A contract-incentivized heterogeneous data transmission method and device

By building a contract incentive mechanism and reinforcement learning model on drones, the bandwidth limitation and data quality problems of drones in crowdsourcing perception tasks are solved, and the upload of high-quality data and the meeting of delay requirements are achieved.

CN118450404BActive Publication Date: 2024-11-19BEIJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410874711.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-02
Publication Date
2024-11-19
Estimated Expiration
2044-07-02

AI Technical Summary

Technical Problem

When performing crowdsourcing perception tasks, drones face bandwidth limitations and unstable communication quality under high dynamics, resulting in data upload delays that cannot meet the needs. At the same time, the traditional linear pricing incentive mechanism is easy to attract malicious UAV to submit low-quality data, which affects the Task quality.

Method used

Using a heterogeneous data transmission method with contract incentives, we can build an optimization problem model by obtaining the configuration parameters of the drone cluster and the task parameters released by the task perception platform, and use the reinforcement learning model to determine the optimal contract to ensure that the drone is based on semantic compression factors. Extract key information and upload it, reduce bandwidth requirements, and incentivize the submission of high-quality data through reputation values.

Benefits of technology

It effectively improves the overall quality of crowdsourcing perception tasks, avoids concealment of low-quality data, meets the delay requirements of perception tasks, and reduces the demand for transmission bandwidth.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118450404B_ABST
    Figure CN118450404B_ABST
Patent Text Reader

Abstract

The present invention provides a contract-incentivized heterogeneous data transmission method and device, which relates to the technical field of communication. The method uses the reputation value of a drone as a recruitment reference item to recruit drones that can provide high-quality data for the current perception task, thereby avoiding the concealment of low-quality perception semantic data and effectively improving the overall quality of the crowdsourcing perception task. In addition, the contract between the task perception platform and the drone cluster is provided with a semantic compression factor for each type of drone when performing the current perception task and a task reward for each type of drone after completing the current perception task. Therefore, the drone performing the task can extract key semantic information from the original perception data based on the semantic compression factor in the optimal contract and then upload it to the task perception platform, reducing the demand for transmission bandwidth, thereby meeting the delay requirement of the current perception task.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of communications, and in particular to a contract-incentivized heterogeneous data transmission method and device. Background Art

[0002] Mobile Crowd Sensing (MCS) networks are emerging technologies that provide extensive sensing services by collecting heterogeneous data through rich sensors deployed on mobile user devices. The successful completion of MCS is highly dependent on the participation of mobile workers, but mobile workers often need to consume their limited perception and computing resources (such as memory, battery, and time) to complete crowdsourcing tasks. Therefore, in MCS, the perception platform needs to play the role of a management center, responsible for assigning tasks and paying rewards to mobile workers to compensate for the perception costs of mobile employees.

[0003] Unmanned Aircraft Vehicle (UAV) is an emerging technology with great potential in industrial scenarios and daily life, and it has the ability to provide a wide range of applications and services. Due to the rapid development of artificial intelligence (AI) technology, coupled with the flexible, responsive and low-cost characteristics of UAV itself, the unmanned swarm-assisted MCS system has become a new option, opening up new horizons for various smart city applications with perception needs.

[0004] However, UAVs face bandwidth limitations and unstable communication quality under high dynamics, and may not be able to meet the latency requirements of heterogeneous data uploads for perception tasks. In addition, the traditional linear pricing incentive mechanism provided to UAVs is likely to attract malicious UAVs to submit low-quality perception data or false information to defraud rewards, seriously affecting the overall quality of crowdsourcing perception tasks. Summary of the invention

[0005] The purpose of the present invention is to provide a contract-incentivized heterogeneous data transmission method and device to improve the overall quality of crowdsourced perception tasks, while reducing the demand for transmission bandwidth of drones and meeting the latency requirements of perception tasks.

[0006] In a first aspect, the present invention provides a contract-incentivized heterogeneous data transmission method, comprising: obtaining configuration parameters of a drone cluster and task parameters of a current perception task issued by a task perception platform; wherein the configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: maximum tolerable delay and minimum required data volume; the higher the reputation value of the drone, the greater the probability of executing the current perception task; based on the configuration parameters and the task parameters, constructing an optimization problem model of a contract between the task perception platform and the drone cluster; wherein the content of the contract includes: the semantic compression factor of each type of drone when executing the current perception task and the value of each drone after completing the current perception task The optimization goal of the optimization problem model is to maximize the utility of the task perception platform to the drone cluster under the constraints of data volume and drone utility; the optimization problem model is solved by using a reinforcement learning model to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster completes the current perception task based on the optimal contract; wherein the state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and a binary vector used to characterize whether the contract of the current time slot satisfies the drone utility constraint; the action of the agent is the contract of the next time slot; the immediate reward of the agent is the utility of the task perception platform to the drone cluster.

[0007] In an optional embodiment, obtaining the reputation value of each drone includes: obtaining the data quality of the perception semantic data uploaded by all drones participating in executing the target historical perception task; wherein the target historical perception task represents any task among all historical perception tasks in which the target drone participates; the target drone represents any drone in the drone cluster; calculating the average data quality of the target historical perception task based on the data quality; calculating the difference between the data quality of the target drone and the average data quality to obtain the target data quality difference of the target drone; determining the change amplitude of the reputation update factor after the target drone executes the target historical perception task based on the target data quality difference to obtain the target change amplitude; determining the reputation update factor of the target drone based on the target data quality difference and the target change amplitude after the target drone completes each historical perception task; and calculating the reputation value of the target drone based on the reputation update factor.

[0008] In an optional embodiment, based on the configuration parameters and the task parameters, an optimization problem model of the contract between the task perception platform and the drone cluster is constructed, including: constructing a time-consuming model for the target drone to complete the current perception task based on the amount of perception semantic data of the target class drone to which the target drone belongs, a first expression of the semantic compression factor, and the position data of the target drone; constructing a first binary variable model for characterizing whether the time-consuming model of the target drone meets the maximum tolerable delay; constructing a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to a preset reputation quantile; constructing a time-consuming model for the target drone based on the perception semantic data of the target class drone, a first expression of the semantic compression factor, and the position data of the target drone; constructing a first binary variable model for characterizing whether the time-consuming model of the target drone meets the maximum tolerable delay; constructing a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to the preset reputation quantile; constructing a time-consuming model for the target drone based on the perception semantic data of the target class drone, a first expression of the semantic compression factor, and the position data of the target drone ... first binary variable model for characterizing whether the time-consuming model of the target drone meets the maximum tolerable delay; constructing a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to the preset reputation quantile; constructing a second binary variable model for characterizing whether the reputation value of the target drone reaches the The utility model of the task perception platform for the target class UAV is constructed based on the amount of perceived semantic data of the target class UAV, the first binary variable model, the second binary variable model, the reputation value of the target UAV, the first expression of the semantic compression factor of the target class UAV and the second expression of the task reward; the utility model of the target class UAV is constructed based on the amount of perceived semantic data of the target class UAV, the first expression and the second expression; the optimization problem model is constructed based on the utility model of the task perception platform for the target class UAV, the utility model of the target class UAV, the amount of perceived semantic data of the target class UAV, the first expression and the minimum required data amount.

[0009] In an optional embodiment, based on the amount of perception semantic data of the target class drone to which the target drone belongs, a first expression of the semantic compression factor, and the position data of the target drone, a time-consuming model for the target drone to complete the current perception task is constructed, including: obtaining the workload and CPU operating frequency of the encoder and decoder in the full-size semantic extraction model; based on the workload, the CPU operating frequency, and the first expression of the semantic compression factor of the target class drone, constructing a first sub-time-consuming model for the target class drone to perform semantic extraction; obtaining the position data of the task perception platform, the bandwidth allocated to the target drone by the task perception platform, and the transmitter power of the target drone; rate; based on the position data of the target UAV and the position data of the task perception platform, determining the small-scale fading index of the target UAV; based on the bandwidth of the target UAV, the transmitter power and the small-scale fading index, calculating the data transmission rate of the target UAV uploading the perception semantic data to the task perception platform; based on the amount of perception semantic data of the target class UAV, the first expression of the semantic compression factor and the data transmission rate, constructing a second sub-time consumption model for the target UAV to upload the perception semantic data to the task perception platform; based on the first sub-time consumption model and the second sub-time consumption model, determining the time consumption model for the target UAV to complete the current perception task.

[0010] In an optional implementation, the utility model of the mission perception platform for the target class UAV is expressed as: ;in, represents the semantic compression factor of the i-th drone in the current perception task m’s contract, represents the task reward of the i-th drone in the current perception task m contract, represents the preset weight parameter, represents the first binary variable model, , represents the time-consuming model for the target UAV n to complete the current perception task m, represents the maximum tolerable delay of the current sensing task m, represents the second binary variable model, , represents the reputation value of the target drone n, Indicates the reputation value corresponding to the preset reputation quantile, represents the semantic fidelity of the i-th type of UAV performing the current perception task m, , , , , All represent preset model parameters. Represents the amount of perception semantic data of the i-th type of UAV.

[0011] In an optional embodiment, the utility model of the target type UAV is expressed as: ;in, represents the energy consumed by the i-th type of UAV to perform the current sensing task m, represents the hardware energy coefficient of the drone, represents the CPU operating frequency of the encoder in the full-size semantic extraction model, represents the workload of the encoder in the full-size semantic extraction model.

[0012] In an optional implementation, the optimization problem model is expressed as:

[0013]

[0014]

[0015]

[0016] ;

[0017] in, represents the minimum required data volume of the current perception task m, i≠i', represents the task reward of the i'th type of drone in the current perception task m contract, Represents the semantic compression factor of the i'th type of drone in the contract of the current perception task m.

[0018] In a second aspect, the present invention provides a contract-incentivized heterogeneous data transmission device, comprising: an acquisition module, used to acquire configuration parameters of a drone cluster and task parameters of a current perception task issued by a task perception platform; wherein the configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: maximum tolerable delay and minimum required data volume; the higher the reputation value of the drone, the greater the probability of executing the current perception task; a construction module, used to construct an optimization problem model of a contract between the task perception platform and the drone cluster based on the configuration parameters and the task parameters; wherein the content of the contract includes: the semantic compression factor of each type of drone when executing the current perception task and the probability of completing the current perception task The optimization goal of the optimization problem model is to maximize the utility of the task perception platform to the drone cluster under the constraints of data volume and drone utility. A solving module is used to solve the optimization problem model using a reinforcement learning model to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster completes the current perception task based on the optimal contract. The state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and a binary vector used to characterize whether the contract of the current time slot satisfies the drone utility constraint. The action of the agent is the contract of the next time slot. The immediate reward of the agent is the utility of the task perception platform to the drone cluster.

[0019] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, and when the processor executes the computer program, the steps of the heterogeneous data transmission method for contractual incentives described in any one of the aforementioned embodiments are implemented.

[0020] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the heterogeneous data transmission method for contractual incentives described in any of the aforementioned embodiments is implemented.

[0021] The present invention provides a contract-incentivized heterogeneous data transmission method, which uses the reputation value of drones as a recruitment reference item to recruit drones that can provide high-quality data for the current perception task, avoiding the concealment of low-quality perception semantic data, and effectively improving the overall quality of crowdsourcing perception tasks. In addition, the contract between the task perception platform and the drone cluster is provided with a semantic compression factor for each type of drone when performing the current perception task and a task reward for each type of drone after completing the current perception task. Therefore, the drones performing the task can extract key semantic information from the original perception data based on the semantic compression factor in the optimal contract and then upload it to the task perception platform, reducing the demand for transmission bandwidth, thereby meeting the delay requirements of the current perception task. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 A flowchart of a contract-incentivized heterogeneous data transmission method provided by an embodiment of the present invention;

[0024] Figure 2 A theoretical model diagram of a perception system consisting of a task perception platform and a drone cluster provided in an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of the interaction between a task perception platform and a drone cluster provided by an embodiment of the present invention;

[0026] Figure 4 A schematic diagram of a semantic extraction model architecture deployed on a perception UAV provided in an embodiment of the present invention;

[0027] Figure 5 A schematic diagram showing the comparison of the utility of the task perception platform between the method of the present invention and the linear pricing method;

[0028] Figure 6 A functional module diagram of a heterogeneous data transmission device for contract incentives provided by an embodiment of the present invention;

[0029] Figure 7 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions 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 part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0031] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0032] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0033] Drones will consume a lot of computing, energy and other resources when performing perception tasks. In the absence of appropriate incentives, drones as mobile workers may be reluctant to participate in crowdsourced perception tasks, which directly affects the quality and efficiency of task completion. Therefore, it is necessary to establish an effective incentive mechanism to give certain rewards to drones that complete perception tasks.

[0034] The linear pricing method is widely used as a mainstream incentive mechanism. Its purpose is to establish a linear relationship of y=kx+b between the contribution of workers and the benefits they can obtain, so as to motivate workers to make more contributions and obtain more rewards. However, this method may attract malicious UAVs to submit low-quality perception data or false information to defraud rewards. In addition, due to the high-speed mobility of UAVs and some harsh environmental factors, the quality of heterogeneous perception data may be affected. These subjective or objective factors may reduce the quality of heterogeneous data uploaded by some UAVs. Selecting these low-quality UAVs for perception tasks will not only waste corresponding resources, but also affect the overall quality of crowdsourcing perception tasks.

[0035] In addition, the large amount of multimedia heterogeneous data (such as images and videos) collected by UAVs faces the challenge of bandwidth limitations and unstable communication quality under high dynamics, which may not meet the requirements of heterogeneous data upload latency for perception tasks. In view of this, an embodiment of the present invention provides a contract-incentive heterogeneous data transmission method to alleviate the technical problems involved above.

[0036] Embodiment 1

[0037] Figure 1A flowchart of a contract-incentivized heterogeneous data transmission method provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method specifically comprises the following steps:

[0038] Step S102, obtaining the configuration parameters of the drone cluster and the task parameters of the current perception task released by the task perception platform.

[0039] Figure 2 A theoretical model diagram of a perception system consisting of a task perception platform and a UAV cluster provided in an embodiment of the present invention. The perception initiator sends a perception request to the task perception platform, and then the task perception platform issues a perception task to the UAV cluster based on the request. After the UAV cluster completes the perception task, the task perception platform feeds back the task data to the perception initiator.

[0040] In the embodiment of the present invention, before using the drone cluster to perform a perception task, it is first necessary to obtain its configuration parameters and the task parameters of the current perception task to be performed. The configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: the maximum tolerable delay and the minimum required data volume.

[0041] In practical applications, when the parameters of the sensors deployed on the drone are fixed, the amount of information data that the drone can perceive is the same. Therefore, the embodiment of the present invention is based on the amount of semantic data perceived by the drone. , the drones in the drone cluster are classified into Class I, , ,in, Represents the amount of perception semantic data of the i-th drone in the drone cluster.

[0042] Whenever a perception task is completed, the task perception platform will evaluate the reputation value of the perception UAV. Compared with the reasonable completion of the perception task, the unreasonable behavior of the perception UAV will have a greater impact on the final effect of the perception task, so the reputation value of the perception UAV will gradually increase after the perception behavior is reasonably completed; however, when an unreasonable behavior or low-quality perception is completed, the reputation value of the perception UAV will drop significantly. In order to avoid the concealment of low-quality perception semantic data, the embodiment of the present invention uses the reputation value of the drone as a recruitment reference item for the perception task. The higher the reputation value of the drone, the greater the probability of performing the current perception task, thereby providing effective protection for the overall quality of the crowdsourcing perception task.

[0043] When the task-aware platform When receiving a perception request, the task perception platform Publish crowdsourcing perception tasks to the perception UAV cluster In the network, , represents the total number of UAVs in the UAV cluster, Task-aware platform A collection of crowdsourced sensing tasks published in the UAV swarm network. Figure 3 As shown, for the mth perception task, the task perception platform provides a set of indicators ,in represents the maximum delay that the perception task m can tolerate, that is, the maximum tolerable delay, represents the minimum amount of data required for the perception task m, that is, the minimum required data amount, In order to motivate the perception UAV to participate in the perception task, the task perception platform A set of contracts published to the perception UAV cluster, ,in, represents the semantic compression factor of the i-th UAV when performing perception task m, , Indicates the task perception platform after completing the perception task m The purpose of the embodiment of the present invention is to provide a task perception platform with a task perception parameter based on the configuration parameters of the existing UAV cluster and the task parameters of the current perception task. Determine optimal contract with UAV cluster.

[0044] Step S104, based on the configuration parameters and the mission parameters, construct an optimization problem model of the contract between the mission perception platform and the drone cluster.

[0045] Among them, the contents of the contract include: the semantic compression factor of each type of drone when performing the current perception task and the task reward of each type of drone after completing the current perception task; the optimization goal of the optimization problem model is to maximize the utility of the task perception platform for the drone cluster under the constraints of data volume and drone utility.

[0046] Specifically, under the premise of known configuration parameters and mission parameters, the embodiment of the present invention takes the content of the contract as the unknown parameters to be optimized, the amount of data required for the perception mission and the utility of the drone as constraints, and maximizes the utility of the mission perception platform for the drone cluster as the goal, and constructs an optimization problem model of the contract between the mission perception platform and the drone cluster.

[0047] The semantic compression factor of each type of drone when performing the current perception task is set in the known contract The semantic compression factor is used to determine the semantic extraction capability of the perception UAV. In the embodiment of the present invention, a lightweight semantic extraction model is deployed on the UAV instead of a full-size semantic extraction model. represents the amount of perceptual semantic data of the i-th type of drone, then the full-size semantic extraction model undergoes semantic compression factor After lightweight processing, the size of the semantic information uploaded to the task perception platform when the i-th type of UAV completes the perception task m is In the case of applying a semantic compression factor, the embodiment of the present invention also uses semantic fidelity to quantify the semantic extraction capability of the semantic extraction model after being lightweighted by the compression factor.

[0048] The data volume constraint condition is that the sum of the data volumes uploaded by all UAVs participating in the current perception task is greater than the minimum required data volume of the current perception task; in order to truly reveal the fidelity of the semantic extraction model deployed by the UAV to prevent the occurrence of similar concealment, the embodiment of the present invention sets the constraints on the utility of the UAV according to the contract theory, including: individual rationality constraint (IR) and incentive compatibility constraint (IC).

[0049] Among them, the individual rationality constraint is specifically: the utility of the drone participating in the current perception task is non-negative. The incentive compatibility constraint is specifically: the drone will only get the maximum utility when it chooses a contract that meets its own type. The interpretation of the incentive compatibility constraint is as follows: For type UAV, choose the corresponding type of contract The utility of choosing other contracts is greater than or equal to The utility of the UAV is the maximum utility when it chooses a contract that matches its own type (i≠i'). The maximum here means that compared with the utility when choosing other types of contracts, the utility of the UAV is the maximum when choosing its own type of contract. The utility of the UAV is determined by the task reward given by the task perception platform and the energy consumption of completing the perception task.

[0050] In an embodiment of the present invention, the utility of the mission perception platform for the drone cluster is the sum of the utility of the mission perception platform for each type of drone, and the utility of the mission perception platform for the i-th type of drone is the sum of the satisfaction of the mission perception platform for each type of drone minus the reward given to the i-th type of drone by the mission perception platform, and the satisfaction is positively correlated with "the product of the drone's reputation value, semantic fidelity and the amount of perceived semantic data".

[0051] Step S106, using the reinforcement learning model to solve the optimization problem model, to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster can complete the current perception task based on the optimal contract.

[0052] Among them, the state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and the binary vector used to characterize whether the contract of the current time slot meets the utility constraint of the drone; the action of the agent is the contract of the next time slot; the immediate reward of the agent is the utility of the task perception platform for the drone cluster.

[0053] The embodiment of the present invention does not specifically limit the selection of reinforcement learning models. Users can choose according to actual conditions. For example, double-delay deep deterministic reinforcement learning can be used to design the optimal contract. First, the above optimization problem is established as a Markov decision process (MDP):

[0054] State space: According to the above introduction, there are IR and IC constraints in the contract design problem, so the state of the agent in time slot t is defined as ,in, represents the contract between the task perception platform and the UAV cluster in time slot t, is the semantic compression factor vector of time slot t, is the task reward vector given by the task perception platform to the drone cluster in the t-time slot, is a binary vector used to characterize whether the contract of the current time slot satisfies the drone utility constraint. , indicates whether the contract design of time slot t violates the IR constraint in the drone utility constraint, Indicates whether the contract design of time slot t violates the IC constraint in the drone utility constraint. Alternatively, if the i-th drone does not violate the IR constraint, then The value is 1; if violated, The value is 0. If the i-th type of drone does not violate the IC constraint, then The value is 1; if violated, The value is 0.

[0055] Action space: In the current perception task m, the task perception platform needs to design a set of contracts to motivate the drone to complete the perception task, the action of the agent in time slot t , that is, the contract for the next time slot.

[0056] Instant rewards: After the agent takes an action, the task-aware platform can immediately obtain an instant reward that maximizes its utility. , the instant reward at time t is the utility of the task perception platform for the drone cluster.

[0057] For the infinite state space and action space in the above MDP, the embodiment of the present invention can use a dual-delay deep deterministic reinforcement learning iterative solution including a real-time actor network and two real-time critic networks. Each iteration round will reset the crowdsourcing perception environment and the initial state. In this process, the task perception platform can obtain continuously optimized contract content. After completing the specified round of iterations, the optimal contract can be obtained.

[0058] Each iteration round includes a preset number of iteration steps. In any round, the iteration of the current round will be terminated in advance only when the IR and IC constraints of all perception UAVs are met, or the time reaches the maximum delay that the perception task can tolerate. Otherwise, the next round of iteration can only be entered after reaching the preset number of iteration steps. This design can try to ensure that the IR and IC constraints of the perception UAV are met in each iteration of successfully completing the perception task.

[0059] The embodiment of the present invention provides a contract-incentivized heterogeneous data transmission method, which uses the reputation value of drones as a recruitment reference item to recruit drones that can provide high-quality data for the current perception task, thereby avoiding the concealment of low-quality perception semantic data and effectively improving the overall quality of crowdsourcing perception tasks. In addition, the contract between the task perception platform and the drone cluster is provided with a semantic compression factor for each type of drone when performing the current perception task and a task reward for each type of drone after completing the current perception task. Therefore, the drones performing the task can extract key semantic information from the original perception data based on the semantic compression factor in the optimal contract and then upload it to the task perception platform, reducing the demand for transmission bandwidth, thereby meeting the latency requirements of the current perception task.

[0060] In an optional implementation, in the above step S102, obtaining the reputation value of each drone specifically includes the following steps:

[0061] Step S1021, obtaining the data quality of the perception semantic data uploaded by all drones participating in executing the target historical perception task; wherein the target historical perception task represents any task among all historical perception tasks participated in by the target drone; and the target drone represents any drone in the drone cluster.

[0062] Specifically, the quality of data uploaded by drone n after performing the kth historical perception task is calculated by the following formula: ;in, represents the original information collected by drone n after performing the kth historical perception task, , which means after compression factor A lightweight pair of encoders and decoders performs semantic extraction and restoration of information.

[0063] Step S1022, calculating the average data quality of the target history perception task based on the data quality.

[0064] After calculating the data quality of the perception semantic data uploaded by all drones participating in the target history perception task according to the above steps, the average data quality is taken as the average data quality of the target history perception task That is, if there are 5 UAVs participating in the target history perception task, then the average data quality of the 5 UAVs is the average data quality of the target history perception task.

[0065] Step S1023, calculating the difference between the data quality of the target UAV and the average data quality, and obtaining the target data quality difference of the target UAV.

[0066] Step S1024, determining the change range of the reputation update factor after the target UAV performs the target history perception task based on the target data quality difference, and obtaining the target change range.

[0067] Specifically, the target data quality difference of UAV n performing the kth historical perception task is ,when When UAV n performs the kth historical perception task, the change in the reputation update factor is ; When Δ When UAV n performs the kth historical perception task, the change in the reputation update factor is . and 0< < , indicating that the change in reputation when the data quality does not meet the standard is greater than the change in reputation when the quality is qualified. The value of is specifically limited, and users can set it according to actual needs.

[0068] Step S1025, determining the reputation update factor of the target UAV based on the target data quality difference and the target change amplitude after the target UAV performs each historical perception task.

[0069] In the embodiment of the present invention, the formula for the reputation update factor of drone n is: ; where y represents the total number of all historical perception tasks performed by the target UAV, represents the time attenuation weight, where ,It can be seen that compared with tasks completed in the past, the perceived tasks completed recently have a greater impact on reputation.

[0070] Step S1026, calculating the reputation value of the target drone based on the reputation update factor.

[0071] In the embodiment of the present invention, the reputation of the drone n is its reputation update factor The formula for the reputation value of drone n is: ; In the formula, except In addition, the remaining parameters are related to the preset parameters of the Richard curve. The Richard curve is a logical function widely used in UAV behavior modeling. is a parameter related to the asymptote, Related to the maximum rate of change of the asymptote, is the growth rate, Determines the maximum growth rate, initial reputation value With parameters related, .

[0072] In an optional implementation, the above step S104, based on the configuration parameters and the mission parameters, constructs an optimization problem model of the contract between the mission perception platform and the drone cluster, specifically including the following steps:

[0073] Step S1041, based on the amount of perception semantic data of the target class drone to which the target drone belongs, the first expression of the semantic compression factor and the position data of the target drone, a time-consuming model for the target drone to complete the current perception task is constructed.

[0074] Specifically, the time it takes for the target UAV to perform semantic extraction in the process of completing the current perception task is affected not only by the relevant hardware parameters of the encoder and decoder (the semantic model consists of two parts, the encoder and the decoder), but also by the semantic compression factor. After the semantic extraction is completed, the time it takes to transmit data to the task perception platform is related to the amount of perceived semantic data, the semantic compression factor, and its location. Therefore, based on the above data, a time-consuming model for the target UAV to complete the current perception task can be constructed.

[0075] Step S1042: construct a first binary variable model for characterizing whether the time consumption model of the target UAV satisfies the maximum tolerable delay.

[0076] Specifically, the first binary variable model is expressed as: ;in, It represents the time-consuming model for UAV n to complete the current perception task m, represents the maximum tolerable delay of the current perception task m. In other words, through the binary variable It can express whether UAV n can complete the current perception task within the maximum delay that the mission perception platform can tolerate.

[0077] Step S1043: construct a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to the preset reputation quantile.

[0078] Specifically, the first binary variable model is expressed as: ;in, represents the set of drone cluster reputations, express The preset reputation quantile of , represents the reputation value corresponding to the preset reputation quantile. For example, if ξ is 0.8, then It represents the reputation value of the top 80% of the drone cluster reputation in descending order. That is to say, when a perception task m is released, the task perception platform can determine whether to allow drone n to participate in the current perception task based on its reputation value.

[0079] Step S1044, based on the amount of perception semantic data of the target class UAV, the first binary variable model, the second binary variable model, the reputation value of the target UAV, the first expression of the semantic compression factor of the target class UAV and the second expression of the task reward, construct a utility model of the task perception platform for the target class UAV.

[0080] In an optional implementation, the utility model of the mission perception platform for the target type UAV is expressed as: .

[0081] in, represents the semantic compression factor of the i-th type of UAV in the contract of the current perception task m (that is, the first expression of the semantic compression factor of the i-th type of UAV), represents the task reward of the i-th drone in the current perception task m contract (that is, the second expression of the task reward of the i-th drone), represents the preset weight parameter, represents the first binary variable model, , represents the time-consuming model for the target UAV n to complete the current perception task m, represents the maximum tolerable delay of the current perception task m, represents the second binary variable model, , represents the reputation value of the target drone n, Indicates the reputation value corresponding to the preset reputation quantile, represents the semantic fidelity of the i-th type of UAV performing the current perception task m, , , , , All represent preset model parameters. Represents the amount of perception semantic data of the i-th type of UAV.

[0082] Step S1045: constructing a utility model of the target class UAV based on the amount of perception semantic data of the target class UAV, the first expression, and the second expression.

[0083] In an optional implementation, the utility model of the target class UAV is expressed as: .

[0084] in, represents the energy consumed by the i-th type of UAV to perform the current perception task m, represents the hardware energy coefficient of the drone, represents the CPU operating frequency of the encoder in the full-size semantic extraction model, Represents the workload of the encoder in the full-size semantic extraction model.

[0085] Step S1046, constructing an optimization problem model based on the utility model of the task perception platform for the target class UAV, the utility model of the target class UAV, the perception semantic data volume of the target class UAV, the first expression and the minimum required data volume.

[0086] Based on the macroscopic description of the optimization problem model in step S104 above, in an optional implementation manner, the optimization problem model is expressed as:

[0087]

[0088]

[0089]

[0090] .

[0091] in, represents the minimum required data volume of the current perception task m, i≠i', represents the task reward of the i'th type of drone in the current perception task m contract, Represents the semantic compression factor of the i'th type of drone in the contract of the current perception task m. Indicates data volume constraints; Represents IR constraints; Represents IC constraints.

[0092] In an optional implementation, the above step S1041, based on the amount of perception semantic data of the target class drone to which the target drone belongs, the first expression of the semantic compression factor and the position data of the target drone, constructs a time-consuming model for the target drone to complete the current perception task, specifically including the following steps:

[0093] Step S10411, obtaining the workload and CPU operating frequency of the encoder and decoder in the full-size semantic extraction model.

[0094] Step S10412, based on the workload, the CPU operating frequency and the first expression of the semantic compression factor of the target type UAV, construct a first sub-time-consuming model for the target type UAV to perform semantic extraction.

[0095] Specifically, the semantic model consists of two parts: an encoder and a decoder. Figure 4 As shown in Figure 1, ϕ and φ represent the semantic encoder deployed on the perception UAV and the semantic decoder deployed on the task perception platform, respectively. denote the workload of the encoder and decoder when using the full-size semantic extraction model, represents the size of semantic information extracted by the i-th type of UAV using the full-size semantic extraction model (i.e., the amount of perceived semantic data), and the full-size model is compressed by the semantic factor After lightweighting, the workload of the encoder and decoder and the size of semantic information are expressed as: .

[0096] For the semantic compression factor The first sub-time-consuming model for executing single-image semantic extraction (including semantic encoding on the drone side and semantic decoding on the task perception platform side) for a drone n is: ;in, Respectively represent the operating frequencies of the CPU at the encoder and decoder ends.

[0097] Step S10413, obtaining the location data of the mission perception platform, the bandwidth allocated by the mission perception platform to the target UAV, and the transmitter power of the target UAV.

[0098] Optionally, the mission perception platform may choose to evenly distribute the total bandwidth to each drone in the drone cluster, or may distribute it unevenly according to specified rules based on actual conditions. The embodiment of the present invention does not specifically limit the bandwidth allocation method. The transmitter power of the target drone is a fixed configuration of the drone and can be directly obtained from the configuration parameters.

[0099] Step S10414, determining the small-scale fading index of the target UAV based on the position data of the target UAV and the position data of the mission perception platform.

[0100] Specifically, when UAV n performs the mth perception task, UAV n and the task perception platform may communicate through line-of-sight communication (LoS) or non-line-of-sight communication (NLoS). The small-scale fading index of the target UAV is composed of the small-scale fading index component of line-of-sight communication Small-scale fading exponential component of non-line-of-sight communication Therefore, it is necessary to calculate the probability of UAV n communicating with the mission perception platform in LoS and NLoS and , + =1.

[0101] in, , a and b are environmental parameters that vary according to the communication environment, Indicates the height The elevation angle between the UAV n and the mission perception platform, represents the distance between UAV n and the mission perception platform.

[0102] ; ;in, Indicates the carrier frequency, represents the speed of light, and Represent the attenuation factor index of LoS and NLoS respectively, and Represent the path loss index of LoS and NLoS respectively, represents a random variable that follows a Ricean distribution, Represents a random variable that obeys the Rayleigh distribution. For different drones, the value of the random variable is different.

[0103] In the embodiment of the present invention, the small-scale fading index of drone n is calculated by the following formula: .

[0104] Step S10415, based on the bandwidth, transmitter power and small-scale fading index of the target UAV, calculate the data transmission rate at which the target UAV uploads the perception semantic data to the mission perception platform.

[0105] According to Shannon's formula, the data transmission rate of UAV n uploading the perception semantic data to the mission perception platform is: ;in, represents the bandwidth of drone n, represents the transmitter power of UAV n, represents the small-scale fading exponent of UAV n, represents the white noise variance.

[0106] Step S10416, based on the amount of perception semantic data of the target UAV, the first expression of the semantic compression factor and the data transmission rate, construct a second sub-time-consuming model for the target UAV to upload the perception semantic data to the task perception platform.

[0107] The second sub-time-consuming model of UAV n uploading the perception semantic data to the task perception platform is expressed as: ;in, The first expression representing the semantic compression factor of the i-th type of drone, represents the amount of perception semantic data of the i-th type of UAV, represents the amount of perception semantic data uploaded by the i-th type of UAV to the mission perception platform, It represents the data transmission rate when UAV n uploads the perception semantic data to the mission perception platform.

[0108] Step S10417, based on the first sub-time consumption model and the second sub-time consumption model, determine the time consumption model for the target UAV to complete the current perception task. .

[0109] In order to verify the effect of the solution provided by the embodiment of the present invention, the method of generating contracts in the embodiment of the present invention is compared with the linear pricing method. Figure 5 This is a schematic diagram comparing the utility of the task-aware platform of the method of the present invention and the linear pricing method. Obviously, through reinforcement learning training, the utility of the task-aware platform has been greatly improved compared to the linear pricing method.

[0110] In summary, the present invention extracts key semantic information from the original information collected by the UAV by deploying a semantic extraction algorithm on the UAV, and reduces the transmission bandwidth by reducing the transmission data to meet the latency requirements of the perception task. The present invention also designs an incentive mechanism based on contract theory for UAV crowdsourcing perception, which reveals the true semantic fidelity of the semantic extraction model deployed by the UAV, prevents the UAV from concealing its own semantic extraction model type in exchange for additional reward value, and considers the reputation value of the UAV while incentivizing the UAV to participate in crowdsourcing tasks. When recruiting mobile workers, the reputation value of the UAV is used as a reference item to recruit high-quality UAVs, thereby obtaining high-quality perception data.

[0111] Embodiment 2

[0112] An embodiment of the present invention also provides a contract-incentivized heterogeneous data transmission device, which is mainly used to execute the contract-incentivized heterogeneous data transmission method provided in the above-mentioned embodiment 1. The following is a specific introduction to the contract-incentivized heterogeneous data transmission device provided in the embodiment of the present invention.

[0113] Figure 6is a functional module diagram of a heterogeneous data transmission device for contractual incentives provided by an embodiment of the present invention, such as Figure 6 As shown, the device mainly includes: an acquisition module 10, a construction module 20, and a solution module 30, wherein:

[0114] The acquisition module 10 is used to obtain the configuration parameters of the drone cluster and the task parameters of the current perception task released by the task perception platform; wherein the configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: the maximum tolerable delay and the minimum required data volume; the higher the reputation value of the drone, the greater the probability of executing the current perception task.

[0115] Construction module 20 is used to construct an optimization problem model of the contract between the task perception platform and the drone cluster based on the configuration parameters and the task parameters; wherein the content of the contract includes: the semantic compression factor of each type of drone when performing the current perception task and the task reward of each type of drone after completing the current perception task; the optimization goal of the optimization problem model is to maximize the utility of the task perception platform for the drone cluster under the constraints of data volume and drone utility.

[0116] The solution module 30 is used to solve the optimization problem model using the reinforcement learning model to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster can complete the current perception task based on the optimal contract; wherein, the state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and a binary vector used to characterize whether the contract of the current time slot satisfies the drone utility constraint; the action of the agent is the contract of the next time slot; the immediate reward of the agent is the utility of the task perception platform for the drone cluster.

[0117] The embodiment of the present invention provides a contract-incentivized heterogeneous data transmission device, which uses the reputation value of drones as a recruitment reference item to recruit drones that can provide high-quality data for the current perception task, thereby avoiding the concealment of low-quality perception semantic data and effectively improving the overall quality of crowdsourcing perception tasks. In addition, the contract between the task perception platform and the drone cluster is provided with a semantic compression factor for each type of drone when performing the current perception task and a task reward for each type of drone after completing the current perception task. Therefore, the drone performing the task can extract key semantic information from the original perception data based on the semantic compression factor in the optimal contract and then upload it to the task perception platform, reducing the demand for transmission bandwidth, thereby meeting the latency requirement of the current perception task.

[0118] Optionally, the acquisition module 10 includes:

[0119] The acquisition unit is used to obtain the data quality of the perception semantic data uploaded by all drones participating in the target historical perception task; wherein the target historical perception task refers to any task among all the historical perception tasks participated in by the target drone; and the target drone refers to any drone in the drone cluster.

[0120] The first calculation unit is used to calculate the average data quality of the target history perception task based on the data quality.

[0121] The second calculation unit is used to calculate the difference between the data quality of the target UAV and the average data quality to obtain the target data quality difference of the target UAV.

[0122] The first determination unit is used to determine the change range of the reputation update factor after the target UAV performs the target history perception task based on the target data quality difference, so as to obtain the target change range.

[0123] The second determination unit is used to determine the reputation update factor of the target UAV based on the target data quality difference and the target change amplitude after the target UAV performs each historical perception task.

[0124] The third calculation unit is used to calculate the reputation value of the target drone based on the reputation update factor.

[0125] Optionally, the building block 20 includes:

[0126] The first construction unit is used to construct a time-consuming model for the target drone to complete the current perception task based on the amount of perception semantic data of the target class drone to which the target drone belongs, the first expression of the semantic compression factor, and the position data of the target drone.

[0127] The second construction unit is used to construct a first binary variable model for characterizing whether the time consumption model of the target UAV satisfies the maximum tolerable delay.

[0128] The third construction unit is used to construct a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to a preset reputation quantile.

[0129] The fourth construction unit is used to construct a utility model of the task perception platform for the target class UAV based on the perception semantic data volume of the target class UAV, the first binary variable model, the second binary variable model, the reputation value of the target UAV, the first expression of the semantic compression factor of the target class UAV and the second expression of the task reward.

[0130] The fifth construction unit is used to construct a utility model of the target class UAV based on the perception semantic data volume of the target class UAV, the first expression and the second expression.

[0131] The sixth construction unit is used to construct an optimization problem model based on the utility model of the task perception platform for the target type UAV, the utility model of the target type UAV, the perception semantic data volume of the target type UAV, the first expression and the minimum required data volume.

[0132] Optionally, the first building block is specifically used for:

[0133] Get the encoder and decoder workload and CPU frequency for the full-size semantic extraction model.

[0134] Based on the first expression of workload, CPU operating frequency and semantic compression factor of target UAV, the first sub-time-consuming model of semantic extraction of target UAV is constructed.

[0135] The location data of the mission awareness platform, the bandwidth allocated by the mission awareness platform to the target UAV, and the transmitter power of the target UAV are obtained.

[0136] Based on the position data of the target UAV and the position data of the mission perception platform, the small-scale fading index of the target UAV is determined.

[0137] Based on the bandwidth, transmitter power and small-scale fading index of the target UAV, the data transmission rate at which the target UAV uploads the perception semantic data to the mission perception platform is calculated.

[0138] Based on the amount of perceived semantic data of the target UAV, the first expression of the semantic compression factor and the data transmission rate, a second sub-time-consuming model for the target UAV to upload perceived semantic data to the task perception platform is constructed.

[0139] Based on the first sub-time consumption model and the second sub-time consumption model, the time consumption model for the target UAV to complete the current perception task is determined.

[0140] Optionally, the utility model of the mission perception platform for target UAVs is expressed as: .

[0141] in, represents the semantic compression factor of the i-th drone in the current perception task m’s contract, represents the task reward of the i-th drone in the current perception task m contract, represents the preset weight parameter, represents the first binary variable model, , represents the time-consuming model for the target UAV n to complete the current perception task m, represents the maximum tolerable delay of the current perception task m, represents the second binary variable model, , represents the reputation value of the target drone n, Indicates the reputation value corresponding to the preset reputation quantile, represents the semantic fidelity of the i-th type of UAV performing the current perception task m, , , , , All represent preset model parameters. Represents the amount of perception semantic data of the i-th type of UAV.

[0142] Optionally, the utility model of the target class UAV is expressed as: .

[0143] in, represents the energy consumed by the i-th type of UAV to perform the current perception task m, represents the hardware energy coefficient of the drone, represents the CPU operating frequency of the encoder in the full-size semantic extraction model, Represents the workload of the encoder in the full-size semantic extraction model.

[0144] Optionally, the optimization problem model is expressed as:

[0145]

[0146]

[0147]

[0148] .

[0149] in, represents the minimum required data volume of the current perception task m, i≠i', represents the task reward of the i'th type of drone in the current perception task m contract, Represents the semantic compression factor of the i'th type of drone in the contract of the current perception task m.

[0150] Embodiment 3

[0151] See also Figure 7 An embodiment of the present invention provides an electronic device, which includes: a processor 60, a memory 61, a bus 62 and a communication interface 63, wherein the processor 60, the communication interface 63 and the memory 61 are connected via the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.

[0152] The memory 61 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 63 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0153] The bus 62 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0154] Among them, the memory 61 is used to store programs, and the processor 60 executes the program after receiving the execution instruction. The method executed by the device defined by the process disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0155] The processor 60 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 60. The above processor 60 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 61, and the processor 60 reads the information in the memory 61 and completes the steps of the above method in combination with its hardware.

[0156] A computer program product of a contract-incentivized heterogeneous data transmission method and apparatus provided in an embodiment of the present invention includes a computer-readable storage medium storing a non-volatile program code executable by a processor, wherein the instructions included in the program code can be used to execute the method described in the previous method embodiment. For specific implementation, please refer to the method embodiment, which will not be repeated here.

[0157] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0158] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that can be executed by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., various media that can store program codes.

[0159] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0160] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0161] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0162] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions 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 they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A contract-incentivized heterogeneous data transmission method, It is characterized in that include: Acquire the configuration parameters of the drone cluster and the task parameters of the current perception task issued by the task perception platform; wherein the configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: maximum tolerable delay and minimum required data volume; the higher the reputation value of the drone, the greater the probability of executing the current perception task; Based on the configuration parameters and the task parameters, an optimization problem model of the contract between the task perception platform and the drone cluster is constructed; wherein the content of the contract includes: a semantic compression factor of each type of drone when performing the current perception task and a task reward for each type of drone after completing the current perception task; the optimization goal of the optimization problem model is to maximize the utility of the task perception platform for the drone cluster under the constraints of data volume and drone utility; The optimization problem model is solved by using a reinforcement learning model to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster completes the current perception task based on the optimal contract; wherein the state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and a binary vector used to characterize whether the contract of the current time slot satisfies the utility constraint of the drone; the action of the agent is the contract of the next time slot; and the immediate reward of the agent is the utility of the task perception platform to the drone cluster.

2. The contract-incentivized heterogeneous data transmission method according to claim 1, It is characterized in that Get the reputation value of each drone, including: Acquire the data quality of the perception semantic data uploaded by all drones participating in executing the target historical perception task; wherein the target historical perception task represents any task among all the historical perception tasks participated in by the target drone; and the target drone represents any drone in the drone cluster; Calculating the average data quality of the target history perception task based on the data quality; Calculating the difference between the data quality of the target UAV and the average data quality to obtain a target data quality difference of the target UAV; Determine the change range of the reputation update factor after the target UAV performs the target history perception task based on the target data quality difference, and obtain the target change range; Determining a reputation update factor of the target UAV based on the target data quality difference and the target change amplitude after the target UAV performs each historical perception task; The reputation value of the target drone is calculated based on the reputation update factor.

3. The contract-incentivized heterogeneous data transmission method according to claim 1, It is characterized in that Based on the configuration parameters and the mission parameters, an optimization problem model of the contract between the mission perception platform and the drone cluster is constructed, including: Based on the amount of perception semantic data of the target class drone to which the target drone belongs, the first expression of the semantic compression factor, and the position data of the target drone, a time-consuming model for the target drone to complete the current perception task is constructed; Constructing a first binary variable model for characterizing whether the time consumption model of the target UAV satisfies the maximum tolerable delay; Constructing a second binary variable model for characterizing whether the reputation value of the target drone reaches the reputation value corresponding to a preset reputation quantile; Based on the amount of perception semantic data of the target drone, the first binary variable model, the second binary variable model, the reputation value of the target drone, the first expression of the semantic compression factor of the target drone, and the second expression of the task reward, construct a utility model of the task perception platform for the target drone; Building a utility model of the target class UAV based on the amount of perception semantic data of the target class UAV, the first expression, and the second expression; The optimization problem model is constructed based on the utility model of the task perception platform for the target class UAV, the utility model of the target class UAV, the perception semantic data volume of the target class UAV, the first expression and the minimum required data volume.

4. The contract-incentivized heterogeneous data transmission method according to claim 3, It is characterized in that Based on the amount of perception semantic data of the target class drone to which the target drone belongs, the first expression of the semantic compression factor, and the position data of the target drone, a time-consuming model for the target drone to complete the current perception task is constructed, including: Get the workload and CPU frequency of the encoder and decoder in the full-size semantic extraction model; Based on the workload, the CPU operating frequency and a first expression of the semantic compression factor of the target class UAV, a first sub-time-consuming model for executing semantic extraction of the target class UAV is constructed; Acquire the location data of the mission sensing platform, the bandwidth allocated by the mission sensing platform to the target UAV, and the transmitter power of the target UAV; Determining a small-scale fading index of the target UAV based on the position data of the target UAV and the position data of the mission sensing platform; Based on the bandwidth of the target UAV, the transmitter power and the small-scale fading index, calculating the data transmission rate at which the target UAV uploads the perception semantic data to the mission perception platform; Based on the amount of perception semantic data of the target drone, the first expression of the semantic compression factor and the data transmission rate, construct a second sub-time-consuming model for the target drone to upload the perception semantic data to the task perception platform; Based on the first sub-time consumption model and the second sub-time consumption model, a time consumption model for the target UAV to complete the current perception task is determined.

5. The contract-incentivized heterogeneous data transmission method according to claim 3, It is characterized in that The utility model of the mission perception platform for the target UAV is expressed as: ; in, represents the semantic compression factor of the i-th drone in the current perception task m’s contract, represents the task reward of the i-th drone in the current perception task m contract, represents the preset weight parameter, represents the first binary variable model, , represents the time-consuming model for the target UAV n to complete the current perception task m, represents the maximum tolerable delay of the current sensing task m, represents the second binary variable model, , represents the reputation value of the target drone n, Indicates the reputation value corresponding to the preset reputation quantile, represents the set of drone cluster reputations, express The preset reputation quantile of , represents the semantic fidelity of the i-th type of UAV performing the current perception task m, , , , , All represent preset model parameters. Represents the amount of perception semantic data of the i-th type of UAV.

6. The contract-incentivized heterogeneous data transmission method according to claim 5, It is characterized in that The utility model of the target class UAV is expressed as: ; in, represents the energy consumed by the i-th type of UAV to perform the current sensing task m, represents the hardware energy coefficient of the drone, represents the CPU operating frequency of the encoder in the full-size semantic extraction model, represents the workload of the encoder in the full-size semantic extraction model.

7. The contract-incentivized heterogeneous data transmission method according to claim 6, It is characterized in that The optimization problem model is expressed as: ; in, represents the minimum required data volume of the current perception task m, i≠i', represents the task reward of the i'th type of drone in the current perception task m contract, Represents the semantic compression factor of the i'th type of UAV in the contract of the current perception task m.

8. A contract-incentivized heterogeneous data transmission device, It is characterized in that include: An acquisition module is used to acquire the configuration parameters of the drone cluster and the task parameters of the current perception task issued by the task perception platform; wherein the configuration parameters include: the amount of perception semantic data of each type of drone in the drone cluster, the reputation value and location data of each drone; the task parameters include: maximum tolerable delay and minimum required data volume; the higher the reputation value of the drone, the greater the probability of executing the current perception task; A construction module is used to construct an optimization problem model of a contract between the task perception platform and the drone cluster based on the configuration parameters and the task parameters; wherein the content of the contract includes: a semantic compression factor of each type of drone when performing the current perception task and a task reward of each type of drone after completing the current perception task; the optimization goal of the optimization problem model is to maximize the utility of the task perception platform for the drone cluster under the constraints of data volume and drone utility; A solution module is used to solve the optimization problem model using a reinforcement learning model to obtain the optimal contract between the task perception platform and the drone cluster, so that the drone cluster completes the current perception task based on the optimal contract; wherein the state of the agent in the current time slot in the reinforcement learning model includes: the contract of the current time slot and a binary vector used to characterize whether the contract of the current time slot satisfies the utility constraint of the drone; the action of the agent is the contract of the next time slot; and the immediate reward of the agent is the utility of the task perception platform to the drone cluster.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor. It is characterized in that When the processor executes the computer program, the steps of the contract-incentivized heterogeneous data transmission method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, which, when executed by a processor, implement the heterogeneous data transmission method for contractual incentives as described in any one of claims 1 to 7.