Air pollution exposure alarm system based on mobile phone positioning

Abstract

The invention belongs to the field of air pollution detection and alarm, and discloses an air pollution exposure alarm system based on mobile phone positioning, and an air quality index prediction method based on mobile phone positioning is characterized by comprising the following steps: collecting a user report position; if an air pollution monitoring point exists in a preset radius range of the reporting position, reporting data of the monitoring point as a result, otherwise, searching for the air pollution monitoring point with a four-vector meeting a threshold value requirement in a preset maximum limiting radius by using a four-vector method; calculating the air quality index of the reporting position by using an interpolation method; if the air pollution monitoring point position with the four-direction vector meeting the threshold value requirement cannot be found within the preset maximum limiting radius, a user is prompted to modify the reporting position. According to the method, the air quality of the positioning position is more accurately predicted by using a linear interpolation method, so that a user can accurately obtain the air quality data of the positioning position of a mobile phone.

Description

Technical Field

[0001] The present invention belongs to the field of air pollution detection and alarm, and more specifically, relates to an air pollution exposure alarm system based on mobile phone positioning. Background Art

[0002] Air pollution has been a matter of great concern in recent years. A growing body of research confirms its link to disease and can even trigger the onset of certain conditions, such as chronic obstructive pulmonary disease and asthma. For vulnerable populations, rapid monitoring of changes in air quality is crucial. Currently, data from local authorities is the primary method for monitoring air quality. This method is not always timely and is not sensitive to sudden changes in environmental data. Summary of the Invention

[0003] In view of the above defects or improvement needs of the prior art, the purpose of the present invention is to provide an air pollution exposure alarm system based on mobile phone positioning. Compared with the traditional method, the present invention uses a linear interpolation method to make a more accurate prediction of the air quality of the positioning location, so that the user can accurately obtain the air quality data of the mobile phone positioning location. The present invention uses multi-site data fusion to analyze the environment around the mobile phone positioning in real time, and issue a timely alarm when the air quality changes suddenly. In addition, the present invention can further establish a real-time correction system for the air quality data of the precise coordinate point based on the real-time acquisition system of the unmanned vehicle (or drone) and the deep regression model, so that the user can more accurately obtain the air quality data of the mobile phone positioning location.

[0004] To achieve the above object, according to one aspect of the present invention, a method for predicting air quality index based on mobile phone positioning is provided, which is characterized by comprising the following steps:

[0005] (1) collecting user-reported locations;

[0006] (II) If there is an air pollution monitoring point within the preset radius of the reported location, then the data of the monitoring point is used as the air quality index of the reported location, and the process jumps to step (V);

[0007] If there is no air pollution monitoring point within the preset radius of the reported location, then proceed to step (III);

[0008] (III) Select the two air pollution monitoring points closest to the reported location, record them as A and B respectively, take the reported location as the origin O, and obtain the vector The unit vector of ; then, calculate The projection size of the unit vector in the east, west, south and north directions is respectively recorded as OA ,West OA ,SouthOA ,north OA ; Calculate at the same time The projection size of the unit vector in the east, west, south and north directions is recorded as east OB ,West OB ,South OB ,north OB Among them, East OA Hexi OA At least one of them is 0, OA ,north OA At least one of them is 0, East OB ,West OB At least one of them is 0, OB ,north OB At least one of is 0;

[0009] (IV) If max{East OA ,East OB}≥0.5, max{West OA ,West OB}≥0.5, max{South OA ,South OB}≥0.5, max{North OA ,north OB If the four inequalities}≥0.5 are simultaneously established, the air quality index at point O is calculated based on the air quality indexes monitored at A and B using the interpolation method, and the process jumps to step (V);

[0010] Otherwise, take point O as the center and find another air pollution monitoring point C within the preset maximum limit radius so that: max{east OA ,East OB ,East OC}≥0.5, max{West OA ,West OB ,West OC}≥0.5, max{South OA ,South OB ,South OC}≥0.5, max{North OA ,north OB ,north OC}≥0.5, these four inequalities hold true at the same time; among them, OC ,West OC ,South OC ,north OC for The projection size of the unit vector in the east, west, south and north directions, east OC Hexi OC At least one of them is 0, OC ,north OCAt least one of them is 0; then, the air quality index at point O is calculated based on the air quality indexes monitored by A, B, and C using the interpolation method, and the process goes to step (V);

[0011] If C cannot be found within the preset maximum limit radius, the user is prompted to modify the reported location and the method is exited;

[0012] (V) Sending the air quality index of the reported location obtained in step (II) or the air quality index of point O obtained in step (IV) to the user's mobile phone.

[0013] As a further preferred embodiment of the present invention, in step (IV), the air quality index at point O is calculated based on the air quality index monitored by A and B based on the interpolation method, specifically: the air quality indexes monitored by A and B are AQI A , AQI B ,

[0014] If points A, O, and B are collinear, then the air quality index AQI at point O is O satisfy:

[0015]

[0016] Among them, |AB| represents the distance between points A and B, and |OA| represents the distance between points O and A;

[0017] If points A, O, and B are not collinear, then

[0018]

[0019] in, |OA| represents the distance between points O and A, and |OB| represents the distance between points O and B.

[0020] As a further preferred embodiment of the present invention, in step (IV), the air quality index at point O is calculated based on the air quality index obtained by monitoring A, B, and C based on the interpolation method, specifically: the air quality indexes obtained by monitoring A, B, and C are AQI A , AQI B , AQI C ,

[0021] First, calculate the air quality index AQI at point D based on the interpolation method D , point D is the intersection of the line AB and the line OC:

[0022]

[0023] Among them, |AB| represents the distance between points A and B, and |DA| represents the distance between points D and A;

[0024] Then, based on the interpolation method, the air quality index AQI at point O is calculated O :

[0025]

[0026] Among them, |CD| represents the distance between points C and D, and |OC| represents the distance between points O and C.

[0027] As a further preferred embodiment of the present invention, in step (II), the preset radius is 500m.

[0028] As a further preferred embodiment of the present invention, in step (IV), the preset maximum restriction radius is 5 km.

[0029] As a further preferred embodiment of the present invention, step (V) further includes: when the air quality index of the reported location or the air quality index at point O exceeds a preset warning value, issuing an air pollution exposure alarm via the user's mobile phone;

[0030] Preferably, the air pollution monitoring points are used to monitor the concentrations of air pollutants PM2.5, PM10 and / or SO2.

[0031] As a further preference of the present invention, step (V) further includes data correction processing:

[0032] The real air quality index of the reported location is collected by unmanned vehicles and / or drones, recorded as AQI real ; The air quality index of the reported location obtained in step (II) is recorded as AQI calc ,

[0033] like No correction is required;

[0034] like Then, adjust the preset radius range and jump to step (II);

[0035] The air quality index at point O calculated in step (IV) is recorded as AQI calc ,

[0036] like No correction is required;

[0037] like Then, AQI calc Input the trained regression model and use the AQI output by the model adj As the modified O-point air quality index;

[0038] The regression model is obtained by collecting multiple groups of AQIreal with AQI calc The samples are trained, and each AQI in these samples is calc with AQI real One to one correspondence.

[0039] As a further preference of the present invention, the regression model is a linear regression model or a random forest regression model.

[0040] According to another aspect of the present invention, the present invention provides an air quality index prediction system based on mobile phone positioning, including a memory and a processor, wherein the memory stores a computer program, and is characterized in that the processor implements the steps of the above method when executing the computer program.

[0041] According to another aspect of the present invention, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0042] Compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0043] (1) Improve data accuracy and real-time performance: By combining mobile phone positioning technology and linear interpolation methods, the system can provide users with more accurate air quality predictions at their location.

[0044] The present invention can enhance user experience, allowing users to quickly and conveniently query the air quality conditions at their current location through their mobile phones, thus meeting the public's growing demand for healthy living and environmental protection.

[0045] The air quality index prediction method of this invention is based on a four-dimensional vector threshold of 0.5. This means that the location information of a monitoring point in a particular direction is considered useful only if the projection magnitude in that direction is at least 0.5. This threshold balances computational efficiency and prediction accuracy, helping to select the monitoring points that have the greatest impact on the final AQI estimate while eliminating information sources that may introduce significant errors.

[0046] (2) The present invention can provide an intelligent error judgment and correction mechanism to ensure that the information obtained by users is reliable and enhance the user's trust and satisfaction. Especially in areas where there is a lack of direct monitoring data, based on the present invention, the mobile monitoring capabilities of unmanned vehicles (or drones) can be preferably utilized to achieve dynamic data collection and real-time correction, significantly improving the accuracy and timeliness of air quality data. Taking unmanned vehicles as an example, unmanned vehicles can dynamically collect data in areas where fixed stations cannot cover or data is insufficient, and feed these real-time data back to the system, so that methods such as linear interpolation can be used to make predictions based on more comprehensive and up-to-date information. For those cases where linear interpolation still cannot achieve sufficient accuracy, the predicted value can be further corrected by a deep regression model. For example, random forest regression or neural network models can adjust the predicted value based on multiple features (such as time, geographical location, meteorological conditions, etc.) to make it closer to the actual situation. The system of the present invention is flexible in design and can be adjusted according to the specific needs and geographical characteristics of different cities. Whether it is a city center or a remote area, the effective coverage of the monitoring network can be guaranteed by adjusting the layout of monitoring sites and the patrol strategy of unmanned vehicles. Furthermore, the model can be flexibly updated. For example, as more high-quality data accumulates, the regression model can be regularly retrained and optimized to ensure it maintains optimal performance. In particular, the present invention can employ linear regression or random forest regression models to refine predictions. Regularly updating the model to reflect the latest environmental trends effectively ensures long-term prediction accuracy.

[0047] (3) In addition, the present invention can collect a large amount of real-time and accurate air quality data, providing valuable first-hand information for environmental protection departments, which is conducive to policy makers making more scientific and reasonable environmental protection decisions, effectively addressing air pollution problems, and protecting public health. The present invention integrates multiple advanced technologies such as the Internet of Things, big data, and artificial intelligence, demonstrating the innovative application of "Internet + Internet of Things" in the field of environmental protection, providing a reference model for the design and implementation of similar environmental monitoring systems, and has a strong demonstration effect and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the flow of the air pollution exposure warning system of the present invention.

[0049] Figure 2 This is a flow chart of a method for calculating the air quality at a location using a mobile phone in the air pollution exposure warning system of the present invention.

[0050] Figure 3 Schematic diagram of the real-time update steps of the interpolation method based on unmanned vehicles. DETAILED DESCRIPTION

[0051] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0052] Example 1

[0053] In general, if Figure 1 As shown, the air pollution exposure warning system based on mobile phone positioning of the present invention is based on the user's reported location obtained by the user's mobile phone positioning, first judging whether it is within a preset fixed radius R0 ( Figure 1 In the example, R0 is preset to 500 meters; of course, R0 can be adjusted according to the actual situation) Whether there are monitoring points (such as monitoring stations):

[0054] (a) If it exists, the air quality index (AQI) obtained by monitoring the monitoring point will be fed back to the user's mobile phone as the air quality index of the user's reported location. That is to say, when using the data service, the user's mobile phone will report the latitude and longitude information of its current location. After obtaining the user's latitude and longitude information, it will first confirm whether there is a collection site within 500 meters of the user. If there is a collection site, the data of the site can be directly returned, which is the air pollution data obtained by the user based on the mobile phone positioning information. Regarding the preset fixed radius R0, it can be adjusted and pre-set from the following perspectives (that is, in order to realize the process of the system confirming whether there is a collection site within 500 meters after the user's mobile phone reports its location, and adjusting this distance range according to the strategy, the following rules can be used to decide how to select the most appropriate air quality monitoring data):

[0055] (a1) Fixed radius search: A fixed search radius (for example, 500 meters) is set by default to find the nearest air quality monitoring station around the user. The reason why 500m is selected as the preset radius in this embodiment is based on a comprehensive consideration of the urban structure, population density, and the distribution of air quality monitoring stations. A smaller radius can provide more accurate location-related data, but may limit the number of data sources; while a larger radius increases the amount of available data, but may result in a decrease in accuracy. Therefore, the present invention selects 500 meters as the initial search radius by analyzing the data accuracy and availability in different scenarios.

[0056] Or: (a2) Dynamically adjust the search radius: If no monitoring station is found within the default radius, the search radius can be gradually expanded until at least one monitoring station is found or the maximum limit (such as 5 kilometers) is reached. Conversely, if it is found that there are too many monitoring stations within the default radius, the radius can be narrowed to reduce the amount of calculation and increase the response speed. The reason why the preset maximum limit radius is set to 5 kilometers in this embodiment is that it can maximize the data coverage while ensuring the system response speed. If it is set too small, it may not be possible to find enough monitoring points for effective interpolation calculations; conversely, if it is set too large, it will increase unnecessary computational complexity and may introduce greater errors.

[0057] Alternatively: (a3) ​​Density-based adjustment: Predetermine the optimal search radius for different areas based on factors such as urban population density and traffic flow. For example, a smaller radius can be used in high-density areas such as the city center, while a wider radius can be used in suburban areas.

[0058] Alternatively: (a4) Historical data analysis: Use historical data to assess the effectiveness of monitoring station distribution within a specific area. For areas where alarm failures or inaccurate data are common, consider adding new fixed monitoring stations or optimizing the patrol routes of unmanned vehicles.

[0059] (a1)-(a4) are four methods for determining the search radius, and several strategies can be used simultaneously. If all four methods for determining the search radius fail, start step (b):

[0060] (b) If it does not exist, calculate it using the four-square vector method.

[0061] As for the four-way vector method, Figure 2 As shown, it includes the following steps:

[0062] (b1) First, find the two monitoring points (such as monitoring stations) closest to the positioning location. These two monitoring points are recorded as A and B. At this time, the four-way vector value is calculated based on the location information of A and B. Specifically, the four-way vector value refers to the establishment of four-way coordinates with the user's reported location as the origin and the four directions of east, south, west and north as the axes. Based on the lines connecting A and B with the origin, the vectors are obtained respectively. The unit vectors of (i.e., draw an arrow pointing to A or B with a unit length of 1, and the modulus of each unit vector is 1), are recorded as two fixed point vectors. For a unit vector, the four-way vector value of a vector is the projection size of the vector in the east, west, south and north directions respectively (if the projection in a certain direction does not exist, the projection value in that direction is recorded as 0), and the projections in the east and west directions do not exist at the same time, and the projections in the south and north directions do not exist at the same time.

[0063] (b2) Decompose the two fixed point position vectors mentioned above into four-directional vectors to obtain the four-directional component values ​​of the user's reported position; The four-way vector values ​​of the unit vector are denoted as east OA ,West OA ,South OA ,north OA ,Will The four-way vector values ​​of the unit vector are denoted as east OB ,West OB ,South OB ,north OB ,

[0064] (i) If max{East OA ,East OB}≥0.5, max{West OA ,West OB}≥0.5, max{South OA ,South OB}≥0.5, max{North OA ,north OB If the four inequalities}≥0.5 hold true at the same time, then points A and B meet the requirements and can be directly interpolated between them.

[0065] (ii) If max{East OA ,East OB}≥0.5, max{West OA ,West OB}≥0.5, max{South OA ,South OB}≥0.5, max{North OA ,north OB}≥0.5When one or more of the four inequalities do not hold, it is necessary to find another monitoring point C (for example, it can be found through iteration) so that: The four-way vector values ​​of the unit vector are recorded as east OC ,West OC ,South OC ,north OC ,max{East OA ,East OB ,East OC}≥0.5, max{West OA ,West OB ,West OC}≥0.5, max{South OA ,South OB ,South OC}≥0.5, max{North OA ,north OB ,north OC}≥0.5These four inequalities hold true at the same time.

[0066] It is worth noting that when searching for point C, the maximum distance between point C and the origin does not exceed the preset maximum distance L max (e.g. 5 km). Otherwise, the current prediction is considered a failure, the location of the point is reported, and the user is guided to move to another test point. In addition, if point C cannot be found to make max{east OA ,East OB ,East OC}≥0.5, max{West OA ,West OB ,West OC}≥0.5, max{South OA ,South OB ,South OC}≥0.5, max{North OA ,north OB ,north OC}≥0.5, if all four inequalities hold true at the same time (for example, when points A and B are located exactly within 30° east of the north (or south) and 30° west of the north (or south) of the origin, respectively), the current prediction is considered a failure, the location of the point is reported, and the user is guided to move and change the test point.

[0067] For example:

[0068] Assume that the user's location is the origin O, and the two nearest air quality monitoring stations are located at A and B. For simplicity, we assume that the two stations are on the east-west or north-south axis, so that there is no need for complex vector decomposition (of course, if both stations are not on the coordinate axis, this can be achieved by using the four-way vector method in step (b) above to find C).

[0069] For example, in this embodiment:

[0070] User position O: (0,0)

[0071] Monitoring station A is located at (-3,0), which is 3 units west of the user.

[0072] Monitoring station B is located at (0,4), which is 4 units north of the user.

[0073] Step 1: Calculate the relative distance and direction to each site

[0074] For monitoring station A: Because it is located due west of the user, The four-way vector value of the unit vector: East OA is 0, West OA 1, South OA 0, North OA is 0.

[0075] For monitoring station B: It is located in the north direction of the user, so The four-way vector value of the unit vector: East OB is 0, West OB 0, South OB 0, North OB is 1.

[0076] Step 2: Check whether the values ​​of the four components are all greater than or equal to 0.5

[0077] Now we need to check the vector Are the four-way vector values ​​of the unit vector greater than or equal to 0.5? OA ,East OB}=0,max{West OA ,West OB}=1,max{South OA ,South OB}=0,max{North OA ,north OB}=1, and both the east component and the south component are less than 0.5, then a new site C can be found in the intersection of the range of 60 degrees to the east and the range of 60 degrees to the south from the origin.

[0078] Step 3: Predict air quality using monitoring station data

[0079] Once the closest station is determined or the auxiliary station C is found, the system estimates the air quality index (AQI) of the user's current location based on the data of these stations through linear interpolation or interpolation methods known in the art.

[0080] Take linear interpolation as an example: Linear interpolation is a method for creating a new estimated value between two known data points. In this real-time air pollution data query system based on mobile phone positioning, linear interpolation is used to predict the air quality index (AQI) at the user's location.

[0081] For two given points A(x1,y1) and B(x2,y2), where x represents the geographic coordinates (such as longitude and latitude) and y represents the corresponding air quality index (AQI), the linear interpolation formula is:

[0082]

[0083] Here x is the user's location and y is the AQI value at the user's location that we want to estimate.

[0084] When using linear interpolation for prediction, the following sub-steps may be included:

[0085] i. Determine the nearest site: First, find the two air quality monitoring sites closest to the user.

[0086] ii. Calculate weights: Based on the distance between the user and each site, calculate the weight of each site's impact on the AQI at the user's location.

[0087] iii. Apply the formula: Substitute the coordinates and AQI values ​​of these stations into the above linear interpolation formula to calculate the AQI value at the user's location.

[0088] For example, suppose the user's current location is P (39.9042°N, 116.4074°E). There are two monitoring stations A and B closest to P, located at A (39.9040°N, 116.4070°E) and B (39.9045°N, 116.4080°E), with corresponding AQIs of 80 and 90, respectively. As an example, we can first consider only latitude changes and apply the linear interpolation formula:

[0089]

[0090] AQI P =80+4

[0091] AQI P =84

[0092] Therefore, based on linear interpolation, the AQI at the user's current location is approximately 84.

[0093] To specifically illustrate how to iterate to ensure that the component values ​​in all four directions are greater than or equal to 0.5 and that the maximum distance limit of no more than 5 kilometers is followed when searching for point C, we can describe it in detail as follows. This process is an iterative process that continues until the conditions are met or the current prediction fails. The specific iterative process can be as follows:

[0094] S1. Initialization may include the following sub-steps (the relevant calculation process can be found in the previous text):

[0095] S11. Take the user position O as the origin.

[0096] S12. Determine the locations of monitoring stations A and B.

[0097] S13. Using the initial calculation of the four-way vector based on the relationship between A and B and the origin, we get the vector The unit vector of .

[0098] S2. The first iteration is for inspection and adjustment, and may include the following sub-steps:

[0099] S21. Calculate the initial four-way components: Based on the positions of monitoring stations A and B, calculate the four-way components (east, south, west, and north) of each station relative to the user.

[0100] S22. Evaluate component values: Check whether the absolute values ​​of the components in all four directions are greater than or equal to 0.5, that is, max{east OA ,East OB}≥0.5, max{West OA ,West OB}≥0.5, max{South OA ,South OB}≥0.5, max{North OA ,north OB}≥0.5. If all four inequalities hold true, then the subsequent steps are skipped and the air quality estimation process is performed directly (i.e., step S4). If any component value in any direction is less than 0.5 (i.e., one of the four inequalities does not hold true), the process proceeds to the next step S23.

[0101] S23. Select a new site C: For each direction where the component value is less than 0.5 (i.e., max{a certain direction OA ,same direction OB}<0.5), search for the nearest monitoring station C within 60 degrees up and down or left and right from the origin (for example, when a direction is east or west, search within 60 degrees up and down from the origin; when a direction is south or north, search within 60 degrees left and right from the origin). Make sure that the distance between the found station C and the origin does not exceed the preset maximum distance L max If the preset maximum distance L is exceeded max , the current prediction fails and the user is reminded to change location.

[0102] S3. After selecting the new site C, perform the following iterations:

[0103] S31. Update four-way components: Once a new site C is found, recalculate the four-way components of all involved sites (now including C).

[0104] S32. Evaluate the component values ​​again: Similar to sub-step S22, check whether the updated four-way component values ​​are all greater than or equal to 0.5 (ie, check max{east OA ,East OB ,East OC}≥0.5, max{West OA ,West OB ,West OC}≥0.5, max{South OA ,South OB ,South OC}≥0.5, max{NorthOA ,north OB ,north OC}≥0.5). If they all hold true, the iteration ends and the air quality estimation is prepared (i.e., step S4). If there is still a component value less than 0.5, but no more sites can be selected as C, or all possible C points exceed the preset maximum distance L. max range, the current prediction fails and the user is reminded to change the location.

[0105] S4. Air quality estimation: using the data from all selected stations and estimating the air quality index (AQI) at the user's location through linear interpolation or an interpolation method known in the prior art, and feeding it back to the user through a mobile phone application.

[0106] (i) When max{East OA ,East OB}≥0.5, max{West OA ,West OB}≥0.5, max{South OA ,South OB}≥0.5, max{North OA ,north OB When these four inequalities are true at the same time, only two points A and B are needed to perform interpolation calculation:

[0107] (i-1) If points A, O, and B are collinear (point O must be between points A and B), then,

[0108]

[0109] Among them, |AB| represents the distance between points A and B, and |OA| represents the distance between points O and A;

[0110] (i-2) If points A, O, and B are not collinear, then, also based on the interpolation method:

[0111]

[0112] in, |OA| represents the distance between points O and A, and |OB| represents the distance between points O and B;

[0113] (ii) When there is a point C and max{east OA ,East OB ,East OC}≥0.5, max{West OA ,West OB ,West OC}≥0.5, max{South OA ,SouthOB ,South OC}≥0.5, max{North OA ,north OB ,north OC When the four inequalities}≥0.5 hold simultaneously, the data from the three sites A, B, and C can be combined for calculation: first, connect line AB and extend it into a straight line, then connect line OC and extend it into a straight line, with the intersection of the two lines at D. First, calculate the estimated value of point D using linear interpolation of line AB, and then calculate the estimated value of point O using line CD to obtain the AQI value at the user's location, that is:

[0114]

[0115] Furthermore, if a prediction fails, the system can record locations where sufficiently accurate data cannot be obtained and recommend that users move to a different location to retry obtaining air quality information. These failure points can then be further statistically analyzed to optimize the layout of monitoring stations or adjust patrol routes for unmanned vehicles in the future.

[0116] In addition, Example 1 is based on an air monitoring base station. There are many types of air monitoring base stations. They can be environmental monitoring departments of government departments that disclose data to the outside world, or they can be air quality data collection units placed at fixed locations in the city. Their main function is to report coordinate information and air quality data. The data reporting period can be once every 1 minute, and their energy supply method can be solar energy or various other forms.

[0117] Example 2

[0118] The real-time air pollution data query system based on mobile phone positioning in the present invention has solved how to calculate the air pollution index of the positioning location in Example 1 (as illustrated in Example 1, it can rely on selecting an air quality monitoring base station near the positioning, and through strategy selection, confirm the air quality data near the positioning). This Example 2 will focus on how to improve the accuracy of air pollution data calculation, that is, how to obtain more accurate data based on the acquisition of air quality data.

[0119] In this embodiment, Figure 3 As shown, a method for calculating the air pollution data at the positioning location is used to collect air quality data and request calculation of air quality data in real time within the system coverage area through unmanned vehicles or other unmanned vehicles, and to correct the calculation method of the air pollution data at the positioning location by combining the position calculation strategy with the regression model trained by the collected data.

[0120] This embodiment primarily determines the error value of the aforementioned embodiment 1 and uses regression and mapping methods to correct the calculated air pollution value. This primarily involves the mobile unit collecting real-time location and air pollution data, invoking a data acquisition service to obtain the real-time calculated value and calculation method (of course, if there is a monitoring device within 500 meters, the data is directly obtained from the monitoring device, without the need for interpolation). This data is then reported to the database along with the two previously mentioned data points.

[0121] Based on the collected data, for direct data acquisition within 500 meters, the difference between the real-time value and the value obtained by calling the service is calculated. If the difference is less than 5%, the 500-meter direct data acquisition strategy is feasible, and the distance parameter is retained. If the difference is large, the distance needs to be adjusted promptly to further narrow the scope of the direct data acquisition method.

[0122] For values ​​obtained by the calculation method and the real-time data, calculate the difference using the same method described above. If the average difference is less than 5%, the strategy can be temporarily retained. If the difference is greater than 5%, a regression method is used to calculate the mapping relationship between the real-time data and the data obtained by the service. Regression methods include but are not limited to linear regression and random forest regression.

[0123] Specifically, the following steps may be included:

[0124] (1) Data preparation

[0125] Suppose we have n sample points, each sample point includes:

[0126] y i real : Real-time air quality index (AQI) obtained by unmanned vehicles or other means.

[0127] y i calc : The AQI value predicted by the server obtained through calculation methods (such as linear interpolation).

[0128] (2) Calculating the average difference may include the following sub-steps:

[0129] (2.1) Calculate the difference of a single sample point: For each sample point i, calculate its absolute percentage error (APE):

[0130]

[0131] (2.2) Calculate the average difference of all sample points: Find the average value of the APE of all sample points, that is, the Mean Absolute Percentage Error (MAPE):

[0132]

[0133] When the service determines that additional mapping is needed, when the user calls the service, after obtaining the calculated value according to the above method, the final air pollution data needs to be calculated through the regression mapping relationship. The data at this time is considered to be the data finally delivered to the user.

[0134] (3) Decision logic:

[0135] (3.1) Determine whether the threshold condition is met:

[0136] If MAPE < 5%, it is considered that the data obtained by the calculation method is consistent with the real-time data, and the current strategy can be temporarily retained.

[0137] If MAPE ≥ 5%, further processing is required to improve the prediction accuracy.

[0138] (4) Mapping using regression method:

[0139] (4.1) Construct the training set: y of all the above sample points i calc and y i real Compose a dataset for training a regression model.

[0140] (4.2) Select and train regression models: You can choose a variety of regression models to try, such as linear regression, random forest regression, etc.

[0141] (4.2.1) Take linear regression as an example: Assume that y i real and y i calc If there is a linear relationship between them, the best fitting line y=ax+b can be found by the least squares method, where x represents y i calc , y represents y i real .

[0142] (4.2.2) Let's take random forest regression as an example: If the nonlinear relationship is more complex, consider using random forest regression. It consists of multiple decision trees and can capture more complex patterns. Each tree can be trained on a portion of the data, and the average prediction of all trees is taken as the output.

[0143] (4.3) Evaluate model performance: After training the model, its performance can be evaluated on an independent test set to ensure that the model is not overfitting. Common evaluation indicators include mean square error (MSE), R 2 Scoring, etc.

[0144] (4.4) Applying the model to modify the prediction: When the user requests the latest air quality information, the initial prediction value y is first obtained by the original calculation method. calc , and then input it into the trained regression model to get the adjusted predicted value y adj In this way, even if the original calculation method has a certain deviation, the predicted value after adjustment by the regression model should be closer to the actual value.

[0145] (4.5) Continuous optimization: As time goes by and more data is accumulated, the regression model is retrained regularly to adapt to possible changing trends and maintain the accuracy of the prediction.

[0146] Example 3

[0147] This embodiment is aimed at the locations where prediction failed in embodiment 1, and the failed locations are counted. Fixed collection units can be newly added at locations where the failed locations are concentrated.

[0148] Example 4

[0149] This embodiment is a method for Example 1, and the following supporting devices are provided:

[0150] The working process of the device is as follows:

[0151] The user wears the smart terminal on their body or carries it with them in other ways. Before use, the smart terminal is activated. The user registers and sets personalized data on the mobile phone. The user settings and user data are uploaded once and recorded in the server database.

[0152] When a user is out and about, they carry a portable device and turn it on. The device uploads collected air quality data at regular intervals. The uploaded message includes not only air pollution data but also longitude and latitude information. Upon receiving the data, the server processes it and links it with data from the environmental monitoring center. This completes the air quality data record. Once the server stores the data, it calls the intelligent prediction service once during the database's calculation cycle. Upon detecting an anomaly and meeting the indicator, an alert is pushed. This generates a valid alert on both the client and the smart terminal.

[0153] In addition to the server performing a data forecast when the data is updated, the smart terminal will also interact with the server (RPC) query at the set time. At this time, the smart terminal will send the latitude and longitude information of the location to the server. After receiving the latitude and longitude information, the server will match it based on the principle of the closest range. If the minimum difference between the location information in the database and the location information reported by the smart terminal is greater than 5 kilometers, the RPC will return an error message. After receiving the error message, the smart terminal will actively trigger an environmental data upload. At this time, it mainly determines whether the reported pollutant information exceeds the index, that is, whether the air pollution index exceeds the threshold. If the threshold is exceeded, an alarm will be issued.

[0154] Furthermore, in addition to periodic RPC data pulling, the smart terminal can also be initiated by the user. The user can initiate an environmental quality judgment by clicking a button on the smart terminal.

[0155] Based on the above embodiments, it is not difficult to see that the present invention, when applied in practice, may include the following steps:

[0156] (1) The user reports the latitude and longitude of the current location through the mobile application.

[0157] (2) The system receives the information and starts a search algorithm to find all available air quality monitoring stations within a set initial radius (e.g., 500 meters).

[0158] (3) If a monitoring station is found, the data of the station is directly returned to the user; if not found, the search radius is adjusted according to the preset strategy and the above operation is repeated.

[0159] (4) Once the monitoring station closest to the user is determined, the AQI value at the user's location will be calculated using linear interpolation or other methods.

[0160] In addition, the system can also record the results of each query for subsequent optimization of search strategies.

[0161] Application examples:

[0162] Application Example 1: Real-time query system for air pollution data based on mobile phone positioning

[0163] Background: In a medium-sized city, the government has deployed several fixed air quality monitoring stations and deployed several unmanned vehicles equipped with air quality sensors to patrol the city and collect data. City residents widely use a mobile app called "AirQCheck" to check the air quality conditions in their current location.

[0164] Step 1: User positioning and preliminary data acquisition

[0165] User Zhang Wei opens the "AirQCheck" app, and the app uses GPS positioning technology to determine Zhang Wei's current latitude and longitude location as (39.9042°N, 116.4074°E).

[0166] After receiving the location information, the system immediately checks to see if there are any known air quality monitoring stations within 500 meters of Zhang Wei. Assume that there is station A within this range, and its most recent uploaded air quality index (AQI) is 85.

[0167] Step 2: Assisted correction of unmanned vehicle data

[0168] The system dispatches the nearest unmanned vehicle (300 meters from Zhang Wei) to collect data and install an air quality sensor. The vehicle collects air quality data and calculates a value based on its location coordinates. The collected value is compared with the calculated value. If the difference exceeds 5%, the system proceeds to the next step. The system determines whether the collected point is within 500 meters of the fixed location data. If so, the system checks whether the difference is less than 5%. Otherwise, the fixed distance is adjusted. If the difference is less than 5%, the current status is maintained; otherwise, the fixed distance is adjusted until the condition is met. The system calculates the mapping between the collected and calculated values ​​to facilitate subsequent corrections and adjustments.

[0169] In this way, the data from the unmanned vehicle is incorporated into the correction system, and the weight is flexibly adjusted according to the degree of difference between the data and the fixed monitoring station data, thereby achieving more accurate air quality predictions.

[0170] Step 3: Error assessment and strategy adjustment

[0171] The system recorded the data discrepancies during this query and compared them with historical data. It found that the average discrepancy between the strategy of directly acquiring data within 500 meters and the actual autonomous vehicle data was less than 5%, meeting the accuracy requirements.

[0172] The system also analyzes the differences between the values ​​obtained by the calculation method and the actual measured values ​​of the unmanned vehicle, confirming that the mapping relationship model (such as the linear regression model) can effectively reduce the prediction error in future queries and ensure the accuracy and stability of long-term data.

[0173] Step 4: Feedback and optimization suggestions

[0174] The "AirQCheck" application displayed the final calculated AQI value of 87.2 to Zhang Wei, and prompted that the air quality in the area was good.

[0175] Based on the contribution of the unmanned vehicle data in this query, the system background automatically records the possible service blind spots in the area, and considers adding fixed monitoring stations or adjusting the patrol routes of unmanned vehicles in areas where large data deviations frequently occur in the future to continuously optimize data accuracy and service coverage.

[0176] This example demonstrates how to combine mobile phone positioning, fixed monitoring station data, unmanned vehicle dynamic data collection, and advanced data analysis technology to provide users with high-precision, real-time air pollution data query services.

[0177] In summary, this invention improves data accuracy and real-time performance: by combining mobile phone positioning technology with linear interpolation, the system can provide users with more accurate location-specific air quality forecasts. In areas where direct monitoring data is lacking, the mobile monitoring capabilities of unmanned vehicles enable dynamic data collection and real-time correction, significantly improving the accuracy and timeliness of air quality data.

[0178] Unmanned vehicles can dynamically collect data in areas not covered by fixed stations or where data is insufficient, and feed this real-time data back to the system, allowing methods such as linear interpolation to make predictions based on more comprehensive and up-to-date information. The following is a detailed description of how unmanned vehicles can improve data accuracy and real-time performance:

[0179] Strategies to improve data accuracy and real-time performance

[0180] 1. Mobile monitoring of unmanned vehicles

[0181] Patrol route planning: Based on factors such as urban structure, traffic flow, and historical pollution patterns, unmanned vehicle patrol routes are pre-planned to ensure they can cover important areas that are difficult to reach by fixed monitoring stations.

[0182] Dynamically adjust patrol routes: Use machine learning algorithms to analyze current air quality and weather forecast data, intelligently adjust the patrol routes of unmanned vehicles, and prioritize areas where problems may exist (such as high pollution risks).

[0183] Real-time data upload: The unmanned vehicle is equipped with advanced sensors that can collect data on various pollutants including PM2.5, PM10, SO2, etc. in real time and upload it to the cloud server immediately.

[0184] 2. Data fusion and correction

[0185] Real-time data integration: Combines data uploaded by unmanned vehicles with data from fixed monitoring stations to form a more complete and accurate air quality map.

[0186] Linear interpolation optimization: When the user is located between two fixed monitoring stations, in addition to using the data from these two stations, you can also add data points collected when the unmanned vehicle passed through the area most recently to provide more accurate linear interpolation results.

[0187] Application of Deep Regression Models: For situations where linear interpolation still fails to achieve sufficient accuracy, deep regression models can be used to further refine the predicted values. For example, random forest regression or neural network models can adjust the predicted values ​​based on various characteristics (such as time, location, and weather conditions) to make them more accurate to the actual situation.

[0188] 3. Feedback mechanism and continuous improvement

[0189] User feedback: Users are encouraged to provide information about their perception of air quality, such as smell, visual observation, etc., as an additional source of data to help validate and calibrate the model.

[0190] Model update: As more high-quality data accumulates, the regression model is regularly retrained and optimized to ensure that it is always performing at its best.

[0191] Anomaly detection and processing: Establish an automated anomaly detection system. Once abnormal fluctuations in data in a certain area are detected, an unmanned vehicle will be immediately dispatched to investigate the cause and collect the latest data to ensure data continuity and reliability.

[0192] Enhanced user experience: Users can quickly and conveniently check the air quality status of their current location via their mobile phone, meeting the public's growing demand for healthy living and environmental protection. The system's intelligent error detection and correction mechanism ensures the reliability of the information users receive, enhancing user trust and satisfaction.

[0193] Optimize resource allocation: Through statistical analysis of failure points, the system can intelligently identify monitoring blind spots and guide the deployment of new fixed collection units, optimizing the layout of the urban air quality monitoring network, avoiding blind investment of resources, and improving overall monitoring efficiency and coverage.

[0194] Promoting scientific environmental decision-making: The system collects a large amount of real-time and accurate air quality data, providing valuable first-hand information for environmental protection departments, helping policymakers make more scientific and reasonable environmental decisions, effectively address air pollution problems, and protect public health.

[0195] Technological innovation and model promotion: This invention integrates multiple advanced technologies such as the Internet of Things, big data, and artificial intelligence, demonstrating the innovative application of "Internet + Internet of Things" in the field of environmental protection. It provides a reference model for the design and implementation of similar environmental monitoring systems and has a strong demonstration effect and promotion value.

[0196] Adaptable and scalable: The system's flexible design allows for adjustments based on the specific needs and geographic characteristics of different cities. Whether in urban centers or remote areas, effective coverage of the monitoring network can be ensured by adjusting the layout of monitoring sites and the patrol strategy of unmanned vehicles. Furthermore, as technology advances, the system can easily integrate more advanced predictive models and data processing algorithms, maintaining technological leadership and service effectiveness.

[0197] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting air quality index based on mobile phone positioning, characterized in that: The following steps are involved: (1) collecting user-reported locations; (II) If there is an air pollution monitoring point within the preset radius of the reported location, then the data of the monitoring point is used as the air quality index of the reported location, and the process jumps to step (V); If there is no air pollution monitoring point within the preset radius of the reported location, then proceed to step (III); (III) Select the two air pollution monitoring points closest to the reported location, record them as A and B respectively, take the reported location as the origin O, and obtain the vector The unit vector of ; then, calculate The projection size of the unit vector in the east, west, south and north directions is respectively recorded as OA ,West OA ,South OA ,north OA ; Calculate at the same time The projection size of the unit vector in the east, west, south and north directions is recorded as east OB ,West OB ,South OB ,north OB Among them, East OA Hexi OA At least one of them is 0, OA ,north OA At least one of them is 0, East OB ,West OB At least one of them is 0, OB ,north OB At least one of is 0; (IV) If max{East OA ,East OB }≥0.5, max{West OA ,West OB }≥0.5, max{South OA ,South OB }≥0.5, max{North OA ,north OB If the four inequalities}≥0.5 are simultaneously established, the air quality index at point O is calculated based on the air quality indexes monitored at A and B using the interpolation method, and the process jumps to step (V); Otherwise, take point O as the center and find another air pollution monitoring point C within the preset maximum limit radius so that: max{east OA ,East OB ,East OC }≥0.5, max{West OA ,West OB ,West OC }≥0.5, max{South OA ,South OB ,South OC }≥0.5, max{North OA ,north OB ,north OC }≥0.5, these four inequalities hold true at the same time; among them, OC ,West OC ,South OC ,north OC for The projection size of the unit vector in the east, west, south and north directions, east OC Hexi OC At least one of them is 0, OC ,north OC At least one of them is 0; then, the air quality index at point O is calculated based on the air quality indexes monitored by A, B, and C using the interpolation method, and the process goes to step (V); If C cannot be found within the preset maximum limit radius, the user is prompted to modify the reported location and the method is exited; (V) Sending the air quality index of the reported location obtained in step (II) or the air quality index of point O obtained in step (IV) to the user's mobile phone.

2. The method according to claim 1, wherein: In step (IV), the air quality index at point O is calculated based on the interpolation method using the air quality indexes monitored by A and B. Specifically, the air quality indexes monitored by A and B are AQI A , AQI B , If points A, O, and B are collinear, then the air quality index AQI at point O is O satisfy: Among them, |AB| represents the distance between points A and B, and |OA| represents the distance between points O and A; If points A, O, and B are not collinear, then in, |OA| represents the distance between points O and A, and |OB| represents the distance between points O and B.

3. The method according to claim 1, wherein: In step (IV), the air quality index at point O is calculated based on the air quality index monitored by A, B, and C using the interpolation method. Specifically, the air quality indexes monitored by A, B, and C are AQI A , AQI B , AQI C , First, calculate the air quality index AQI at point D based on the interpolation method D , point D is the intersection of the line AB and the line OC: Among them, |AB| represents the distance between points A and B, and |DA| represents the distance between points D and A; Then, based on the interpolation method, the air quality index AQI at point O is calculated O : Among them, |CD| represents the distance between points C and D, and |OC| represents the distance between points O and C.

4. The method according to claim 1, wherein: In step (II), the preset radius is 500m.

5. The method according to claim 1, wherein: In step (IV), the preset maximum restriction radius is 5 km.

6. The method according to claim 1, wherein: Step (V) further includes: when the air quality index of the reported location or the air quality index at point O exceeds a preset warning value, issuing an air pollution exposure alarm via the user's mobile phone; Preferably, the air pollution monitoring points are used to monitor the concentrations of air pollutants PM2.5, PM10 and / or SO2.

7. The method according to claim 1, wherein: Step (V) also includes data correction processing: The real air quality index of the reported location is collected by unmanned vehicles and / or drones, recorded as AQI real ; The air quality index of the reported location obtained in step (II) is recorded as AQI calc , like No correction is required; like Then, adjust the preset radius range and jump to step (II); The air quality index at point O calculated in step (IV) is recorded as AQI calc , like No correction is required; like Then, AQI calc Input the trained regression model and use the AQI output by the model adj As the modified O-point air quality index; The regression model is obtained by collecting multiple groups of AQI real with AQI calc The samples are trained, and each AQI in these samples is calc with AQI real One to one correspondence.

8. The method according to claim 7, wherein: The regression model is a linear regression model or a random forest regression model.

9. An air quality index prediction system based on mobile phone positioning, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images