A Crowdsourcing Spectrum Monitoring Method Based on Two-Stage User Selection

Abstract

The present invention discloses a crowdsourcing spectrum monitoring method based on two-stage user selection. Aiming at the problems of high construction cost and limited coverage of traditional radio spectrum monitoring methods, a crowdsensing incentive mechanism is introduced to achieve accurate and low-cost spectrum monitoring. The method constructs a system model based on the crowdsensing incentive mechanism, divides the user selection process into two stages according to whether users need to deliberately move to complete the sensing task, so as to recruit users beneficial to the platform. The method defines task priorities and joint probabilities, and establishes a platform utility function. In the first stage, a group of non-mobile users are selected from all users to complete the sensing task by the way during their daily life routes. In the second stage, a group of mobile users are selected from the remaining users to change their original routes and move to specific areas to complete the sensing tasks not completed by the non-mobile users. When the sensing budget is limited, the winning users are selected through two stages to maximize the number of tasks completed.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and particularly to a crowd-sourced spectrum monitoring method based on two-stage user selection. Background Art

[0002] With the rapid development of wireless communication technologies, higher requirements are put forward for the use and management of radio spectrum resources. As the carrier of information transmission in radio communication, radio spectrum resources play an irreplaceable role in today's information age. In recent years, the state has paid increasing attention to radio spectrum resources. As a national strategic resource, radio spectrum is uniformly allocated and used by the state. In life, the phenomenon of illegal use of radio spectrum occurs from time to time. Radio spectrum management is an effective method to solve the problem of non-compliant use of radio spectrum. Radio spectrum monitoring is an important part of radio spectrum management. It can provide a basis for radio management, so as to better manage radio spectrum resources. By analyzing the monitoring data, the occupancy situation of the spectrum can be known, and then the spectrum resources can be scientifically planned and reasonably configured to solve the congestion of spectrum resources and achieve the efficient and reasonable utilization of spectrum resources.

[0003] Traditional radio spectrum monitoring methods collect spectrum monitoring data by placing large and expensive monitoring stations. Due to the sparse deployment locations of the monitoring stations and limited monitoring ranges, to obtain accurate monitoring data, it is necessary to densely place monitoring devices. In this way, spectrum monitoring will undoubtedly lead to very high construction costs. Therefore, it is unrealistic to achieve accurate monitoring through traditional spectrum monitoring methods. In today's society, with the rapid growth of the application of intelligent terminals, using mobile crowd sensing to motivate users to participate in sensing tasks can obtain sensing data with a wide coverage range and reduce the cost of traditional radio spectrum monitoring.

[0004] Crowd sensing is a data acquisition mode that combines crowdsourcing and mobile sensing devices. It uses widely existing mobile devices (such as smart phones, wearable devices, tablet computers, etc.) to form an interactive and participatory sensing network. The sensing tasks are published to the participants in the network through the sensing platform to collect data information. The sensing data collected by crowd sensing has the advantages of multi-source heterogeneity, wide coverage range, high scalability, etc., and has received extensive attention from the academic and industrial circles at home and abroad and is applied in multiple fields, such as air quality monitoring, traffic monitoring, etc. Since participants need to consume communication resources such as the battery power, computing resources, and data traffic of their mobile devices to complete sensing tasks, and other costs, such as time, also need to be paid during the sensing process, without appropriate rewards, participants are not willing to participate in sensing activities gratuitously. The crowd sensing platform compensates the costs of sensing users by designing a reasonable incentive mechanism, motivates enough participants to actively join the sensing activities, improves the participation rate, and thus obtains accurate sensing data. Summary of the Invention

[0005] To overcome the deficiencies of the prior art, the objective of the present invention is to provide a crowdsourcing spectrum monitoring method based on two-stage user selection. The sensing platform recruits appropriate users to complete sensing tasks through two processes: non-mobile user selection and mobile user selection, thereby solving the problems of high economic cost and limited coverage of traditional radio spectrum monitoring methods.

[0006] Technical solution: The crowdsourcing spectrum monitoring method based on two-stage user selection according to the present invention is used in a radio spectrum monitoring scenario and includes:

[0007] Divide the sensing area into n sub-areas, and there are m candidate users participating in sensing to form a user set W. Define the activity of sensing the wireless signal strength of each sub-area published by the sensing platform as a sensing task, that is, there are n sensing tasks;

[0008] Users who can complete the sensing task without changing their routes in daily life are defined as non-mobile users; users who need to change their daily routes and move to specific sensing areas to complete the sensing task are defined as mobile users;

[0009] The users selected within the sensing time enter the corresponding sub-areas to complete the corresponding sensing tasks; the sensing platform pays a certain reward to the winning users who provide sensing results;

[0010] Under the condition that the total incentive reward does not exceed the budget, recruit winning users through two stages: non-mobile user selection and mobile user selection to maximize the number of completed tasks;

[0011] Merge the task sets completed in the first and second stages to obtain the total completed task set. The total completed task set is the set of the total visited sub-areas. Obtain the wireless signal strength of the visited sub-areas through the smart terminal devices carried by the winning users to complete the spectrum monitoring of the sub-areas.

[0012] Further, the method for non-mobile user selection in the first stage is as follows:

[0013] Initialize the non-mobile user set FW to be empty, calculate the utility brought by all users in the user set W to the sensing platform, and Umax represents the maximum utility and is initialized to 0;

[0014] Select the user in the user set W that brings the maximum utility to the sensing platform and has uncompleted tasks as the winning user and include it in the non-mobile user set. Each time a winning user is selected, Umax is updated to the utility of the winning user; the uncompleted tasks refer to the sub-areas that the user does not visit;

[0015] The sensing platform pays each non-mobile user the same fixed reward until the budget is insufficient or there are no eligible non-mobile users left; the condition means that the utility brought by the user to the sensing platform is greater than Umax and there are unfinished tasks.

[0016] Furthermore, for the selection of mobile users in the second stage, the method is as follows:

[0017] Initialize the set of mobile users SW to be empty, and select a group of winning candidate users from the remaining user set W\FW Complete the sensing tasks that have not been completed by non-mobile users by moving to a specific area; specifically:

[0018] Subtract the cost in the first stage from the total platform budget to get the remaining budget for recruiting users in the second stage; each remaining user willing to move submits a task-bid pair to the sensing platform;

[0019] Within the remaining budget, the sensing platform selects winning users by comparing the utility-bid ratios of each remaining user and incorporates the winning users into the set of mobile users SW;

[0020] The selected mobile users perform the sensing tasks and report the sensing results, and the sensing platform pays the selected participating users.

[0021] Furthermore, in the selection of non-mobile users in the first stage, select users whose utility brought to the sensing platform is greater than Umax and who have unfinished tasks as target users, and judge whether the budget B for the sensing tasks released by the platform is not less than the fixed reward I paid to the target users;

[0022] If B≥I holds, then incorporate the target users as winning users into the set of non-mobile users FW, incorporate the tasks completed by the winning users into the set of tasks completed in the first stage T c1 , and remove the winning users from the user set W, subtract the reward for recruiting the winning users from the budget B, that is, B = B - I; continue to calculate the utility brought by each user in the user set W to the platform until there are no users whose utility brought to the sensing platform is greater than Umax and who have unfinished tasks;

[0023] If B≥I does not hold, then the selection of non-mobile users in the first stage ends and enters the second stage.

[0024] Furthermore, in the selection of mobile users in the second stage, the users willing to move in the user set W\FW submit a task-bid pair to the sensing platform, calculate the utility-bid ratio and sort it, and select the user with the largest utility-bid ratio as the target user;

[0025] Judge whether the remaining budget B' of the platform is not less than the bid of the selected target user

[0026] If holds, then the target user w u is included as the winning user in the set SW of mobile users, and the tasks completed by the winning user are included in the set T of tasks completed in the second stage c2 , and the winning user is removed from the set W\FW of users, and the reward required to recruit the winning user is subtracted from the remaining budget B', that is

[0027] Continue to calculate the utility-bid ratio and select the user with the largest utility-bid ratio as the target user until does not hold, then the selection of mobile users in the second stage ends.

[0028] Furthermore, the users in the set W\FW who are willing to move submit a task-bid pair

[0029]

[0030] to the platform, where U(·) represents the platform utility, FW represents the set of winning users in the first stage, U(FW∪{w j ) represents the utility brought to the platform by recruiting user w j and FW; Umax represents the maximum utility, with an initial value of 0, and each time a non-mobile user is selected, Umax is updated to the utility of that non-mobile user; T c1 represents the set of tasks completed in the first stage, represents the set of tasks completed by user w j .

[0031] Furthermore, the users in the set W\FW who are willing to move submit a task-bid pair to the platform, where is the set of tasks that user w u wants to complete, is the bid price for all tasks in the set u of the perception tasks completed by user w , and the utility-bid ratio of user w u is expressed as:

[0032]

[0033] In the formula, U(SW∪{w u )-U(SW) represents the utility brought to the platform by recruiting user w j ;

[0034] Calculate the utility-bid ratio of all users in the set W\FW and sort them, and select the user with the largest utility-bid ratio as the target user.

[0035] Furthermore, the platform utility calculation method is as follows:

[0036]

[0037] Where Priority(t i ) represents the priority of the task , T represents the set of sensing tasks, and JointPro(t i , W f ) is the set of winning users W f 's joint probability of completing task t i ;

[0038] The priority calculation formula is:

[0039]

[0040] Where LE(t i ) represents the position entropy of task t i , and its calculation formula is:

[0041]

[0042] Where represents the set of candidate users for accessing task t i , represents 's total number of times that all candidate users access task t i , represents 's number of times that candidate user w u accesses task t i ;

[0043] The joint probability calculation formula is:

[0044]

[0045] Where represents the probability that user w u completes task t i , and its calculation formula is:

[0046]

[0047] Where represents the number of times that user w u appears in sub-region l i , that is, the number of times that user w u completes the sensing task t i ; The number of times that user w u appears in sub-region l i once means that the sensing task t iUpon completion, user w u appears in sub-region l i with probability k, that is, user w u completes sensing task t i with probability k is expressed as:

[0048]

[0049] User w u appears in sub-region l i at least once is expressed as:

[0050]

[0051] The utility brought to the platform by each recruited winning user is expressed as:

[0052]

[0053] Beneficial effects: Compared with the prior art, the significant advantages of the present invention are:

[0054] 1. The method of the present invention aims to maximize the number of task completions. By selecting appropriate users for tasks through two stages: non-mobile user selection and mobile user selection, it takes into account the advantages of both opportunistic sensing and participatory sensing modes, and maximizes the number of task completions at a lower cost.

[0055] 2. In the mobile user selection stage, a reverse auction is introduced. The sensing platform acts as the only buyer, and all users willing to move act as sellers. The sellers play a game, calculate the utility-bid ratio of all sellers and sort them, and the sellers with a larger ratio are recruited by the sensing platform as winners to complete the sensing task.

[0056] 3. Different from traditional spectrum detection methods, the method of the present invention introduces crowd sensing technology into the radio spectrum monitoring process, and combines an incentive mechanism to encourage participants to join the sensing task and upload sensing results, effectively saving the cost of traditional spectrum monitoring methods. Brief Description of the Drawings

[0057] Figure 1 is a schematic flow diagram of the method of the present invention;

[0058] Figure 2 is a schematic system model diagram of the method of the present invention;

[0059] Figure 3 is a simulation result diagram of the relationship between the budget and the number of tasks completed;

[0060] Figure 4 is a simulation result diagram of the relationship between each region and the number of task completions when the budget B = 15;

[0061] Figure 5 It is a simulation result graph showing the relationship between each area and the number of task completions when the budget B = 30;

[0062] Figure 6 It is the wireless signal strength of the spectrum monitoring sub - area. Specific implementation manners

[0063] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0064] The system model of the present invention is as Figure 2 , the scenario of the system is a square area with a side length of 400m×400m. The entire sensing area is divided into n sub - areas, and the activity of a user sensing the signal strength of each sub - area is called a sensing task. There are m candidate users participating in sensing distributed within the system scenario. Each candidate user can complete the sensing task when arriving at different sub - areas. In order to motivate the candidate users to sense the signal strength of the sub - areas, the sensing platform (i.e., the base station) will pay a certain reward to the winning users who provide the sensing results.

[0065] Suppose the entire sensing area is divided into n sub - areas of the same size, and the set of sub - areas is denoted as L = {l 1 , l 2 ,..., l i ,..., l n}, where l i represents the i - th sub - area. Defining the activity of sensing the wireless signal strength of each sub - area as a "task", that is, there are n sensing tasks, and the set of sensing tasks is denoted as T = {t 1 , t 2 ,..., t i ,..., t n}, where t i represents the i - th sensing task. Once the selected user enters the corresponding sub - area during the sensing time, the corresponding sensing task can be completed.

[0066] The objective of the present invention is to maximize the number of completed tasks by recruiting winning users in two stages under the condition that the total incentive reward does not exceed the budget.

[0067] The total set of candidate users is defined as W = {w 1 , w 2 ,..., w j ,..., w m}, where w jDenote the j-th candidate user. Each candidate user can complete the sensing task in two ways. One is opportunistic sensing, which can complete the sensing task by the way in their daily life without changing the route. The other is participatory sensing, which needs to change the daily life route and actively move to a specific sensing area to complete the sensing task. The selected user who completes the sensing task in the first way is called a non-mobile user, and the one who completes the sensing task in the second way is called a mobile user.

[0068] Based on the above description, the user selection problem of the present invention is formally defined as follows.

[0069] In the first stage, a set of winning candidate users is predicted through historical records to complete the sensing task, where FW is called the set of non-mobile users. Within the budget, the utility brought by the user to the platform is calculated through the defined utility formula. The users whose utility brought to the sensing platform is greater than Umax and who have unfinished tasks are selected as winning users and included in the set of non-mobile users. The sensing platform pays each non-mobile user the same fixed reward I until the budget is insufficient or there are no eligible non-mobile users.

[0070] In the second stage, a set of winning candidate users is selected from the remaining user set W\FW to complete the sensing tasks that have not been completed by non-mobile users so far by moving to specific areas, where SW is called the set of mobile users. Each remaining user willing to move submits a task-bid pair where is the set of tasks that user w u wants to complete, is the bid price for all tasks in the set of sensing tasks u completed by user w. The remaining budget obtained by subtracting the cost in the first stage from the total budget B is used to recruit users in the second stage. Within the remaining budget, the sensing platform selects winning users by comparing the utility-bid ratio of each remaining user and includes them in the set of mobile users. The selected mobile users execute the sensing tasks and report the sensing results, and the base station pays the selected participating users. Denote the set of completed tasks. For the platform, the optimization goal is to maximize the number of completed tasks under the condition that the total incentive reward does not exceed the total budget, that is, to maximize |T

[0071] |. The optimization problem of the present invention can be expressed as: c |.

[0072] Maximize|T c | (1)

[0073]

[0074] The selected user of the present invention appears in sub-region l i indicates that the sensing task t i is completed. Therefore, within the sensing time, it is necessary to predict the probability that each candidate user appears in different sub-regions at least once, that is, the probability that each candidate user completes different sensing tasks at least once. Count the number of times each candidate user w u appears in different sub-regions l i in several time periods in the historical record, and calculate the average value, which represents the number of times the candidate user w u appears in sub-region l i , denoted as Suppose obeys a non-homogeneous Poisson distribution. Then, within the sensing time, the probability that the candidate user w u appears in sub-region l i k times is:

[0075]

[0076] Therefore, the probability that the candidate user w u appears in sub-region l i at least once can be obtained as:

[0077]

[0078] Correspondingly, the probability that the candidate user w u completes task t i can be obtained as:

[0079]

[0080] Since the base station pays the same reward to the target users (non-mobile users) selected in the first stage, the sensing platform tends to select candidate users with as many visited sub-regions as possible. The number of times each candidate user visits different sub-regions is different, that is, the probabilities of different tasks being completed are different. The present invention defines the priority of each task according to the number of times users in different sub-regions are visited. Simply put, the more times a task is visited, the lower the priority assigned to this task, that is, the priority of the task is inversely proportional to the number of times the candidate user is visited. To further describe the task priority, the concept of location entropy is introduced. A task with more visits has a high location entropy (low priority), and conversely, a task with fewer visits has a low location entropy (high priority). The location entropy (Location Entropy, LE) of task t i is defined as:

[0081]

[0082] where Denote the access task as t i , the set of candidate users Denote the total number of times that all candidate users in i access the task t Denote the number of times that candidate user w u accesses the task t i .

[0083] Based on the concept of location entropy, the priority of each task is defined as:

[0084]

[0085] The platform utility is defined as:

[0086]

[0087] The utility brought to the platform by recruiting each winning user is defined as:

[0088]

[0089] where JointPro(t i ,W f ) is the joint probability that a set of selected users W f completes the task t i . According to formula (5), the probability that user w u completes the task t i is The joint probability that the task t i is completed by a set of selected users W f can be obtained as:

[0090]

[0091] Finally, as Figure 1 shown, the inventive method process based on the above optimization problem is as follows:

[0092] (1) Initialization: The set of sub-regions is represented by L = {l 1 ,l 2 ,...,l i ,...,l n}, the total number of sub-regions is n, the set of users is represented by W = {w 1 ,w 2 ,...,w j ,...,w m}, and the total number of candidate users is m;

[0093] (2) The base station issues sensing tasks: Sense the wireless signal strength of each sub-region. The set of sensing tasks is T = {t1 , t 2 ,..., t i ,..., t n} is used to represent that the total number of sensing tasks is n;

[0094] (3) The first stage (non-mobile user selection): FW represents the set of winning users in the first stage and is initially an empty set, T c1 represents the set of completed tasks in the first stage and is initially an empty set, Umax represents the maximum utility and is initially 0; for each non-mobile user selected, Umax is updated to the utility of that non-mobile user;

[0095] (4) Calculate the utility brought by each user in the user set W to the platform, and select users whose utility to the sensing platform is greater than Umax and who have unfinished tasks as target users;

[0096] (5) Judge whether the system budget B is not less than the fixed reward I distributed to the target users;

[0097] (5.1) If B≥I holds, then include the target users as winning users in the non-mobile user set FW, include the tasks completed by the winning users in the set of completed tasks T c1 , and remove the winning users from the user set W, subtract the reward for recruiting the winning users from the budget B, that is, B = B - I, and jump to step (4) to continue execution until there are no users whose utility to the sensing platform is greater than Umax and who have unfinished tasks;

[0098] (5.2) If B≥I does not hold, then the selection of non-mobile users in the first stage ends and the second stage is entered;

[0099] (6) The second stage (mobile user selection): B' represents the remaining system budget, T' represents the set of unfinished tasks, and B' and T' are obtained according to the situation of recruiting winning users in the first stage, that is, B' = B, T' = T - T c1 ;

[0100] (7) SW represents the set of winning users in the second stage and is initially an empty set, T c2 represents the set of completed tasks in the second stage and is initially an empty set;

[0101] (8) Mobile users in the user set W\FW who are willing to move submit a task - price pair to the sensing platform, calculate the utility - price ratio and sort it, and select the user with the largest utility - price ratio as the target user;

[0102] (9) Judge whether the remaining system budget B' is not less than the price of the target user

[0103] If If it holds, then target user w u is included as the winning user in the set of mobile users SW, and the tasks completed by the winning user are included in the set of tasks completed in the second stage T c2 , and the winning user is removed from the set of users W\FW, and the reward required to recruit the winning user is subtracted from the remaining budget B', that is jump to step (8) and continue to execute until does not hold, then the selection of mobile users in the second stage ends;

[0104] (10) Combine the sets of tasks completed in the first and second stages to obtain the total set of completed tasks, and the process of the system selecting the winning user ends. The total set of completed tasks is the total set of visited sub-regions. The wireless signal strength of the visited sub-region is obtained through the smart terminal device carried by the winning user to complete the spectrum monitoring of the sub-region.

[0105] In summary, the present invention proposes a crowd-sourced spectrum monitoring method based on two-stage user selection for the spectrum monitoring scenario, in combination with the crowd-sensing technology. This method constructs a system model based on the crowd-sensing incentive mechanism, divides the user selection process into two stages according to whether the user needs to move deliberately to complete the sensing task, in order to recruit users beneficial to the platform. The method of the present invention defines task priorities and joint probabilities, establishes a platform utility function. In the first stage, a group of non-mobile users are selected from the total set of users to complete the sensing task by the way in their daily life routes. In the second stage, a group of mobile users are selected from the remaining users to change their original routes and move to specific areas to complete the sensing tasks not completed by the non-mobile users. By selecting the winning users in two stages under the limited sensing budget, the number of completed tasks is maximized.

[0106] As Figure 3 shown, under different system budgets, the number of completed tasks of the crowd-sourced spectrum monitoring method based on two-stage user selection increases with the increase of the budget; as Figure 4 and Figure 5 shown, after the sensing ends, with the increase of the budget, the number of areas visited by users increases, and the number of times the area tasks are completed also increases; spectrum monitoring is reflected in whether there are users visiting the sub-region. If a sub-region is visited by a user, the spectrum monitoring of that sub-region is completed. As Figure 6 shown, after the sensing ends, the spectrum monitoring is the wireless signal strength of the different sub-regions sensed.

[0107] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted to include the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0108] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. A crowd-sourced spectrum monitoring method based on two-stage user selection, characterized in that, specifically including: Dividing the sensing area into n sub-areas, there are m candidate users participating in sensing to form a user set W, and defining the activity of sensing the wireless signal strength of each sub-area published by the sensing platform as a sensing task, that is, there are n sensing tasks; Users who can complete the sensing task without changing their routes in daily life are defined as non-mobile users; users who need to change their daily routes and move to a specific sensing area to complete the sensing task are defined as mobile users; The users selected within the sensing time enter the corresponding sub-areas to complete the corresponding sensing tasks; the sensing platform pays a certain reward to the winning users who provide sensing results; Under the condition that the total incentive reward does not exceed the budget, recruit winning users through two stages of non-mobile user selection and mobile user selection to maximize the number of completed tasks; Merge the task sets completed in the first and second stages to obtain the total completed task set. The total completed task set is the set of the total visited sub-areas. Obtain the wireless signal strength of the visited sub-areas through the smart terminal devices carried by the winning users to complete the spectrum monitoring of this sub-area; The first stage of non-mobile user selection, the method is as follows: Initialize the non-mobile user set FW to be empty, calculate the utility brought by all users in the user set W to the sensing platform, and Umax represents the maximum utility and is initialized to 0; Select the user in the user set W who brings the maximum utility to the sensing platform and has uncompleted tasks as the winning user and include it in the non-mobile user set. Each time a winning user is selected, Umax is updated to the utility of this non-mobile user; the uncompleted task mentioned refers to the sub-areas that the user does not visit; The sensing platform pays the same fixed reward to each non-mobile user until the budget is insufficient or there are no eligible non-mobile users; the condition refers to that the utility brought by the user to the sensing platform is greater than Umax and there are uncompleted tasks; The platform utility calculation method is: Among them, Priority(t i ) represents the priority of the task , T represents the set of perception tasks, JointPro(t i , W f ) is the set of winning users W f to complete the task t i 's joint probability; The priority calculation formula is: where \(LE(t\) i ) represents the location entropy of task \(t\) i , and its calculation formula is as follows: Among them represents the set of candidate users for accessing task t i , represents the total number of times that all candidate users in i access task t represents the number of times that candidate user w in u accesses task t i ; The joint probability calculation formula is: Among them represents the probability that user w u completes task t i The calculation formula is as follows: Among them represents the number of times user w u appears in sub-region l i , that is, the number of times user w u completes sensing task t i ; within the sensing time, the number of times user w u appears in sub-region l i once means that sensing task t i is completed. Then, the probability that user w u appears in sub-region l i k times, that is, the probability that user w u completes sensing task t i k times is expressed as: User w u appears in sub-region l i The probability of occurring at least once is expressed as: The utility brought by each recruited winning user to the platform is expressed as: The second stage of mobile user selection, the method is as follows: Initialize the set of mobile users SW to be empty, and select a group of winning candidate users from the remaining user set W\FW Complete the sensing tasks that have not been completed by non-mobile users by moving to specific areas; specifically: Subtract the cost of the first stage from the total platform budget to get the remaining budget for recruiting users in the second stage; each remaining user willing to move submits a task-bid pair to the sensing platform; Within the remaining budget range, the sensing platform selects winning users by comparing the utility-bid ratios of each remaining user and includes the winning users in the mobile user set SW; The selected mobile users execute the sensing tasks and report the sensing results, and the sensing platform pays rewards to the selected participating users.

2. The crowd-sourced spectrum monitoring method based on two-stage user selection according to claim 1, characterized in that, In the first stage of non-mobile user selection, select users whose utility brought to the sensing platform is greater than Umax and have uncompleted tasks as target users, and judge whether the budget B for the sensing tasks released by the platform is not less than the fixed reward I paid to the target users; If B≥I holds, the target user is included as the winning user in the non-mobile user set FW, and the tasks completed by the winning user are included in the first-phase completed task set T c1 , and the winning user is removed from the user set W, and the reward for recruiting the winning user is subtracted from the budget B, that is, B = B - I; continue to calculate the utility brought by each user in the user set W to the platform until no user brings a utility greater than Umax to the sensing platform and there are users with uncompleted tasks; If B≥I does not hold, the selection of non-mobile users in the first stage ends and enters the second stage.

3. The crowd-sourced spectrum monitoring method based on two-stage user selection according to claim 1, characterized in that, in the second-stage mobile user selection, the users in the user set W\FW who are willing to move submit a task-bid pair to the sensing platform, calculate the utility-bid ratio and sort it, and select the user with the largest utility-bid ratio as the target user; Determine whether the remaining budget B' of the platform is not less than the quote b of the selected target user wu ; If B'≥b wu holds, then the target user w u is included as the winning user in the mobile user set SW, and the tasks completed by the winning user are included in the second-phase completed task set T c2 , and the winning user is removed from the user set W\FW, and the reward required to recruit the winning user is subtracted from the remaining budget B', that is, B' = B' - b wu ; Continue to calculate the utility-offer ratio and select the user with the maximum utility-offer ratio as the target user until B'≥b wu If it does not hold, the selection of mobile users in the second stage ends.

4. The crowd-sourced spectrum monitoring method based on two-stage user selection according to claim 1, characterized in that, the selection of the target user in the first stage is screened by the following conditions: U(FW ∪ {w j}) > Umax && |T c1 ∪ T wj | > |T c1 | Among them, U(i) represents the platform utility, FW represents the set of users who win in the first stage, and U(FW ∪ {w j}) represents the utility brought by recruiting user w j and FW to the platform; Umax represents the maximum utility, with an initial value of 0. Each time a non-mobile user is selected, Umax is updated to the utility of that non-mobile user; T c1 represents the set of tasks completed in the first stage, and j represents the set of tasks completed by user w 5. The crowd-sourced spectrum monitoring method based on two-stage user selection according to claim 1, characterized in that, Users in the set \(W\setminus FW\) who are willing to move submit a task-bid pair to the platform where is the set of tasks that user \(w\) u wants to complete, is the bid price for all tasks in the set of sensing tasks that user \(w\) u completes, and the utility-bid ratio of user \(w\) is expressed as: u ​ where \(U(SW\cup\{w u \}) - U(SW)\) represents the utility brought by recruiting user \(w u \) to the platform; calculate the utility-bid ratio of all users in the user set W\FW and sort it, and select the user with the largest utility-bid ratio as the target user.

Images