A method for online collection of regional data

Abstract

The present invention discloses an online collection method for regional data. A value table for data providers is pre-established, a gridded target area is divided into a general area and a key area, data providers are selected for data collection, collection data of the target area is obtained, and the movement trajectory of the data providers is obtained. The general area coverage rate and the key area coverage rate of each data provider are calculated, and the current income and the maximum expected income in the future of each data provider are calculated based on the current income and the maximum expected income in the future of each data provider. The value of each data provider is calculated according to the current income and the maximum expected income in the future, and the pre-established value table is updated. The data providers of the next round are selected randomly or by searching the value table. The above process is repeated, and the collected data are integrated to complete the data collection of the target area. The method can solve the cold start problem in data collection and realize special and urgent data collection tasks.

Description

Technical Field

[0001] The invention relates to an online collection method for regional data, belonging to the technical field of data collection. Background Art

[0002] With the development of science and technology, sensor technology, communication technology and wireless transmission technology have also developed rapidly. In recent years, due to the explosive popularity of wireless intelligent terminal devices, wireless intelligent terminal devices have been widely used to perform sensing tasks, which has caused great attention to mobile crowd sensing technology.

[0003] Traditionally, data collection is often done by using specific large sensors to collect data in the target area, which is undoubtedly laborious and short-lived, and consumes manpower and financial resources. With the rise of mobile crowd sensing technology, it has brought new methods for collecting data in specific geographic areas and defined a new paradigm for regional data collection. Mobile personnel can intentionally or unintentionally reach a certain location through their own trajectory routes, and then collect and upload certain data of the area through sensors in mobile devices. The task publisher can obtain an overall perception map of the target area within a certain time and space range, which is more efficient and covers a wider range than traditional regional collection.

[0004] Existing data collection methods using crowd trajectories are basically based on data collection based on the historical trajectories of known data providers over a period of time. In some urgent data collection tasks, we hope to obtain data in the target area without any historical trajectories, and pay more attention to the data in the more key areas of the target area. Frequent collection of data in these areas can improve the reliability of data in these areas. At the same time, we also hope that the selected data providers can provide data in key areas of the target area while taking into account the coverage of the target area. How to quickly collect the above-mentioned data in the target area without the historical trajectory of the data provider needs to be solved.

[0005] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgement or any form of suggestion that the information constitutes the prior art already known to ordinary technicians in this field. Summary of the invention

[0006] The purpose of the present invention is to overcome the deficiencies in the prior art and provide an online collection method for regional data to solve the technical problem that in the prior art, data collection is performed using the historical trajectory of a data provider, and it is difficult to achieve rapid data collection in an emergency task without the historical trajectory of the data provider.

[0007] In order to solve the above technical problems, the present invention is implemented by adopting the following technical solutions:

[0008] An online regional data collection method, comprising:

[0009] Establish a value table for data providers in advance, grid the target area, and divide the target area into key areas and general areas;

[0010] According to the size of the target area, a specified number of data providers are randomly selected to collect data from the target area, and the collected data of the target area is obtained;

[0011] Obtain and calculate the key area coverage rate and general area coverage rate of each data provider based on the movement trajectory of the data provider in the target area;

[0012] Calculate the current revenue and the maximum expected revenue of each data provider in the future based on the coverage of key areas and general areas of the selected data providers;

[0013] Calculate the value of each data provider based on the current income and the maximum expected income in the future, and complete the update of the pre-established value table;

[0014] Randomly select or look up the value table to select the data provider for the next round of data collection in the target area;

[0015] Repeat multiple rounds of selecting data providers to collect data for the target area, integrate the data collected by multiple rounds of data providers, and complete the data collection for the target area.

[0016] As a preferred technical solution of the present invention, a method for calculating the regional coverage of the data provider includes:

[0017] Obtain the data collection time of each round of data providers, and divide the data collection time into multiple time slots of equal length;

[0018] Obtain the location information of the data provider in each time slot, obtain the number of times the data provider passes through the key area and general area in each round of data collection, and establish a trajectory statistical matrix;

[0019] The calculation of the general area coverage and key area coverage of the data provider is completed based on the trajectory statistical matrix.

[0020] As a preferred technical solution of the present invention, the calculation formula of the key area coverage rate of the data provider is:

[0021]

[0022] Among them, CD is the coverage rate of the mobile trajectory of the data provider in the key area; X is the trajectory statistical matrix based on which the coverage rate is calculated, and |X| is the number of rows in the trajectory statistical matrix based on which the coverage rate is calculated; is the data provider; m is the number of general area grids, n is the number of key area grids; j represents the column in the trajectory statistics matrix, and j from m+1 to m+n represents the key area grids represented by the last n columns in the trajectory statistics matrix; is the number of times the i-th data provider passes through the j-th grid;

[0023] The general regional coverage of the data provider is calculated as:

[0024]

[0025] Among them, CP is the coverage rate of the mobile trajectory of the data provider in the general area; j represents the column in the trajectory statistical matrix, j from 1 to m represents the general area grid represented by the first m columns in the trajectory statistical matrix, and j from m+1 to m+n represents the key area grid in the trajectory statistical matrix.

[0026] As a preferred technical solution of the present invention, the formula for calculating the current income of the data provider is:

[0027]

[0028] in, is the kth data provider in the set p; p is the set of selected data providers; For the selected data provider Current income; The data provider whose current revenue has been calculated and the next data provider whose current revenue is to be calculated A collection of; For collection The current earnings of data providers in; is the current income of the data providers in set S before calculating the current income of the next data provider;

[0029]

[0030] Among them, S is the set of data providers whose current returns have been calculated in this round; CP is the general area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S; CD is the key area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S;

[0031]

[0032] in, is the set of data providers whose current revenue has been calculated in this round; CP is the set The trajectory statistics matrix of the data provider is the general area coverage calculated based on; CD is the set The coverage of key areas is calculated based on the trajectory statistics matrix of the data provider.

[0033] As a preferred technical solution of the present invention, the maximum expected future benefit of the data provider is the highest value possessed by the remaining unselected data providers.

[0034] As a preferred technical solution of the present invention, the calculation formula of the data provider value is:

[0035]

[0036] Among them, lr is the learning rate; For washing in case of selecting S the original value of If you select S, select the current value of For the remaining data providers with the highest value; For choosing The highest value that the remaining data providers can provide in the case of

[0037] As a preferred technical solution of the present invention, in a round of updating of the value table, the order of calculating the values ​​of the data providers is consistent with the order of calculating the benefits of the data providers in this round.

[0038] As a preferred technical solution of the present invention, the method of selecting the data provider in a round of data provider selection includes:

[0039] Randomly generate a fixed value between 0 and 1 as the exploration probability;

[0040] Randomly generate a number of random numbers between 0 and 1 corresponding to the number of data providers;

[0041] Compare the random number with the exploration probability, and select from the value table or all data providers based on the comparison results;

[0042] Wherein, when the random number is less than the exploration probability, select from all data providers;

[0043] When the random number is greater than or equal to the exploration probability, the data provider is selected by looking up the value table.

[0044] As a preferred technical solution of the present invention, the probability of random selection from data providers who have participated in the data collection is 30%, and the probability of random selection from data providers who have not participated in the data collection is 70%.

[0045] As a preferred technical solution of the present invention, the method of selecting a data provider by searching a value table includes:

[0046] Find the corresponding row of the value table according to the selected data provider;

[0047] Determine the remaining optional data providers according to the corresponding rows of the value table;

[0048] The values ​​of the remaining optional data providers are compared, and the data provider with the highest value is selected to complete the selection of the data provider.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] By dividing the importance of the regions and using the trajectories of people on them to collect data in specific areas, data providers are selected to complete the task without any historical movement trajectories of data providers. This can solve the cold start problem in data collection and enable special and urgent data collection tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of the present invention;

[0052] Figure 2 It is a flow chart of selecting a data provider through a value table in the present invention. DETAILED DESCRIPTION

[0053] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0054] like Figure 1 and Figure 2 As shown, an online method for collecting regional data includes:

[0055] A value table for data providers is established in advance, the target area is gridded, and the target area is divided into key areas and general areas.

[0056] When the target area is gridded and divided into general areas and key areas, R is used to represent the set of all gridded areas. ,in, represents the grid of the general area, m represents the number of grids representing the general area, Represents the grid of the key area, n represents the number of grids in the key area, the pre-established initial value table is zero, each row of the table represents whether each data provider is selected to form a combination, and the column represents each data provider, that is, the value of each data provider is zero before data collection is performed.

[0057] According to the size of the target area, a specified number of data providers are randomly selected to collect data for the target area, and the collected data of the target area is obtained.

[0058] The number of data providers is determined according to the actual situation. P represents the set of selected data providers. , represents the kth data provider in P, and z represents the number of selected data providers.

[0059] In each round, data providers will collect data on the target area. The next round can obtain better coverage of the target area and key areas compared with the previous round. After multiple rounds of data collection, users can obtain reliable data on key areas and available general area data in the target area.

[0060] Obtain and calculate the key area coverage and general area coverage of each data provider based on the movement trajectory of the data provider in the target area.

[0061] The regional coverage calculation method for data providers includes:

[0062] Obtain the data collection time of each round of data providers, and divide the data collection time into multiple time slots of equal length;

[0063] Obtain the location information of the data provider in each time slot, obtain the number of times the data provider passes through the key area and general area in each round of data collection, and establish a trajectory statistical matrix;

[0064] The calculation of the general area coverage and key area coverage of the data provider is completed based on the trajectory statistical matrix.

[0065] Based on the key area coverage and general area coverage of the selected data providers, the current revenue and the maximum expected revenue in the future of each data provider are calculated.

[0066] The profit is calculated by the data providers in a round one by one in a random order, and the current profit of each data provider is calculated based on the current profit of the data provider before it.

[0067] S represents the set of data providers that have calculated the current returns in this round. The trajectories of data providers in S are represented by the trajectory matrix L(S). Since the trajectory is associated with the additional time, the time range of the trajectory is represented as equal-length time slots. , represents the kth time slot;

[0068]

[0069] Among them, |S| represents the number of data providers in S, ; represents the position of the i-th data provider in the j-th time slot, where i represents the i-th data provider in S and j represents the j-th time slot. Each row in the trajectory matrix L(S) represents the trajectory of each data provider in S, and each column represents the position of the data provider in S in a certain time slot.

[0070] According to the trajectory matrix, the statistical matrix V(S) can be obtained:

[0071]

[0072] Among them, |S| represents the number of data providers in S, Indicates the number of times the i-th data provider passes through the j-th grid (duplicates are not counted), where i represents the i-th data provider in S, and j represents the j-th grid. The first m columns of the statistical matrix represent the number of general area grids experienced by the data provider, and the last n columns represent the number of key area grids experienced. Each row of the statistical matrix V(S) represents the number of times the trajectory of each data provider in S passes through different grids, and each column represents which data providers in S have experienced a certain grid.

[0073] The formula for calculating the coverage rate of key areas according to the provider is:

[0074]

[0075] Among them, CD is the coverage rate of the mobile trajectory of the data provider in the key area; X is the trajectory statistical matrix based on which the coverage rate is calculated, and |X| is the number of rows in the trajectory statistical matrix based on which the coverage rate is calculated; is the data provider; m is the number of general area grids, n is the number of key area grids; j represents the column in the trajectory statistics matrix, and j from m+1 to m+n represents the key area grids represented by the last n columns in the trajectory statistics matrix; is the number of times the i-th data provider passes through the j-th grid;

[0076] The formula for calculating the general regional coverage of a data provider is:

[0077]

[0078] Among them, CD is the coverage rate of the mobile trajectory of the data provider in the key area; X is the trajectory statistical matrix based on which the coverage rate is calculated, and |X| is the number of rows in the trajectory statistical matrix based on which the coverage rate is calculated; is the data provider; m is the number of general area grids, n is the number of key area grids; j represents the column in the trajectory statistics matrix, and j from m+1 to m+n represents the key area grids represented by the last n columns in the trajectory statistics matrix; is the number of times the i-th data provider passes through the j-th grid;

[0079] The general regional coverage of the data provider is calculated as:

[0080]

[0081] Where CP is the coverage rate of the data provider's mobile trajectory in the general area; X is the trajectory statistical matrix based on which the coverage rate is calculated, and |X| is the number of rows in the trajectory statistical matrix based on which the coverage rate is calculated; is the data provider, representing the row in the trajectory statistics; m is the number of general area grids, n is the number of key area grids; j represents the column in the trajectory statistics matrix, j from 1 to m represents the general area grids represented by the first m columns in the trajectory statistics matrix, and j from m+1 to m+n represents the key area grids in the trajectory statistics matrix; The number of times the i-th data provider passes through the j-th grid.

[0082] Calculate the value of each data provider based on the current income and the maximum expected income in the future, and complete the update of the pre-established value table;

[0083] Data Provider The trajectory is given by the trajectory vector express, Get the statistics vector , the data provider The trajectory vector and statistical vector of are merged into the trajectory matrix and statistical matrix of the data provider whose returns have been calculated, and we get and , and calculate its current income. The formula for calculating the current data provider's income is: and calculate its current income. The formula for calculating the current data provider's income is:

[0084]

[0085] Among them, p is the set of selected data providers; is the kth data provider in set p; For the selected data provider Current income; The data provider whose current revenue has been calculated and the next data provider whose current revenue is to be calculated A collection of; For collection The current earnings of data providers in; is the current income of the data providers in set S before calculating the current income of the next data provider;

[0086]

[0087] Among them, S is the set of data providers whose current returns have been calculated in this round; is the current income of the data providers in the set S before calculating the current income of the next data provider; CP is the general area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S; CD is the key area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S; ; ;

[0088]

[0089] in, It is the set of data providers that have calculated the current returns in this round; For collection The current revenue of the data provider in The trajectory statistics matrix of the data provider is the general area coverage calculated based on; CD is the set The coverage of key areas is calculated based on the trajectory statistics matrix of the data provider; ; .

[0090] It should be noted that in technology and , the trajectory statistics matrices used for the calculation of CP and CD are different, where

[0091] Iterating the above calculation process can complete the calculation of the current income of all data providers in this round.

[0092] The value of a data provider is mainly composed of the data provider's current income and expectations of future income. The expectation of future income refers to the prediction of the impact that the selected data provider will have on the overall coverage and degree of data collection in the future.

[0093] The maximum expected future payoff of a data provider is the highest value among the remaining unselected data providers.

[0094] The calculation formula of data provider value is:

[0095]

[0096] Where S is the set of data providers whose current returns have been calculated; p is the set of selected data providers; is the kth data provider in set p; lr is the learning rate; The data provider whose current revenue has been calculated and the next data provider whose current revenue is to be calculated A collection of; For data providers Current income; For washing in case of selecting S the original value of If you select S, select the current value of For the remaining data providers with the highest value; For choosing The highest value that the remaining data providers can provide in the case of

[0097] It reflects the expectation of future returns, which is specifically expressed as In the case of Transformed into At this time, the data provider with the highest value among the remaining data providers is selected according to the value table , the value of this data provider is .

[0098] In a round of value table updates, the order in which the data providers' values ​​are calculated is consistent with the order in which the data providers' earnings are calculated in this round.

[0099] The updating of the value table includes putting the updated value into the value table. Each row of the value table indicates whether each data provider is selected to form a combination. Find the row corresponding to the set S in the value table and put the value into the first In the column grid, the above process is iterated, and the values ​​of the data providers are updated according to the calculation order of the values ​​of the data providers to complete the update of the value table.

[0100] It should be noted that the updated value table in each round will serve as the basis for selecting data providers in the next round.

[0101] Randomly select or look up the value table to select the data provider for the next round of data collection in the target area;

[0102] The selection methods of data providers in a round of data provider selection include:

[0103] Randomly generate a fixed value between 0 and 1 as the exploration probability;

[0104] Randomly generate a number of random numbers between 0 and 1 corresponding to the number of data providers;

[0105] Compare the random number with the exploration probability, and select from the value table or all data providers based on the comparison results;

[0106] Wherein, when the random number is less than the exploration probability, select from all data providers;

[0107] When the random number is greater than or equal to the exploration probability, the data provider is selected by looking up the value table.

[0108] The probability of random selection from data providers who have participated in the data collection is 30%, and the probability of random selection from data providers who have not participated in the data collection is 70%.

[0109] Methods for selecting data providers by looking up a value table include:

[0110] Find the corresponding row of the value table according to the selected data provider;

[0111] Determine the remaining optional data providers according to the corresponding rows of the value table;

[0112] The values ​​of the remaining optional data providers are compared, and the data provider with the highest value is selected to complete the selection of the data provider.

[0113] By selecting the data providers with the highest value, the coverage of key areas and general areas in the target area can be improved compared to the previous round, thereby completing the data collection of the target area.

[0114] Repeat multiple rounds of selecting data providers to collect data for the target area, integrate the data collected by multiple rounds of data providers, and complete the data collection for the target area.

[0115] Repeating multiple rounds of selecting data providers is a data iteration process to achieve efficient subsequent data collection in the target area, which can effectively improve the efficiency of data collection while avoiding cold start.

[0116] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for online collection of regional data, characterized in that: include: Establish a value table for data providers in advance, grid the target area, and divide the target area into key areas and general areas; According to the size of the target area, a specified number of data providers are randomly selected to collect data from the target area, and the collected data of the target area is obtained; Obtain and calculate the key area coverage rate and general area coverage rate of each data provider based on the movement trajectory of the data provider in the target area; Calculate the current revenue and the maximum expected revenue of each data provider in the future based on the coverage of key areas and general areas of the selected data providers; Calculate the value of each data provider based on the current income and the maximum expected income in the future, and complete the update of the pre-established value table; Randomly select or look up the value table to select the data provider for the next round of data collection in the target area; Repeat multiple rounds of selecting data providers to collect data for the target area, integrate the data collected by multiple rounds of data providers, and complete the data collection for the target area; The regional coverage calculation method for the data provider includes: Obtain the data collection time of each round of data providers, and divide the data collection time into multiple time slots of equal length; Obtain the location information of the data provider in each time slot, obtain the number of times the data provider passes through the key area and general area in each round of data collection, and establish a trajectory statistical matrix; The calculation of the general area coverage and key area coverage of the data provider is completed based on the trajectory statistical matrix.

2. The method for online collection of regional data according to claim 1, characterized in that: The formula for calculating the coverage rate of key areas of the data provider is: ； Among them, CD is the coverage rate of the mobile trajectory of the data provider in the key area; X is the trajectory statistical matrix based on which the coverage rate is calculated, and |X| is the number of rows in the trajectory statistical matrix based on which the coverage rate is calculated; is the data provider, representing the row in the trajectory statistics; m is the number of general area grids, n is the number of key area grids; j represents the column in the trajectory statistics matrix, j from m+1 to m+n represents the key area grids represented by the last n columns in the trajectory statistics matrix; is the number of times the i-th data provider passes through the j-th grid; The general regional coverage of the data provider is calculated as: ； Among them, CP is the coverage rate of the mobile trajectory of the data provider in the general area; j represents the column in the trajectory statistical matrix, j from 1 to m represents the general area grid represented by the first m columns in the trajectory statistical matrix, and j from m+1 to m+n represents the key area grid in the trajectory statistical matrix.

3. The method for online collection of regional data according to claim 2, characterized in that: The formula for calculating the current income of data providers is: ； in, is the kth data provider in the set p; p is the set of selected data providers; For the selected data provider Current income; The data provider whose current revenue has been calculated and the next data provider whose current revenue is to be calculated A collection of; For collection The current earnings of data providers in; is the current income of the data providers in set S before calculating the current income of the next data provider; ； Among them, S is the set of data providers whose current returns have been calculated in this round; CP is the general area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S; CD is the key area coverage rate calculated based on the trajectory statistical matrix of the data providers in the set S; ； in, is the set of data providers whose current revenue has been calculated in this round; CP is the set The trajectory statistics matrix of the data provider is the general area coverage calculated based on; CD is the set The coverage of key areas is calculated based on the trajectory statistics matrix of the data provider.

4. The method for online collection of regional data according to claim 1, characterized in that: The maximum expected future benefit of the data provider is the highest value possessed by the remaining unselected data providers.

5. The method for online collection of regional data according to claim 3 or 4, characterized in that: The calculation formula of data provider value is: ； Among them, lr is the learning rate; For washing in case of selecting S the original value of If you select S, select the current value of For the remaining data providers with the highest value; For choosing The highest value that the remaining data providers can provide in the case of 6. The method for online collection of regional data according to claim 5, characterized in that: In a round of updating of the value table, the order of calculating the values ​​of the data providers is consistent with the order of calculating the benefits of the data providers in this round.

7. The method for online collection of regional data according to claim 1, characterized in that: The method of selecting the data provider in a round of data provider selection includes: Randomly generate a fixed value between 0 and 1 as the exploration probability; Randomly generate a number of random numbers between 0 and 1 corresponding to the number of data providers; Compare the random number with the exploration probability, and select from the value table or all data providers based on the comparison results; Wherein, when the random number is less than the exploration probability, select from all data providers; When the random number is greater than or equal to the exploration probability, the data provider is selected by looking up the value table.

8. The method for online collection of regional data according to claim 7, characterized in that: The probability of random selection from data providers who have participated in the data collection is 30%, and the probability of random selection from data providers who have not participated in the data collection is 70%.

9. The method for online collection of regional data according to claim 7, characterized in that: Methods for selecting data providers by looking up a value table include: Find the corresponding row of the value table according to the selected data provider; Determine the remaining optional data providers according to the corresponding rows of the value table; The values ​​of the remaining optional data providers are compared, and the data provider with the highest value is selected to complete the selection of the data provider.

Images