User selection method in mobile crowd-sensing system

Abstract

Disclosed in the present invention is a user selection method in a mobile crowd-sensing system. The method comprises: on the basis of the distance between the location of each user and a task region location, acquiring a mobility capability and a historical task capability of the user, and performing calculation on the mobility capability and the historical task capability of the user, so as to obtain an effectiveness capability of the user; acquiring a direct interaction evaluation and an indirect recommendation evaluation for each user, and performing calculation on the direct interaction evaluation and the indirect recommendation evaluation, so as to obtain a global credit of the user; performing calculation on the effectiveness capability and the global credit, so as to obtain a comprehensive indicator of each user; and selecting a user on the basis of the comprehensive indicators of the users. A user selection solution provided in the present invention can select a high-quality candidate user, who has great real-time performance and reliable data, as a worker for uploading sensing data in a sensing task, and can not only defend against unreliable data aggregation in a mobile network, but can also meet the requirement of a task releaser for acquiring effective data, thereby achieving the aim of healthy and sustainable development of a mobile crowd-sensing system.

Description

A user selection method in a mobile crowd-sensing system Technical Field

[0001] The present invention belongs to the field of crowd sensing, and in particular relates to a user selection method in a mobile crowd sensing system. Background Art

[0002] Mobile crowdsensing has become a widely used perception paradigm for collecting large amounts of data. Mobile users utilize embedded sensors in their mobile devices (such as cameras, microphones, GPS, and gyroscopes) to collect information about human activity and the surrounding environment. This perception data is then uploaded to cloud servers for further analysis and sharing. However, some tasks in crowdsensing systems require sensory data that cannot tolerate latency. Data collected beyond this time limit is meaningless and can even severely impact the overall effectiveness of the task.

[0003] Furthermore, before users upload their sensory data, the perception platform is unaware of the quality of the data they provide. Users can easily upload false sensory data and incomplete information, leading to deception on the task requester's part, while receiving compensation that is disproportionate to the quality of their service. Over time, task requesters will be less willing to offer appropriate rewards to obtain data, while users who can provide reliable and valid data will gradually dwindle. Furthermore, dishonest users still have the potential to engage in malicious behavior during the task, such as launching poisoning attacks or intentionally delaying computations. This can lead to task failure and significant losses.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a user selection method in a mobile crowd sensing system to solve the problems existing in the above-mentioned prior art.

[0006] To achieve the above object, the present invention provides a user selection method in a mobile crowdsensing system, comprising:

[0007] Obtaining the user's mobility and historical task capabilities based on the distance between the user's location and the task area, and calculating the user's mobility and historical task capabilities to obtain the user's effective capability;

[0008] Obtaining the user's direct interaction evaluation and indirect recommendation evaluation, calculating the user's direct interaction evaluation and indirect recommendation evaluation, and obtaining the user's global credit;

[0009] The effective capability of the user and the global credit of the user are calculated to obtain a comprehensive index of the user, and the user is selected based on the comprehensive index of the user.

[0010] Preferably, the process of obtaining the user's mobility capability includes:

[0011] Obtaining the user's location and the task area location at the same time, calculating the user's location and the task area location based on the Euclidean formula to obtain the distance between the user's location and the task area location;

[0012] The distance between the user's location and the task area and the user's average speed are calculated to obtain the user's mobility.

[0013] Preferably, the process of obtaining the historical task capability includes:

[0014] Get the completed tasks and task submission time in the historical task list maintained by the user;

[0015] The completed tasks and the task submission time are calculated based on the average time slot of the user to obtain the historical task capacity.

[0016] Preferably, the expression for obtaining the user's effective capability is:

[0017] Among them, P i For effectiveness capability, C i For historical mission capability, A i is the mobility capability, N is the total number of users required, n is the nth user, C n is the historical task capability of the nth user.

[0018] Preferably, the direct interaction evaluation is calculated based on the direct interaction evaluation before the task, the total number of interactions between the platform and the worker so far, the time sensitivity factor, and the interval between the mth interaction and the current Mth interaction;

[0019] The indirect recommendation evaluation is calculated based on the user recommendation evaluation, the recommender set, and the personal credit factor.

[0020] Preferably, the expression for obtaining the user's global credit is: R i =δDR i +(1-δ)IR i ;

[0021] Among them, R i For global credit, IR i For indirect recommendation evaluation, DR i is a direct interaction evaluation, and δ is the evaluation factor parameter.

[0022] Preferably, the expression for obtaining the user's comprehensive index is: S i =ln(1+R i *Pi );

[0023] Among them S i It is a comprehensive indicator.

[0024] The technical effects of the present invention are:

[0025] The user selection scheme proposed in the present invention can select high-quality user candidates with strong real-time performance and reliable data as workers who upload perception data in perception tasks. It can not only resist unreliable data aggregation in mobile networks, but also meet the requirements of task publishers to obtain valid data, thereby achieving the goal of healthy and sustainable development of mobile crowd intelligence perception systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0027] FIG1 is a flow chart of a user selection method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0030] Example 1

[0031] As shown in FIG1 , this embodiment provides a user selection method in a mobile crowdsensing system, including:

[0032] Mobile crowdsensing tasks are characterized by timeliness. The effectiveness of perception data decreases with the time it is submitted, and perception data obtained after the task deadline is ineffective. Therefore, the system should impose requirements on the time users can participate in perception tasks to ensure the validity of perception data. Furthermore, to prevent malicious behavior from compromising data quality, the system should ensure that users with a history of dishonest behavior or a high probability of dishonest behavior are not selected for the task. Therefore, this paper uses comprehensive metrics to evaluate user service quality. Task publishers use these comprehensive metrics to select users for reliable perception tasks.

[0033] Since the effectiveness of sensory data decreases as the duration of sensory tasks increases and sensory tasks must be performed within a specified area, certain requirements are placed on the user's mobility. If the user has strong mobility and can reach the designated location in a relatively short time, the quality of the sensory data collected will be high. However, if the user has poor mobility, it will take longer to reach the sensory location, resulting in less usable sensory data.

[0034] In this embodiment, it is assumed that user U i View Perception Task W i The position at which the i =(l i .x,l i .y). Only when the user moves to the specific area of ​​task perception can the perception task be performed. Therefore, in this embodiment, d is used to represent the distance between the user's current location and the nearest location point r in the perception area, which is calculated according to the Euclidean formula, as shown in formula (1). d(l i ,g i )=[(l i .xg i .x) 2 +(l i .yg i .y) 2 ] 1 / 2 (1)

[0035] Assume that the average moving speed of users is fixed and the same, expressed as v i , then the user's mobility capability A i Calculated by formula (2). i =d(l i ,g i ) / v i (2)

[0036] Furthermore, a user's historical task capability can also reflect their actual efficiency in performing perception tasks. A user may have previously participated in perception tasks multiple times, accumulating feedback from each performance, ultimately generating historical task capability. Stronger historical task capabilities indicate a higher average efficiency in collecting perception data and a more stable performance level.

[0037] Each user maintains a historical task list, which records all previously completed tasks and the time when the corresponding tasks were submitted, expressed as in Indicates that the user performs task W nThis embodiment assumes that the perception task is completed when the user finishes uploading the perception data (data transmission time is negligible). There may be a time gap between the user's task submission time and the task deadline. This embodiment uses the user's average time gap to evaluate its historical task capability C. i , calculated by formula (3).

[0038] Where K represents the total number of tasks in the historical task list.

[0039] User's effectiveness capability P i It is determined by the mobility capability and historical mission capability and is calculated by formula (4).

[0040] Where N represents the participation in the perception task W j The total number of people, and They are the perception tasks W j When the user is new, the history task list is empty.

[0041] Dishonest users may exist in the system. These users qualify for sensing tasks through fraudulent means. However, they perform these tasks inefficiently, with errors in their actual sensing locations, resulting in delayed sensing and failure to upload sensing data within the specified timeframe. This results in the actual quality of the uploaded sensing data not matching the expected quality, reducing overall data availability and causing significant losses for the platform and task publishers. Therefore, evaluating user service quality also requires considering both user honesty and the reliability of their sensing data. This paper uses global credit to measure these two aspects.

[0042] A participant's global credit is composed of direct interaction evaluation and indirect recommendation evaluation. Direct interaction evaluation is the result of the requester scoring the participant based on the quality of the perceived data, and is direct feedback on the quality of the perceived data. Indirect recommendation evaluation comes from the recommendations of other participants in the system.

[0043] For user U i The direct interaction evaluation of this task is expressed as DR i , the direct interaction evaluation before this task is expressed as tr i , calculated by formula (5).

[0044] Where M represents the total number of interactions between the platform and the worker so far. Considering that feedback with a short interval from the current evaluation should be better than feedback with a long interval, ε k Represents the time-sensitive factor, which is used to adjust the weight of each feedback with time attributes. T M -T m is the interval between the mth interaction and the current Mth interaction.

[0045] User U i The indirect recommendation evaluation of this task is represented by IR i , user U j For user U i The recommendation evaluation is represented as ir i,j , calculated by formula (6).

[0046] Where S represents the recommender set, which consists of other users who participated in this task, j∈S. Considering that the feedback from users with high credit is more reliable than that from users with low credit, σ i Represents the personal credit factor, which serves as the weight of the recommender's feedback. where R j Indicates the recommender U j Global credit.

[0047] User U i The global credit is represented by R i , calculated by formula (7). i =δDR i +(1-δ)IR i (7)

[0048] Effectiveness and global credit jointly determine the comprehensive index, expressed as S i , used to measure the user's service quality. The stronger the user's effectiveness ability and the higher the global credit, the better the effectiveness and credibility of the perception data can be guaranteed. Therefore, the better the participant's service quality, the greater the comprehensive index value. Effectiveness ability P i and global credit R i Respectively and comprehensive index S i Positively correlated, and the effectiveness capability P i and global credit R i The product of can make the comprehensive index S i The value change trend of is more obvious, which can ensure the clear display of the change of service quality. i When the value of is large, S i The growth space of S becomes smaller, so i The growth rate of S i There is a slight increase near the maximum value. According to the above change law, the comprehensive index is calculated by formula (8). i =ln(1+R i *P i ) (8)

[0049] After calculating the comprehensive indicators in the above steps, the task publisher can select high-quality user candidates with strong real-time performance and reliable data as workers who upload perception data in the perception task.

[0050] The above description is merely a preferred embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A user selection method in a mobile crowd intelligence perception system, characterized in that: The following steps are involved: The user's mobility and historical task capabilities are obtained based on the distance between the user's location and the task area, and the user's mobility and historical task capabilities are calculated to obtain the user's effective capabilities; Obtaining the user's direct interaction evaluation and indirect recommendation evaluation, calculating the user's direct interaction evaluation and indirect recommendation evaluation, and obtaining the user's global credit; The effective capability of the user and the global credit of the user are calculated to obtain a comprehensive index of the user, and the user is selected based on the comprehensive index of the user.

2. The user selection method in the mobile crowd intelligence perception system according to claim 1, characterized in that: The process of obtaining the user's mobility capability includes: Obtaining the user's location and the task area location at the same time, calculating the user's location and the task area location based on the Euclidean formula to obtain the distance between the user's location and the task area location; The distance between the location of the user and the location of the task area and the average speed of the user are calculated to obtain the mobility of the user.

3. The user selection method in the mobile crowd intelligence perception system according to claim 1, characterized in that: The process of acquiring the historical mission capability includes: Get the completed tasks and task submission time in the historical task list maintained by the user; The completed tasks and task submission times are calculated based on the average time slot of the user to obtain historical task capabilities.

4. The user selection method in the mobile crowd intelligence perception system according to claim 2, characterized in that: The expression for obtaining the user's effective capability is: Among them, P i For effective capability, C i A is the historical mission capability. i is the mobility capability, N is the total number of users required, n is the nth user, C n is the historical task capability of the nth user.

5. The user selection method in the mobile crowd intelligence perception system according to claim 1, characterized in that: The direct interaction evaluation is calculated based on the direct interaction evaluation before the task, the total number of interactions between the platform and the worker so far, the time sensitivity factor, and the interval between the mth interaction and the current Mth interaction; The indirect recommendation evaluation is calculated based on the user recommendation evaluation, the recommender set, and the personal credit factor.

6. The user selection method in the mobile crowd intelligence perception system according to claim 1, characterized in that: The expression for obtaining the user's global credit is: i =δDR i +(1-δ)IR i ; Among them, R i For global credit, IR i For indirect recommendation evaluation, DR i is a direct interaction evaluation, and δ is the evaluation factor parameter.

7. The user selection method in the mobile crowd intelligence perception system according to claim 1, characterized in that: The expression for obtaining the user's comprehensive index is: S i =ln(1+R i *P i ); Where S i It is a comprehensive indicator.

Images