the user-end device. Therefore, it is necessary to determine whether the remaining time is less than or equal to the preset time threshold. If the remaining time is less than or equal to the preset time threshold, the processing of the target task by the executing satellite is suspended, and the executing satellite is controlled to migrate the task data corresponding to the target in the task suspension state to the backup satellite.

[0095] Exemplarily, this optional embodiment may include: in the intersatellite link, the network topology between satellites is relatively fixed, so when the satellite is performing a task, it regularly reports the task execution status and the distance from the boundary to the cluster head satellite. The cluster head satellite predicts the time when the satellite is about to fly away from the satellite cluster by the boundary distance, and initiates task migration when the satellite approaches the boundary. The task migration algorithm will send a task migration command to the satellite before it flies away from the cluster based on the satellite departure prediction result. After receiving the command, the satellite suspends the execution of the local task and migrates the task to the cluster head satellite. The cluster head satellite re-scores the satellite nodes in the cluster, selects the best execution node, and redistributes the task to the execution node.

[0096] The task migration process is as follows Figure 5 When the cluster head satellite receives the task to be migrated, it re-scores the satellite nodes in the cluster, selects the execution node according to the node scores in the cluster, and sends the task to the execution node. After receiving the task, the execution node allocates the corresponding computing resources to the task, starts to execute the task, and transmits the task back after the task is completed. The pseudo code for the above task migration algorithm is as follows: Figure 6 shown.

[0097] In yet another optional embodiment, selecting an execution end matching the task information from the user end, the satellite cluster end, and the cloud server end according to the task information may include:

[0098] If the task data volume of the target task is less than the first threshold, determining the user end as the execution end;

[0099] If the task amount of the target task is greater than or equal to the first threshold and less than the second threshold, the satellite cluster end is determined as the execution end;

[0100] If the task amount of the target task is greater than or equal to the second threshold and less than the third threshold, the cloud server end is determined as the execution end.

[0101] This optional embodiment allocates the target task to the execution end with matching threshold value for execution according to the task information of the target task. That is, tasks with small data volume and short processing time are offloaded to the user end for execution, tasks with medium data volume and medium processing time are offloaded to the satellite cluster end for execution, and tasks with large data volume and long processing time are offloaded to the cloud server end for execution.

[0102] In this optional embodiment, specific values ​​of the first threshold and the second threshold can be adaptively selected according to different application scenarios.

[0103] In yet another optional embodiment, if the task data volume of the target task is less than the first threshold, determining the user end as the execution end may include:

[0104] Input the target task into a preset computing simulation model, and obtain a user-side execution time prediction value and a transmission time prediction value corresponding to the target task output by the computing simulation model;

[0105] If the task data volume of the target task is less than the first threshold, and the predicted value of the execution time of the user end is greater than the predicted value of the transmission time, the user end is determined to be the execution end.

[0106] In this optional embodiment, for the target task whose task data volume is less than the first threshold, it is also necessary to determine which execution end is more suitable to assign it to based on the comparison between its execution time and transmission time. Among them, the user-side execution time prediction value is used to predict the time required for the user-side to execute the target task, and the transmission time prediction value is used to predict the time required for the target task to be transmitted between the user-side and the satellite cluster end. If the task data volume of the target task is less than the first threshold, and the user-side execution time prediction value is greater than the transmission time prediction value, it is considered that it is better to assign the target task to the user-side for execution; if the task data volume of the target task is less than the first threshold, and the user-side execution time prediction value is less than or equal to the transmission time prediction value, it is considered that it is better to assign the target task to the satellite cluster end.

[0107] In this optional embodiment, the simulation model is used to analyze the execution time and transmission time of the target task. The simulation model can be a pre-trained artificial intelligence model or a computational model established based on set modeling rules.

[0108] Optionally, the computational simulation model may calculate the predicted value of the user-side execution time by the following formula:

[0109]

[0110] in, Predict the execution time for the user side; is the computing power parameter of the user end; β is the preset computing power scaling factor; T wait The waiting delay of the target task determined according to the serial processing queue of the user end; is the data volume of the target task;

[0111] And, the calculation simulation model calculates the transmission time prediction value through the following formula:

[0112]

[0113] In the above formula, is the transmission time prediction value, is the upload time of the target task to the satellite cluster. The time it takes for the satellite cluster to transmit the processing results corresponding to the target task back to the user end;

[0114] Among them, upload time Determined by the following formula:

[0115]

[0116] In the above formula, is the data transmission rate between the user end and the satellite cluster end; n is the total number of tasks in the sending and receiving state between the user end and the satellite cluster end;

[0117] Return time Determined by the following formula:

[0118]

[0119] In the above formula, μ is the ratio of the preset processing result data volume to the target task data volume.

[0120] The data transmission rate It can be obtained through experiments according to the current application scenario, or it can be set manually.

[0121] In another optional embodiment, the data transmission rate between the user end and the satellite cluster end is Determined by the following formula:

[0122]

[0123] In the above formula, B is the communication bandwidth between the user end and the satellite cluster end, is the data upload capability parameter of the user end, N0 is the Gaussian white noise power, is the channel gain between the user end and the satellite cluster end;

[0124] Among them, the channel gain between the user end and the satellite cluster end is Determined by the following formula:

[0125]

[0126] In the above formula, d m,n is the geographical distance between the user terminal and the satellite cluster terminal, and s is the preset path loss index.

[0127] Embodiment 2

[0128] With the increasing number of computing scenarios and the diversification of task types in cloud-edge-device collaborative networks, the quality of task collaborative offloading algorithms for cloud-edge-device collaborative networks is becoming increasingly critical to the work efficiency of collaborative networks. However, it is difficult and costly to perform task collaborative offloading on real hardware platforms. To this end, simulation models such as computing models, offloading models, channel models, and energy harvesting models have been developed to simulate and evaluate task collaborative offloading algorithms in cloud-edge-device collaborative scenarios.

[0129] In view of the difficulty of existing methods in testing the effectiveness of task collaborative offloading algorithms in an integrated space-ground-air information network, an embodiment of the present invention proposes a new simulation system. The simulation system includes a collaborative computing scenario, a computing platform model, a task model, a computing model, and an offloading model, such as Figure 7 As shown in the figure, the establishment process of the simulation system can be: by abstracting the air-space-ground integrated information network to obtain a collaborative computing scenario, modeling the hardware resources of three computing platforms, namely the cloud computing center, the satellite edge computing platform and the IoT end-side device, to obtain a computing platform model, modeling the tasks generated by the IoT end-side device to obtain a task model, analyzing and modeling the task offloading process to obtain a computing model, and finally modeling the cluster-based satellite cluster collaborative working mode to obtain an offloading model. Using this simulation system, the modeling and simulation of the air-space-ground integrated scenario can be completed quickly and accurately, and the effectiveness verification of the task collaborative offloading algorithm can be finally completed.

[0130] Based on the above simulation model establishment process, an embodiment of the present invention proposes a task collaborative offloading simulation system for an air-ground integrated scenario, which is used to simulate the task collaborative offloading method for the air-ground integrated scenario in Embodiment 1. The simulation system may include:

[0131] Computing platform model, used to model the user side, satellite cluster side and cloud server side;

[0132] The target task model is used to model the target tasks generated by the user end;

[0133] Computational model, used to model the transmission and computation process of the target task;

[0134] A collaborative offloading model is used to model the offloading process of the target task, and obtain the execution time and energy consumption of the target task when it is executed according to the task collaborative offloading method of the air-ground integrated scenario in the first embodiment;

[0135] The non-cooperative offloading model is used to model the process of target task execution by the user end or target task transmission to the cloud server end through the satellite cluster end, and obtain the execution time and energy consumption of the target task when it is executed;

[0136] A comparison model is used to compare the execution time and energy consumption output by the collaborative offloading model and the non-collaborative offloading model, so as to determine the effectiveness of the task collaborative offloading method for the air-ground integrated scenario in Example 1.

[0137] The following are detailed descriptions:

[0138] 1. Computing platform model

[0139] The computing platform models in the simulation scenario include the terminal-side computing platform, the edge-side computing platform, and the cloud-side computing platform.

[0140] The edge computing platform model is a satellite computing platform. As an important part of modern communication technology, satellites can provide more efficient and reliable communication services and data transmission for ground users. In satellite edge computing, satellites can usually be classified according to dimensions such as orbital altitude, orbital type, constellation type and frequency band type. The satellite in the embodiment of the present invention may be a low-earth orbit satellite classified according to orbital altitude. The altitude of low-earth orbit satellites is generally below 1,000 kilometers. They have a shorter orbital period and a faster orbital speed. They are usually used for earth observation, communication, navigation and other applications. Since such satellites are closer to the earth, they can provide lower latency communications and higher bandwidth. The satellite computing platform model established in the embodiment of the present invention represents inter-satellite heterogeneity through differences in satellite computing power and transmission capacity. The parameters include: Two Line Elements, also known as TLE orbital elements, is a general form of orbital element publication, which describes the expression of the position and velocity of a flying body in space; computing power parameter P sat , describing the speed at which the satellite performs calculations; the transmission capacity parameter T Sat , describing the speed at which satellites can transmit data.

[0141] The cloud computing platform model is a cloud server computing platform. Compared with the satellite computing platform, the cloud server computing platform has more abundant computing resources, storage resources and a wider coverage, and has more powerful performance in computing-intensive application processing methods. The cloud server computing platform model established in the embodiment of the present invention has a unified communication distance, computing power and transmission capacity. The specific parameters may include: communication distance L cloud , describing the communication distance between the cloud and the satellite, in meters; computing power parameter P cloud , describing the speed at which cloud computing centers perform calculations; the transmission capacity parameter T cloud , describing the speed at which data is transferred in a cloud computing center.

[0142] The end-side computing platform is the terminal device, which is randomly distributed on the surface of the earth. As the low-earth orbit satellite moves, the terminal devices covered by the satellite can offload computing tasks to the low-earth orbit satellite edge nodes for processing through the satellite-to-ground link. The end-side computing platform model of the embodiment of the present invention has the following characteristics: 1. The end-side device randomly generates a variety of computing tasks according to its own needs; 2. The end-side device has a certain computing power and can select a task offloading strategy according to task requirements; 3. The end-side device can only offload computing tasks to the low-earth orbit satellite edge nodes for execution within the coverage of the low-earth orbit satellite. Terminal device parameters include: location data, including latitude, longitude and altitude; computing power parameter P end , describes the speed at which the terminal device performs task calculations; the transmission capacity parameter T end, describes the speed at which the terminal device transmits data; the task trigger frequency, describes the time interval between the current task and the next task of the terminal device, in seconds.

[0143] During the simulation of the computing platform model, the satellite status is divided into offline, idle, and busy, which respectively represent whether the satellite forms a cluster, whether the satellite has no unfinished tasks, and whether the satellite has unfinished tasks. The terminal status is divided into idle and busy. The idle terminal device can trigger the task immediately, and the busy terminal device is waiting for the task to be completed. The cloud defaults to constant data transmission capacity and computing capacity, which are not affected by the number of tasks.

[0144] 2. Task Model

[0145] In the embodiment of the present invention, in order to make the constructed model easier to analyze, it is assumed that the terminal device triggers tasks at a fixed frequency, and the task triggering frequencies of different terminal devices are different, and the tasks generated by the terminal device are modeled by task type, task data size and task execution time. The specific parameters include: task type, which is used to mark whether the task can be executed on the terminal or satellite; task data size D task , used to describe the amount of data transferred by the task, in MB; standard execution time T task , used to describe the task when the computing power is 1 (P sat =1;P cloud =1;P end =1) on a computing platform, in seconds.

[0146] Through the above modeling method, the end-side device generates a computing task at time t = {1, ..., T}, and the end-side device Only one computing task is generated at time t It can be expressed as in Represents the data size of the task, represents the computing time of the task on the above standard computing platform, Represents the task type of this task.

[0147] The task status during the simulation process is divided into untriggered, waiting, sending, executing, returning and completed. Figure 8As shown. Task untriggered state: The terminal device counts down the task triggering according to its own task triggering frequency. When the terminal is in an idle state, the task it carries is in an untriggered state, and no task and energy consumption calculation is performed. Task waiting state: When the terminal state task meets its own task triggering frequency, the task is triggered, and the first untriggered task in its own untriggered task queue is triggered. If the terminal device is in an idle state, the task skips the waiting state; if the terminal device is in a busy state, the task enters the waiting state. Task sending state: When the terminal starts to trigger the task, the first stage of scheduling is first performed according to the unloading algorithm in the collaborative offloading test. The default collaborative offloading algorithm selects the unloading platform for the waiting task according to the task type and terminal status. If the task is offloaded to its own terminal device for execution, the task sending state is skipped; if the task is offloaded to the cloud computing center or the satellite edge computing platform, the task sending state is entered. Task executing state: When the task is offloaded, it enters the executing state, and the task execution platform processes the task according to its own computing power. The cloud computing center can perform parallel computing without computing power loss, the satellite edge computing platform can perform parallel computing with computing power loss, and the terminal device cannot perform parallel computing. Task return status: When the computing platform completes task execution, if it is a terminal device, it skips the return, otherwise it enters the return status.

[0148] 3. Computational Model

[0149] The simulation scenario of the embodiment of the present invention includes two computing models, namely a local computing model and an edge computing model.

[0150] Local computing model:

[0151] The local computing power and battery capacity of the end-side device are limited, but it can still undertake some computing tasks to improve the overall performance of the system. indicates that, and Because the single-side device can only process tasks serially, the task Each time a simulation time granularity passes in the waiting state, the task increases the waiting delay. The latency generated by local calculation is:

[0152]

[0153] β is the unified computing power scaling factor of the system, T wait The waiting delay of the task.

[0154] The energy consumption generated by local computing on the end device can be obtained based on the above definition: The energy consumption generated locally is:

[0155]

[0156] ε is the system default execution energy consumption coefficient.

[0157] Edge computing model:

[0158] Different from the local computing model, offloading tasks to edge or cloud computing platforms for computing is divided into two stages. In the first stage, the task is in the wireless channel, and the end-side device needs to transmit the content data of the task to the edge computing platform. In this stage, transmission delay and end-side transmission energy consumption are generated. In the second stage, the task has reached the edge computing platform, the edge computing platform receives the task data, and dispatches the corresponding computing resources to execute the task. In this stage, execution delay and task distribution delay are generated.

[0159] For the wireless communication between the end-side device and the edge computing platform, it is assumed that an orthogonal frequency division multiple access scheme is used. The communication between all end-side devices and the edge computing platform uses the same bandwidth B and is orthogonal to each other. At time t, user device n can only maintain a communication connection with one edge computing platform m and initiate an offload request to it. The upload capacity of end-side device n is T end , user uplink transmission rate It can be expressed as:

[0160]

[0161] Where N0 is the Gaussian white noise power. Represents the channel gain between the end-side computing platform and the end-side device, which can be calculated by the following formula:

[0162]

[0163] Among them, d m,n is the geographical distance between the edge computing platform and the end-side device, and s is the path loss index. According to the above formula, the transmission rate can be obtained, and the task in the first stage can be obtained. Upload latency caused by offloading to the edge computing platform Upload energy consumption They are:

[0164]

[0165] Where n represents the total number of tasks in which the terminal is sending and receiving within the current simulation granularity, and γ and δ represent the default transmission energy consumption scaling factors when the terminal sends.

[0166] The result after the calculation is usually very small. The embodiment of the present invention uses a return parameter μ to represent the ratio of the system default task return data volume to the task data volume. According to the above formula, the task return delay can be obtained and task return energy consumption They are:

[0167]

[0168] Where n represents the total number of tasks in which the terminal is sending and receiving within the current simulation granularity, and γ and δ represent the default transmission energy consumption scaling factors when the terminal sends.

[0169] In the second phase, the task It has arrived at the edge computing platform or cloud computing platform. If it is at the edge computing platform, the master node x distributes the task to the actual execution node m through the task offloading scheduling within the cluster. The distribution delay calculation formula in this process is as follows:

[0170]

[0171] When the task After arriving at the actual execution computing platform m, the execution process is carried out. The available computing power of the actual execution computing platform for:

[0172]

[0173] Among them, P m is the total computing capacity of the actual computing platform m, and n is the number of tasks being executed on the computing platform. From the above formula, we can get the task The execution delay generated by offloading to computing platform m is:

[0174]

[0175] Where β represents the uniform computing power scaling factor.

[0176] When the task After completion, the total completion time calculated by the local calculation model and the offload calculation model is:

[0177]

[0178] Task The terminal energy consumption calculation formula is:

[0179]

[0180] 5. Collaborative Unloading Model

[0181] An embodiment of the present invention proposes a two-stage task offloading model suitable for an air-ground-ground integrated scenario. The offloading model reduces the task completion time and the energy consumption of the terminal-side device by offloading part of the tasks to the edge side for execution.

[0182] When the task data is generated in the terminal device, it is executed locally by the ground terminal device or uploaded to the ground cloud computing center through the low-orbit constellation for calculation. This mode is called non-cooperative computing. The simulation model obtained based on the non-cooperative offloading calculation modeling is a non-cooperative offloading model. In this task scenario, there are problems such as weak computing power of the terminal device and high proportion of task transmission delay. The task collaborative offloading model is divided into two stages: the first stage determines the target platform for task offloading according to the task type and the terminal device situation, and can be offloaded to three computing platforms: the terminal side, the edge side, and the cloud side; the core algorithm of the second stage is the intra-cluster task offloading algorithm based on load balancing. In this stage, for the tasks offloaded to the edge side, secondary offloading is performed according to the overall resource status of the satellite cluster.

[0183] In the low-orbit satellite network resource management scenario, computing tasks can be executed locally or offloaded to the server of the low-orbit satellite or the ground cloud computing center for execution. The embodiment of the present invention sets up three types of tasks. In the first stage of offloading, the task needs to be offloaded to the corresponding computing platform according to the task type and the cloud-edge device situation. The discussion according to the task type is as follows:

[0184] For Class I tasks, the calculation time of this type of tasks is short and the data volume is low, and they can be executed on the end-side device. In the first stage of offloading, the end-side device situation and the edge-side device situation will be comprehensively considered, and the algorithm will give priority to offloading the task to the edge-side computing platform for calculation. If there is no low-orbit satellite that can communicate within the communication range of the end-side device that triggers the task, local execution will be given priority; if there is a low-orbit satellite that can communicate within the communication range of the end-side device that triggers the task, the predicted execution delay of the end-side device will be compared with the predicted delay for uploading and returning the task. If the predicted execution delay is lower than the predicted delay for uploading and returning, it is considered better to offload the task to the local execution, otherwise the task will be offloaded to the satellite cluster for execution.

[0185] For Class II tasks, the computing time and data volume of this type of tasks are medium, and they can be executed on edge devices. In the first stage of offloading, the edge device situation is mainly considered, and offloading to the edge computing platform for calculation is prioritized. If there is no low-orbit satellite that can communicate within the communication range of the end-side device that triggers the task, it will be directly offloaded to the cloud-side computing center for execution, and the results will be transmitted back after the task is completed; if there is a low-orbit satellite that can communicate within the communication range of the end-side device that triggers the task, the task will be directly offloaded to the edge satellite cluster for execution, and the results will be transmitted back after the task is completed.

[0186] For Class III tasks, which have long computing time and high data volume, they can only be executed in the cloud computing center. In the first stage of offloading, the tasks are offloaded to the cloud computing center for execution. If there is no low-orbit satellite that can communicate within the communication range of the end-side device that triggers the task, it will wait.

[0187] In the first stage of offloading, if the task is offloaded to the side satellite cluster for execution, the second stage of offloading is performed in the satellite cluster. The second stage of offloading algorithm directly adopts the second stage of offloading algorithm in the first embodiment.

[0188] It can be seen that a new task collaborative offloading simulation model and modeling method for the air-space-ground integrated scenario is implemented in the embodiment of the present invention, which supports lightweight, convenient and scalable implementation of task offloading simulation for the air-space-ground integrated scenario. Specifically, a collaborative computing scenario is obtained by abstracting the air-space-ground integrated information network, the hardware resources of three computing platforms including the cloud computing center, the satellite edge computing platform and the IoT end-side device are modeled to obtain a computing platform model, the tasks generated by the IoT end-side device are modeled to obtain a task model, the task offloading process is analyzed and modeled to obtain a computing model, and finally the cluster-based satellite cluster collaborative working mode is modeled to obtain an offloading model. Compared with the physical simulation, the simulation model obtained by this model and modeling method is more lightweight and convenient while ensuring a certain accuracy, and has stronger scalability and customization.

[0189] Embodiment 3

[0190] The embodiment of the present invention proposes a task collaborative offloading algorithm test platform for low-orbit constellations. In the task collaborative offloading algorithm test platform, it is required to quickly and simply complete the simulation of cloud-edge-end scenarios and perform task collaborative offloading algorithm testing in the simulation. The simulation model and modeling method proposed in the embodiment of the present invention can effectively improve the data processing speed and ensure the simulation accuracy. The specific implementation steps are as follows:

[0191] Scenario modeling: First, define a plurality of computing platforms including a cloud computing center, a low-orbit satellite edge computing platform, and IoT end-side devices. The modeling method proposed by the embodiment of the present invention completes the modeling and simulation of these computing platforms, and models and simulates the tasks generated by the IoT end-side devices. These computing platforms differ in computing power, transmission capacity, storage capacity, device power, etc., and the end-side tasks also differ in task size, task type, task computing amount, etc.

[0192] Satellite simulation: orbit calculation is performed based on the TLE orbit data of the satellite model to obtain the satellite orbit data within the simulation time. During the simulation process, the change of satellite hardware resources is simulated according to the satellite's mission load.

[0193] Task simulation: Perform a series of simulations such as task triggering, scheduling, transmission, execution, and return according to the task triggering time of the task model and the task triggering frequency of the end-side device task model, and obtain the execution time and energy consumption of each task based on the calculation model.

[0194] Collaborative offloading: The tasks triggered in the simulation are offloaded and scheduled according to the offloading model, and task migration simulation is performed on some tasks in the satellite cluster.

[0195] Non-cooperative offloading: Customized modification of the offloading model can realize non-cooperative offloading simulation, and the execution time and energy consumption of the task in the non-cooperative offloading scenario can be obtained, which can be compared with the cooperative offloading to obtain the effectiveness and optimization rate of the tested algorithm.

[0196] Embodiment 4

[0197] See also Fig. 9 , Fig. 9 Schematic diagram of the structure of a task collaborative offloading device for an air-ground integrated scenario disclosed in an embodiment of the present invention. Fig. 9 As shown, the task collaborative offloading device of the air-ground integrated scenario may include:

[0198] A memory 201 storing executable program codes;

[0199] a processor 202 coupled to the memory 201;

[0200] The processor 202 calls the executable program code stored in the memory 201 to execute the steps of the task collaborative offloading method for the air-ground integrated scenario described in the first embodiment of the present invention.

[0201] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0202] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0203] Finally, it should be noted that the task collaborative offloading method, device, and simulation system for an air-ground integrated scenario disclosed in the embodiments of the present invention only disclose preferred embodiments of the present invention, and are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features therein may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A task collaborative offloading method for an air-ground integrated scenario, characterized in that: The method is used in an air-ground integrated system consisting of a user end, a satellite cluster end and a cloud server end, and the method includes: Acquire task information of a target task generated by a user end, wherein the task information includes at least a task type and a task data volume; According to the task information, an execution end matching the task information is selected from the user end, the satellite cluster end and the cloud server end, and the target task is unloaded to the execution end; Wherein, if the execution end is a satellite cluster end, the step of unloading the target task to the execution end includes: Acquire satellite information of each satellite in the satellite cluster, wherein the satellite information includes at least satellite status information and satellite attribute information; An execution satellite matching the mission information is selected from the satellite cluster according to the satellite information, and the target mission is unloaded to the execution satellite.

2. The task collaborative offloading method for the air-ground integrated scenario according to claim 1 is characterized in that: The satellite status information includes the remaining time of unfinished tasks, the remaining data volume of unfinished tasks and the number of tasks; the satellite attribute information includes data transmission capacity parameters and computing capacity parameters; And, selecting an execution satellite matching the mission information from the satellite cluster according to the satellite information includes: According to the satellite attribute information and the mission information, the adaptation value corresponding to each satellite in the satellite cluster is calculated, and the satellite with the largest corresponding adaptation value is determined as the execution satellite.

3. The task collaborative offloading method for the air-ground integrated scenario according to claim 2 is characterized in that: The adaptation value is calculated according to the following formula: Among them, T i represents the remaining execution time of the unfinished task i in the satellite, D i represents the remaining data of the unfinished task i in the satellite, k is the preset mixing parameter, N m is the number of satellite missions, F m is the satellite’s computing power parameter, B m It is the data transmission capability parameter of the satellite.

4. The task collaborative offloading method for the air-ground integrated scenario according to claim 3 is characterized in that: The method further comprises: Acquire the position information of the execution satellite, and calculate the remaining time for the execution satellite to reach the communication edge according to the position information, wherein before the execution satellite reaches the communication edge, the communication between the execution satellite and the user terminal is maintained; after the execution satellite reaches the communication edge, the communication between the execution satellite and the user terminal is disconnected; Determine whether the remaining time is less than or equal to a preset time threshold. If the remaining time is less than or equal to the preset time threshold, suspend the processing of the target task by the executing satellite, and control the executing satellite to migrate the task data corresponding to the target in the task suspension state to the backup satellite, wherein the backup satellite is a satellite that has not reached the communication edge.

5. The task collaborative offloading method for the air-ground integrated scenario according to claim 1 is characterized in that: The method further comprises: selecting an execution end matching the task information from the user end, the satellite cluster end, and the cloud server end according to the task information, including: If the task data volume of the target task is less than a first threshold, determining the user end as the execution end; If the task amount of the target task is greater than or equal to the first threshold and less than the second threshold, the satellite cluster end is determined as the execution end; If the task amount of the target task is greater than or equal to the second threshold and less than the third threshold, the cloud server end is determined to be the execution end.

6. The task collaborative offloading method for the air-ground integrated scenario according to claim 5 is characterized in that: If the task data amount of the target task is less than a first threshold, determining the user end as the execution end includes: Input the target task into a preset computing simulation model, and obtain a user-side execution time prediction value and a transmission time prediction value corresponding to the target task output by the computing simulation model, wherein the user-side execution time prediction value is used to predict the time required for the user-side to execute the target task, and the transmission time prediction value is used to predict the time required for the target task to be transmitted between the user-side and the satellite cluster side; If the task data volume of the target task is less than a first threshold, and the predicted value of the execution time of the user end is greater than the predicted value of the transmission time, the user end is determined to be the execution end.

7. The task collaborative offloading method for the air-ground integrated scenario according to claim 6 is characterized in that: The calculation simulation model calculates the predicted value of the user-side execution time by the following formula: in, Predict the execution time for the user side; is the computing power parameter of the user end; β is the preset computing power scaling factor; T wait The waiting delay of the target task determined according to the serial processing queue of the user end; is the data volume of the target task; And, the calculation simulation model calculates the transmission time prediction value by the following formula: In the above formula, is the transmission time prediction value, is the upload time for uploading the target task to the satellite cluster end, The time it takes for the satellite cluster to transmit the processing results corresponding to the target task back to the user end; Among them, upload time Determined by the following formula: In the above formula, is the data transmission rate between the user end and the satellite cluster end; n is the total number of tasks in the sending and receiving state between the user end and the satellite cluster end; Return time Determined by the following formula: In the above formula, μ is the ratio of the preset processing result data volume to the target task data volume.

8. The task collaborative offloading method for the air-ground integrated scenario according to claim 7 is characterized in that: The data transmission rate between the user end and the satellite cluster end Determined by the following formula: In the above formula, B is the communication bandwidth between the user end and the satellite cluster end, is the data upload capability parameter of the user end, N0 is the Gaussian white noise power, is the channel gain between the user end and the satellite cluster end; Among them, the channel gain between the user end and the satellite cluster end is Determined by the following formula: In the above formula, d m,n is the geographical distance between the user terminal and the satellite cluster terminal, and s is the preset path loss index.

9. A task collaborative unloading device for an air-ground integrated scenario, characterized in that: The device includes: a memory storing executable program code; a processor coupled to the memory; the processor calls the executable program code stored in the memory to execute the task collaborative unloading method for the air-ground integrated scenario as described in any one of claims 1-8.

10. A task collaborative offloading simulation system for an air-ground integrated scenario, characterized in that: The simulation system is used to simulate the task collaborative offloading method of the air-ground integrated scenario according to any one of claims 1 to 8, and the system includes: Computing platform model, used to model the user side, satellite cluster side and cloud server side; A target task model, used to model the target task generated by the user end; Computational model, used to model the transmission and computation process of the target task; A collaborative unloading model is used to model the unloading process of the target task, and obtain the execution time and energy consumption of the target task when the task collaborative unloading method of the air-ground integrated scenario according to any one of claims 1-8 is executed; The non-cooperative offloading model is used to model the process of target tasks being executed by the user end or being transmitted to the cloud server end through the satellite cluster end, and obtain the execution time and energy consumption of the target tasks at this time; A comparison model is used to compare the execution time and energy consumption output by the collaborative offloading model and the non-collaborative offloading model, so as to determine the effectiveness of the task collaborative offloading method for the air-ground integrated scenario described in any one of claims 1-8.

Images