Method and system for constructing air pressure fingerprint model

Abstract

The application discloses a method for constructing an air pressure fingerprint model, which comprises the following steps: sampling based on a sampling device with known horizontal and vertical positions to obtain sample information; constructing a sample information library; defining an influence range of the sampling device and performing grid division based on the influence range; fitting an air pressure baseline of each grid from the sample information library, wherein the air pressure baseline is an air pressure reference value at a specified temperature and height within the grid; sampling sample information in any grid based on a terminal device with known horizontal position, wherein the data volume of the sample information meets a preset quantity, comparing the sample information with a preset environment semantic label template to obtain an environment semantic label of at least one grid; and constructing an air pressure fingerprint model, wherein the air pressure fingerprint model comprises the air pressure baseline of each grid and the environment semantic label of at least one grid. The method can improve the environment perception accuracy in a complex scene by constructing a standardized air pressure baseline field and realizing intelligent environment semantic labeling.

Description

Technical Field

[0001] This application relates to the field of environmental perception, and in particular to a method and system for constructing a barometric fingerprint model. Background Technology

[0002] Currently, technologies using barometric pressure data for environmental sensing or assisted positioning typically rely directly on raw barometer readings. However, raw readings are significantly affected by temperature, equipment altitude, and local microenvironment, leading to a lack of comparability between data collected from different devices and locations, making it difficult to construct a consistent regional pressure field model. Furthermore, related technologies lack continuous learning and adaptive correction mechanisms for dynamic changes in pressure data, making it difficult for models to adapt to long-term pressure field drift caused by seasonal changes, weather variations, or urban environmental modifications once established. Summary of the Invention

[0003] This application provides a method for constructing a barometric fingerprint model. This method improves the accuracy of environmental perception in complex scenarios by constructing a standardized barometric reference field and realizing intelligent annotation of environmental semantics.

[0004] In a first aspect, embodiments of this application provide a method for constructing a barometric fingerprint model, the method comprising:

[0005] Sampling is performed using known horizontal and vertical positioning sampling devices to obtain sample information, which includes at least air pressure data and temperature data.

[0006] Construct a sample information database, which includes information on multiple samples;

[0007] Define the influence range of the sampling equipment, and divide the grid based on the influence range;

[0008] Based on the sample information database, the air pressure baseline of each grid is fitted. The air pressure baseline is the air pressure reference value within the grid at a specified temperature and altitude.

[0009] Based on the known horizontal positioning of the terminal device, sample information is sampled within any grid. The amount of sample information sampled by the terminal device meets the preset quantity. The sample information sampled by the terminal device is compared with the preset environmental semantic label template to obtain the environmental semantic label of at least one grid.

[0010] A barometric fingerprint model is constructed, which includes the barometric baseline of each grid and the environmental semantic label of at least one grid.

[0011] In one embodiment, the construction method further includes:

[0012] The confidence level of the air pressure baseline in any grid in the barometric fingerprint model is assigned, and the confidence level of the air pressure baseline of each grid is inversely proportional to the distance between the nearest sampling device.

[0013] In one embodiment, the construction method further includes:

[0014] Based on the sample information of any grid sampled by the terminal device, the confidence level of the air pressure baseline is adjusted. When the number of times the confidence level of the air pressure baseline of any grid is lower than the preset threshold is greater than or equal to the preset number, the air pressure baseline of the grid is updated.

[0015] In one embodiment, adjusting the confidence level of the air pressure baseline based on sample information from arbitrary grid sampling by the terminal device includes:

[0016] Based on the air pressure value sampled by the terminal device in any grid, the absolute difference between the air pressure value and the air pressure baseline of the corresponding grid is calculated. When the absolute difference is less than or equal to the preset error range, the confidence of the air pressure baseline is increased; when the absolute difference is greater than the preset error range, the confidence of the air pressure baseline is decreased.

[0017] The preset error range is defined by the environmental semantic labels of the grid.

[0018] In one embodiment, when adjusting the confidence level of the air pressure baseline, the confidence level of the air pressure baseline decreases or increases according to a preset step size;

[0019] When the number of times the confidence level of the air pressure baseline of any grid is lower than the preset threshold is greater than or equal to the preset number of times, the air pressure baseline of the grid is updated to: the weighted average between the initial air pressure baseline fitted based on the sample information database and the air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline is adjusted to the first height. The first height is obtained by calculating the average of the height corresponding to the second height and the initial air pressure baseline. The second height is the height at which the terminal device samples the air pressure data. The confidence level of the air pressure baseline of the grid is increased to a value between the current air pressure baseline confidence level and the preset threshold for air pressure baseline confidence level.

[0020] If the confidence level of the air pressure baseline still decreases after the air pressure baseline is updated, the air pressure baseline of the grid will be updated to the average value of the partial air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline will be adjusted to the third height. The third height is the average height of the partial air pressure data sampled by the terminal device for calculating the average value. The partial air pressure data used by the terminal device for calculating the average value satisfies the following condition: the sampling time of the air pressure data is less than the current time threshold.

[0021] In one embodiment, the construction method further includes:

[0022] The sample information collected by the terminal device is compared with the preset environmental semantic label template to obtain the environmental semantic label of at least one grid.

[0023] Based on sample information collected by terminal devices within a specified time range, physical tag data is calculated. Physical tag data includes one or more of the following types: air pressure change rate, absolute air pressure change, temperature change rate, and temperature value.

[0024] The calculated physical label data is compared with the physical label data in the environmental semantic label template. When the similarity is greater than the preset similarity threshold, the environmental semantic label corresponding to the environmental semantic label template is marked for the grid. Based on the similarity, the environmental semantic label confidence score is assigned to the environmental semantic label of the grid. The environmental semantic label confidence score is proportional to the similarity score.

[0025] In one embodiment, the construction method further includes:

[0026] Based on sample information collected by the terminal device and the horizontal positioning at the time of sampling, the air pressure baseline of the corresponding grid is adjusted. Then, based on the adjusted air pressure baseline, the confidence level of the environmental semantic labels is adjusted, including the following steps:

[0027] Based on the set environmental semantic labels, perturbation items are removed from the sample information sampled by the terminal device to obtain air pressure correction data containing altitude information;

[0028] Based on the air pressure correction data and combined with the adjusted air pressure baseline, the vertical height positioning information when obtaining the sample information of the terminal device is calculated.

[0029] Based on the vertical height positioning information, the height information of the air pressure correction data is removed, and the corrected air pressure value when the terminal device is at the height corresponding to the air pressure baseline is calculated.

[0030] Based on the corrected air pressure value, physical label data is calculated. The similarity between the calculated physical label data and the physical label data in the environmental semantic label template is calculated. Based on the similarity, the confidence of the environmental semantic label is adjusted for the grid's environmental semantic label.

[0031] In one embodiment, the construction method further includes:

[0032] Compare the sample information collected by the terminal device in the latest time period with the preset environmental semantic label template to obtain the environmental semantic label of at least one grid and the corresponding environmental semantic label confidence.

[0033] The newly generated environmental semantic label of the terminal device is compared with the existing environmental semantic label of the grid. If the two labels are the same and the confidence of the newly generated environmental semantic label of the terminal device is greater than the first preset confidence threshold, the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model is increased.

[0034] If the newly generated environmental semantic label of the terminal device does not match the existing environmental semantic label of the grid, or if the confidence of the newly generated environmental semantic label of the terminal device is less than the second preset confidence threshold, then the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model will be reduced.

[0035] If the confidence level of the environmental semantic label of any grid in the barometric fingerprint model is lower than the third preset confidence threshold within a specified time range, the environmental semantic label of the corresponding grid in the barometric fingerprint model will be deleted.

[0036] Among the first, second, and third preset reliability thresholds, any two preset reliability thresholds may be equal or unequal.

[0037] In one embodiment, the horizontal air pressure is calculated based on a sample information database using the following formula;

[0038]

[0039] In the formula, T This indicates the temperature sampled by the sampling device. h Indicates the vertical positioning height of the sampling device. This represents the air pressure value sampled by the sampling device. R d This represents the specific gas constant of dry air. The specified temperature represents the air pressure reference value. g =9.81 m / s 2 .

[0040] Secondly, this application also provides a system for constructing a barometric fingerprint model, which constructs a barometric fingerprint model based on the method for constructing a barometric fingerprint model in the first aspect.

[0041] The aforementioned method for constructing a barometric fingerprint model involves collecting sample information containing air pressure and temperature using sampling devices with known horizontal and vertical positioning. After constructing a sample information database, the model is divided into grids based on the influence range of the sampling devices. A pressure baseline for each grid at a specified temperature and altitude is then fitted using the sample information database. Next, sample information meeting the required data volume is collected using a terminal device with known horizontal positioning. This data is compared with an environmental semantic label template to assign environmental semantic labels to the grids, thus constructing a barometric fingerprint model containing the pressure baseline for each grid and at least one environmental semantic label for each grid. This method establishes a unified benchmark spatial pressure field, enabling semantic annotation of environmental features and forming a barometric fingerprint model that combines quantitative benchmarks with qualitative descriptions, providing a standardized data foundation for environmental perception and location services. Attached Figure Description

[0042] Figure 1 This is a flowchart of a method for constructing a barometric fingerprint model in one embodiment;

[0043] Figure 2 This is a flowchart illustrating the calculation of physical tag data in one embodiment;

[0044] Figure 3 This is a flowchart illustrating the adjustment of the confidence level of environmental semantic labels in one embodiment;

[0045] Figure 4 This is a flowchart illustrating the adjustment of the confidence level of environmental semantic labels in another embodiment. Detailed Implementation

[0046] The present application will be described in detail below with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present application. Any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the protection scope of the present application.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0048] This application provides a method for constructing a barometric fingerprint model. This method constructs the barometric fingerprint model based on sample data (including barometric pressure data and / or temperature data) collected by sampling devices. Before the sampling devices collect sample data, each sampling device can be calibrated to ensure that all barometers and thermometers are aligned within the same system, reducing the impact of device differences on the construction of the barometric fingerprint model.

[0049] In one embodiment, such as Figure 1 As shown, a method for constructing a barometric fingerprint model is provided, which includes the following steps:

[0050] Step 101: Sampling is performed using known horizontal and vertical positioning sampling devices to obtain sample information, which includes at least air pressure data and temperature data.

[0051] Specifically, sampling based on known horizontal positioning (e.g., latitude and longitude coordinates) and vertical positioning (e.g., altitude) can refer to collecting environmental parameters using sensing devices with fixed spatial location information (e.g., professional weather stations installed on the roof of specific buildings and precisely mapped).

[0052] During sampling, the sampling equipment can simultaneously record its geographical coordinates and altitude information, and collect at least two physical quantities: air pressure data (i.e., instantaneous readings of atmospheric pressure, usually measured in hectopascals (hPa) or pascals (Pa)) and temperature data (ambient air temperature, usually measured in degrees Celsius (°C)). Air pressure data can refer to the raw atmospheric pressure readings collected by sampling equipment with known horizontal and vertical positioning; these values ​​will vary with altitude, temperature, and weather systems. Temperature data can refer to the air temperature values ​​collected by sampling equipment with known horizontal and vertical positioning.

[0053] Step 102: Construct a sample information database, which includes information on multiple samples;

[0054] Specifically, the core objective of building the sample information database is to clean, standardize the format, synchronize the time, and control the quality of the raw sample information obtained from sampling devices at all known spatial locations (including precise horizontal coordinates and vertical elevations) and then aggregate it into a structured sample information database.

[0055] The sample information database consists of multiple sample information records. Each sample information record includes at least: the unique identifier of the sampling device, the precise timestamp of data collection, the coordinates (longitude and latitude) of the sampling device, the altitude of the sampling device, the corrected air pressure data, and the temperature data.

[0056] Step 103: Define the influence range of the sampling device and perform grid division based on the influence range;

[0057] Specifically, defining the influence range of a sampling device can refer to delineating a representative geographical area for each device's observation data based on its measurement accuracy, environmental complexity, and the spatial continuity of meteorological elements. Beyond this range, the reliability of the data decreases with increasing distance.

[0058] When dividing the grid based on the influence range, the influence radius of the sampling device is taken as an important scale parameter for grid division. In areas with dense equipment and overlapping influence ranges, fine-grained grids (e.g., 50m×50m) can be used to capture micro-changes; in areas with sparse equipment, the edge of the influence range, or isolated areas, coarse-grained grids (e.g., 500m×500m) can be used to reduce uncertainty.

[0059] For example, a weather station deployed in the city center, due to the complex surrounding buildings and diverse microclimates, can have its influence range defined as a radius of 500 meters, and a dense grid can be drawn at this scale; while another weather station in the suburbs is surrounded by flat terrain and a homogeneous environment, and its influence range can reach a radius of 2 kilometers, so the grid drawn can be larger and sparser.

[0060] Step 104: Based on the sample information database, fit the air pressure baseline for each grid. The air pressure baseline is the air pressure reference value within the grid at a specified temperature and altitude.

[0061] Specifically, based on the air pressure data provided by all sampling devices in the sample information database, a spatial interpolation algorithm (such as Kriging interpolation or inverse distance weighting) can be used to fit and calculate the air pressure baseline within each grid cell. The air pressure baseline can uniformly reduce each grid to the air pressure reference value that should exist under a specified standard temperature (e.g., the usual 25°C) and a specified standard altitude (e.g., sea level).

[0062] It should be noted that the process of fitting the pressure baseline for each grid involves constructing a virtual pressure field under standard conditions. This removes the vertical influence of actual ambient temperature fluctuations and terrain elevation differences on the pressure readings, ensuring that the pressure values ​​of all grids are placed on a unified reference plane. Based on the spatial distribution structure of the sampling equipment and data quality weights, a continuous spatial distribution function for pressure values ​​is established. For any grid point, its pressure baseline value is determined by the interpolation result of this function at that location.

[0063] For example, if the standard air pressure of sampling device A on the east side of a certain area is 1015 hPa after correction, and that of device B on the west side is 1010 hPa, then the air pressure baseline of a certain grid point located between A and B may be calculated as 1012 hPa through linear interpolation, and the area far from the connection line between the devices will generate a smooth transition value based on the interpolation surface.

[0064] Step 105: Based on the known horizontal positioning of the terminal device, sample information in any grid is sampled. The amount of sample information sampled by the terminal device meets the preset number. The sample information sampled by the terminal device is compared with the preset environmental semantic label template to obtain the environmental semantic label of at least one grid.

[0065] Specifically, based on the known horizontal positioning of mobile terminal devices (such as smartphones or vehicle sensors), sample information such as air pressure data and temperature data is continuously collected within any grid. When the sampling time or the amount of sample information accumulates to a preset number (such as 50 consecutive hours or more than 1,000 sets of sample information), the sample information sampled by the terminal device is compared with the preset environmental semantic label template.

[0066] The sample information collected by the terminal device is preprocessed to extract physical feature vectors, including quantitative indicators such as air pressure change rate (e.g., the differential mean within 10 seconds), air pressure fluctuation amplitude (e.g., range or standard deviation), and temperature change trend. The physical feature vectors are then matched against a pre-defined environmental semantic label template library in multiple dimensions. Each environmental semantic label template corresponds to a typical environmental phenomenon (e.g., a wind vent template may require an air pressure change rate below -0.5 Pa / s and a simultaneous temperature decrease). A similarity algorithm (e.g., cosine similarity or dynamic time warping) is used to quantify the degree of match between the sample information collected by the terminal device and the environmental semantic label template. If the similarity score between the sample information collected by the terminal device and the environmental semantic label template exceeds a pre-defined threshold (e.g., 0.7), a corresponding environmental semantic label (e.g., wind vent) is assigned to that grid.

[0067] Step 106: Construct a barometric fingerprint model, which includes the barometric baseline of each grid and the environmental semantic label of at least one grid.

[0068] Specifically, the barometric fingerprint model uses a grid as the basic unit. Based on sampling data from known locations, it calculates a standardized barometric baseline value (i.e., a theoretical barometric reference under a unified benchmark) for each grid. Then, using time-series data collected by terminal devices, it uses feature extraction and pattern matching to label the grid with environmental semantic tags such as wind vents and heat sources, and assigns corresponding confidence levels. This forms a hybrid model that integrates quantitative barometric benchmarks, qualitative environmental descriptions, and dynamic confidence assessments, providing a data foundation for high-precision positioning and environmental perception.

[0069] In this embodiment, the method collects sample information including air pressure and temperature using sampling devices with known horizontal and vertical positioning. After constructing a sample information database, it divides the area into grids based on the influence range of the sampling devices and fits the air pressure baseline of each grid at a specified temperature and altitude according to the sample information database. Then, it collects sample information that meets the data volume requirements using terminal devices with known horizontal positioning, compares it with an environmental semantic label template, assigns environmental semantic labels to the grids, and constructs a pressure fingerprint model containing the air pressure baseline of each grid and at least one grid environmental semantic label. This method establishes a unified benchmark spatial pressure field to achieve semantic annotation of environmental features, forming a pressure fingerprint model that combines quantitative benchmarks and qualitative descriptions, providing a standardized data foundation for environmental perception and location services.

[0070] In one embodiment, the horizontal air pressure is calculated based on a sample information database using the following formula;

[0071]

[0072] In the formula, T This indicates the temperature sampled by the sampling device. h Indicates the vertical positioning height of the sampling device. This represents the air pressure value sampled by the sampling device. R d This represents the specific gas constant of dry air. The specified temperature represents the air pressure reference value. g= 9.81 m / s 2 .

[0073] Specifically, the purpose of this formula is to eliminate the influence of the temperature and vertical positioning height h obtained by the sampling equipment on the air pressure value obtained by the sampling equipment. p ( T The influence of this, thus correcting it to a specified temperature with a uniform atmospheric pressure reference value. Standard atmospheric pressure value p ( T ref ).

[0074] Celsius temperature T Convert to Kelvin temperature T k To satisfy the thermodynamic requirements of the gas law; further calculate a correction factor related to altitude and temperature. b ,in g It is the acceleration due to gravity. R d This represents the specific gas constant of dry air. R d The effect of temperature gradient on the vertical distribution of air pressure can be quantified; finally, this coefficient is used to measure the air pressure values ​​sampled by the sampling equipment.p ( T ) The specified temperature is calibrated to the atmospheric pressure reference value. T ref Standard atmospheric pressure value p ( T ref ).

[0075] In this embodiment, the entire process is essentially a standardization of temperature and altitude, enabling meaningful comparison and analysis of air pressure data collected from different locations and environments on a common meteorological reference surface. This is the foundation for building a high-precision, consistent air pressure field model.

[0076] In one embodiment, the construction method further includes:

[0077] The confidence level of the air pressure baseline in any grid in the barometric fingerprint model is assigned, and the confidence level of the air pressure baseline of each grid is inversely proportional to the distance between the nearest sampling device.

[0078] Specifically, in the barometric fingerprint model, the confidence level assignment of the barometric baseline for each grid follows the inverse distance principle, meaning the confidence level is inversely proportional to the spatial distance from the grid to the nearest sampling device. This is because the measurement data from the sampling device is the most direct and reliable, and its influence decreases with increasing spatial distance. The closer the grid is to the sampling device, the stronger the support from the actual data from the sampling device, resulting in a higher confidence level. Conversely, grids farther from the sampling device rely more on mathematical estimations for their barometric baselines, making them more susceptible to interference from unmeasured environmental factors such as terrain undulations and local microclimates, thus receiving lower confidence levels.

[0079] For example, suppose a sampling device S1 is deployed in an industrial park. The confidence levels of grids at different distances from S1 are assigned as follows: Grid A is adjacent to S1 at a distance of 10 meters; its pressure baseline value can be directly derived from measurements by S1, therefore it is assigned a high confidence level, such as 0.95. Grid B is located in the middle of the influence range of S1 at a distance of 300 meters; its pressure baseline can be estimated by interpolation algorithms, which introduces some uncertainty, and it is assigned a medium confidence level, such as 0.75. Grid C is located at the edge of the influence range of S1 at a distance of 800 meters; its pressure baseline has the highest uncertainty and is assigned a low confidence level, such as 0.65.

[0080] In one embodiment, the construction method further includes:

[0081] Based on the sample information of any grid sampled by the terminal device, the confidence level of the air pressure baseline is adjusted. When the number of times the confidence level of the air pressure baseline of any grid is lower than the preset threshold is greater than or equal to the preset number, the air pressure baseline of the grid is updated.

[0082] Specifically, the terminal device reports a set of sample information, which may include air pressure data. The absolute difference between this set of air pressure data and the current air pressure baseline of the grid is calculated, and this difference is compared with a preset error range dynamically defined by the grid's environmental semantic labels. If the difference is within the tolerance range, the air pressure baseline is considered reliable, and the confidence level of the air pressure baseline is increased accordingly. Conversely, if the difference continuously exceeds the tolerance range, it indicates that the air pressure baseline may deviate from the actual environment, and the confidence level of the air pressure baseline is gradually lowered. This preset error range is not fixed but is defined by the environmental semantic labels assigned to the grid. The logic lies in the air pressure fluctuation characteristics of the physical environment represented by different semantic labels. For example, a grid marked as a wind vent, due to its inherent meteorological characteristics allowing for rapid and large air pressure fluctuations, can be assigned a wider preset error range (e.g., ±3hPa); a grid marked as indoors, with its expected very stable air pressure, can be assigned a narrower error range (e.g., ±0.5hPa).

[0083] It should be noted that when the confidence level of the air pressure baseline for a certain grid is found to be below a preset threshold (e.g., 0.5) for a number of consecutive or cumulative evaluation periods, or when this number exceeds a preset number (e.g., 3 times), the air pressure baseline is considered significantly inaccurate and can no longer effectively represent the true air pressure conditions of that grid. In this case, the air pressure baseline for that grid needs to be updated. This can typically be done by calculating and replacing the new air pressure baseline value for that grid based on the air pressure data recently reported by the terminal device, using algorithms such as weighted averaging or re-interpolation.

[0084] In one embodiment, when adjusting the confidence level of the air pressure baseline, the confidence level of the air pressure baseline decreases or increases according to a preset step size;

[0085] When the number of times the confidence level of the air pressure baseline of any grid is lower than the preset threshold is greater than or equal to the preset number of times, the air pressure baseline of the grid is updated to: the weighted average between the initial air pressure baseline fitted based on the sample information database and the air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline is adjusted to the first height. The first height is obtained by calculating the average of the height corresponding to the second height and the initial air pressure baseline. The second height is the height at which the terminal device samples the air pressure data. The confidence level of the air pressure baseline of the grid is increased to a value between the current air pressure baseline confidence level and the preset threshold for air pressure baseline confidence level.

[0086] If the confidence level of the air pressure baseline still decreases after the air pressure baseline is updated, the air pressure baseline of the grid will be updated to the average value of the partial air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline will be adjusted to the third height. The third height is the average height of the partial air pressure data sampled by the terminal device for calculating the average value. The partial air pressure data used by the terminal device for calculating the average value satisfies the following condition: the sampling time of the air pressure data is less than the current time threshold.

[0087] Specifically, by continuously comparing the difference between the air pressure value reported by the terminal device and the current air pressure baseline of the grid, the confidence level is fine-tuned according to a preset fixed step size (e.g., 0.05). When the difference is within the tolerance range defined by the environmental semantic label, the confidence level is increased; when it exceeds the range, the confidence level is decreased. When the air pressure baseline confidence level of any grid is detected to be below a preset threshold (e.g., 0.5) for a set number of consecutive evaluations (e.g., 3 times), a two-stage air pressure baseline and highly collaborative update process is initiated.

[0088] Taking grid F as an example, its initial state is a baseline air pressure of 1012.0 hPa, corresponding to a height of 1 meter, with a confidence level of 0.8. Due to the changes in the local environment caused by the construction of surrounding buildings, the user's usual path has changed (e.g., the opening of connecting corridors between buildings, or the opening of elevated bridges, users often use connecting corridors / elevated bridges and no longer or less frequently pass through the ground). The terminal device continuously collects air pressure data at an average height of 1010.0 hPa at a height of about 15 meters (because the terminal no longer collects ground air pressure data or collects it less frequently, the confidence level of the air pressure baseline at this location gradually decreases until it falls below the air pressure baseline confidence level threshold). Assuming that it continues to decrease after falling below the air pressure confidence level threshold, if the confidence level drops to 0.45 after three evaluations, the first stage update is triggered. At this stage, the air pressure baseline is updated to the weighted average of the initial baseline and the air pressure data reported by the terminal device (1012.0 + 1010.0) / 2 = 1011.0 hPa. Simultaneously, the corresponding height of the baseline is adjusted to the average of the initial height and the terminal sampling height (1 + 15) / 2 = 8 meters, forming a transitional benchmark. The confidence level is reset to the median value between the current value and the threshold (0.45 + 0.5) / 2 = 0.475. It should be noted that when calculating the weighted average between the initial air pressure baseline and the air pressure data reported by the terminal device, the weights of the two data points are adjustable; they can be equal or unequal.

[0089] If, after the update, the terminal continues to report a pressure value of 1010.2 hPa at an altitude of approximately 16 meters, which is still significantly different from the new baseline of 1011.0 hPa, causing the confidence level to further decrease to 0.4, then a more aggressive second-stage update will be initiated. This stage will rely entirely on recent data, selecting the average of all pressure data reported by the terminal within the last 24 hours (e.g., 1010.1 hPa) as the new pressure baseline, and synchronously updating the corresponding altitude to the average altitude at the time of data sampling (e.g., 15.5 meters). Through this gradual correction strategy, a smooth transition from the initial state (1012.0 hPa, 1 meter) to the final state (1010.1 hPa, 15.5 meters) is achieved, ensuring a re-match between pressure and altitude.

[0090] In one embodiment, such as Figure 2 As shown, the construction method also includes the following steps:

[0091] Step 201: Compare the sample information collected by the terminal device with the preset environmental semantic label template to obtain at least one environmental semantic label for a grid.

[0092] Specifically, feature extraction is performed on time-series sample information (air pressure data and temperature data) collected by the mobile terminal within a specific grid, meeting a preset requirement, to calculate feature values. The calculated feature values ​​are then matched against predefined environmental semantic label templates, where each template corresponds to a typical physical phenomenon (e.g., a vent template requires a rapid drop in air pressure and a simultaneous decrease in temperature within a short period). If the matching score exceeds a set threshold, a corresponding environmental semantic label (e.g., vent) is assigned to that grid.

[0093] Step 202: Based on the sample information collected by the terminal device within a specified time range, calculate the physical tag data. The physical tag data includes one or more of the following types: air pressure change rate, absolute value of air pressure change, temperature value, and temperature change rate.

[0094] Specifically, physical tag data can include one or more of the following types: rate of change of air pressure, absolute value of air pressure change, temperature value, and rate of change of temperature. The rate of change of air pressure can be obtained by calculating the ratio of the pressure difference between the start and end times within a time window to time. It is used to perceive the instantaneous trend and speed of air pressure changes and is a core indicator for identifying rapid phenomena such as wind vents. The absolute value of air pressure change refers to the difference between the maximum and minimum air pressure values ​​within the same time window. It is used to capture the overall amplitude of air pressure fluctuations while ignoring direction, helping to determine environmental stability. The rate of change of temperature is similar to the rate of change of air pressure and can reflect the rate of rise and fall of ambient temperature, which is crucial for identifying thermal processes such as heat sources.

[0095] Step 203: Compare the calculated physical label data with the physical label data in the environmental semantic label template. When the similarity is greater than the preset similarity threshold, mark the grid with the environmental semantic label corresponding to the environmental semantic label template. Based on the similarity, assign environmental semantic label confidence to the environmental semantic label of the grid. The environmental semantic label confidence is proportional to the similarity.

[0096] Specifically, the calculated physical label data (such as air pressure change rate, temperature change rate, etc.) is compared with the physical feature standards defined in the pre-stored environmental semantic label templates to calculate similarity. Each environmental semantic label template corresponds to a typical physical phenomenon; for example, the wind vent template may require the air pressure change rate to be below a specific negative threshold and the temperature change rate to be negative. Further, matching algorithms, such as rule-based scoring, can quantify the degree of agreement between the physical label data and the physical label data in the environmental semantic label templates, generating a similarity score.

[0097] When the similarity score exceeds a preset threshold, the observed data is considered sufficient to support the corresponding environmental semantic judgment, and the grid is labeled with the environmental semantic label corresponding to the template. Further, a confidence level is assigned to the newly labeled semantic label based on the similarity score. The core principle is that the confidence level is directly proportional to the similarity; that is, the higher the matching degree, the higher the confidence level. For example, if the similarity between the physical label data and the physical label data of the wind vent template is as high as 90%, the confidence level assigned to this label may be 0.9; if the similarity is just above the threshold (e.g., 65%), the confidence level may be only 0.65.

[0098] In one embodiment, such as Figure 3 As shown, based on the sample information collected by the terminal device and the horizontal positioning at the time of sampling, the air pressure baseline of the corresponding grid is adjusted. Based on the adjusted air pressure baseline, the confidence level of the environmental semantic label is adjusted, including the following steps:

[0099] Step 301: Based on the set environmental semantic labels, remove disturbance terms from the sample information sampled by the terminal device to obtain air pressure correction data containing altitude information;

[0100] Specifically, semantic labels (e.g., air vents and heat sources) are determined based on the horizontal position of the terminal device, and the type and intensity of environmental disturbances are judged by combining the label confidence scores. Furthermore, physical filtering models corresponding to the semantic labels are established; for example, high-pass filtering is used to remove rapid wind pressure fluctuations for the air vent labeled grid, and a temperature compensation algorithm is used to eliminate thermal buoyancy effects for the heat source labeled grid.

[0101] Step 302: Based on the air pressure correction data and combined with the adjusted air pressure baseline, calculate the vertical height positioning information when the terminal device samples the information;

[0102] Specifically, the measured air pressure value of the terminal device after removing environmental disturbances is converted into air pressure-altitude value with the standard air pressure baseline of the corresponding grid. The vertical height positioning information of the terminal device can be inferred from the air pressure difference.

[0103] For example, this process can be performed using the pressure-altitude formula: Δh = (RT / gM) × ln(P0 / P1). Where P0 is the grid baseline pressure, P1 is the terminal correction pressure, R is the gas constant, T is the thermodynamic temperature, g is the gravitational acceleration, and M is the molar mass of air.

[0104] It should be noted that, in terms of obtaining vertical height positioning information, in addition to the back-calculation of air pressure difference, multi-source data such as satellite positioning (such as GPS), inertial navigation or laser ranging can also be integrated.

[0105] Step 303: Based on the vertical height positioning information, remove the height information from the air pressure correction data and calculate the corrected air pressure value when the terminal device is at the height corresponding to the air pressure baseline.

[0106] Specifically, the calculation essentially converts the air pressure values ​​collected by the terminal device at any altitude into equivalent air pressure values ​​on the standard reference altitude plane defined by the air pressure baseline through a physical model. This eliminates the air pressure differences caused by the inconsistency between the height of the terminal device and the height of the reference plane, and generates comparable air pressure data with a unified altitude benchmark.

[0107] For example, based on the vertical difference (Δh) between the height of the terminal device and the reference height corresponding to the barometric pressure baseline, combined with the current ambient temperature, the formula P is used to calculate the vertical difference. corrected = P measured The pressure compensation is calculated using × exp(gMΔh / RT). Where P... measured The terminal correction pressure is given by: g is the acceleration due to gravity, M is the molar mass of air, R is the gas constant, and T is the thermodynamic temperature.

[0108] Step 304: Calculate the physical label data based on the corrected air pressure value, calculate the similarity between the calculated physical label data and the physical label data in the environmental semantic label template, and adjust the environmental semantic label confidence for the grid based on the similarity.

[0109] Specifically, based on the corrected air pressure value of the unified altitude benchmark, physical tag data is calculated through time series analysis, including but not limited to the rate of change of air pressure per unit time (reflecting the dynamic trend of air pressure), the absolute value of air pressure change within a specified time window (characterizing the fluctuation amplitude), and the rate of change of temperature (related to thermal disturbances).

[0110] Furthermore, the calculated physical label data is compared with the threshold range set in the predefined environmental semantic label template to perform multi-dimensional similarity calculation (e.g., using the cosine similarity algorithm) to quantify its matching degree with the typical features of various environmental phenomena (e.g., wind vents and heat sources).

[0111] If the similarity exceeds a preset threshold, the confidence of the corresponding semantic label is adjusted according to the direct proportional relationship between the similarity value and the confidence: high similarity increases the confidence of the environmental semantic label, while low similarity decreases the confidence of the environmental semantic label accordingly.

[0112] In one embodiment, multi-scenario applications based on barometric fingerprint models achieve intelligent decision-making by integrating grid-based barometric pressure baselines with environmental semantic tags: In the positioning field, by comparing measured barometric pressure with baseline data and using semantic tags to correct disturbances, the terminal's altitude positioning accuracy is improved; when planning paths for low-altitude aircraft, semantic tags are used to avoid areas prone to wind disturbances and dynamically adjust flight attitude to balance safety and efficiency; in environmental control, air conditioning systems utilize "vent" and "heat source" tags to intelligently adjust valves and temperature settings, achieving energy saving and comfort optimization; indoor navigation identifies tags such as "gentle slope" and "vent" to construct floor connections, and combines baseline correction to improve cross-floor positioning stability; multi-level parking garages use semantic tags to quickly match floors and entrances / exits, generating precise guidance paths for pick-up services; in tunnel and subway scenarios, seamless switching between indoor and outdoor positioning is achieved based on characteristic pressure change patterns; and for high-altitude operations, real-time monitoring of "vent" tags dynamically assesses risk levels and automatically adjusts work windows and safety strategies. These applications collectively demonstrate the ability of barometric fingerprint models to transform physical sensor data into intelligent environmental decisions.

[0113] In one embodiment, such as Figure 4 As shown, adjusting environmental semantic tags also includes the following steps:

[0114] Step 401: Compare the sample information collected by the terminal device in the latest time period with the preset environmental semantic label template to obtain the environmental semantic label of at least one grid and the corresponding environmental semantic label confidence.

[0115] Specifically, physical label data (such as air pressure change rate and temperature change rate) is obtained based on time-series sample information collected by terminal devices in a specific grid within the latest period (e.g., the past 10 minutes). The physical label data is then matched and calculated in real time with the physical label data in a preset environmental semantic label template library. The degree of conformity between the physical label data and the feature patterns of each template is quantified by a similarity algorithm. If the similarity score of a template exceeds a preset threshold, the corresponding environmental semantic label (e.g., wind vent) is assigned to the grid. The confidence score of the environmental semantic label is generated in a proportional relationship according to the similarity score (e.g., a confidence score of 0.9 corresponds to a similarity of 90%).

[0116] For example, in an indoor navigation application, the terminal device continuously collects air pressure data in the stairwell of a shopping mall, which shows a slow increase in air pressure. If the data is calculated to be highly similar to the slope template (e.g., similarity 85%), the grid is labeled with a slope tag (e.g., confidence 0.85) to assist in floor identification and path planning. If a sudden drop in air pressure and a simultaneous drop in temperature are detected in the lobby area, and the data matches the air vent template (e.g., similarity 92%), a high-confidence air vent tag (e.g., confidence 0.92) is marked to indicate that this is an entrance or ventilation area.

[0117] Step 402: Compare the newly generated environmental semantic label of the terminal device with the existing environmental semantic label of the grid. If the two labels are the same and the confidence of the newly generated environmental semantic label of the terminal device is greater than the first preset confidence threshold, then increase the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model.

[0118] Specifically, when a terminal device samples and generates a new environmental semantic label within a certain grid, it compares the new label with the existing semantic labels of that grid in the model: if the two labels are of the same type (e.g., both are wind vents), and the confidence of the new label is higher than a preset first confidence threshold (e.g., 0.7), then the confidence of the corresponding semantic label of that grid in the model is increased by a fixed step size or based on the difference ratio.

[0119] For example, in a multi-level parking system application, suppose the ramp area between the third and fourth floors is divided into grid P. The barometric fingerprint model has already labeled its environmental semantic tag as a gentle slope, with an initial confidence level of 0.6. When a car equipped with sensors enters the ramp, the terminal device continuously collects barometric pressure data (e.g., pressure slowly decreasing) and calculates feature values ​​to generate a new environmental semantic tag. The new tag is also determined to be a gentle slope, and its confidence level is calculated to be 0.82. Since the new tag is consistent with the existing tag, and the new tag's confidence level of 0.82 is higher than a first preset confidence threshold (e.g., set to 0.75), the confidence level of the gentle slope tag for grid P is increased from 0.6 to 0.7 in increments. This method makes the barometric fingerprint model more certain that the grid is a ramp area. When other terminal devices locate this area, they can more accurately determine that the vehicle is in a floor transition zone, thereby optimizing indoor navigation floor switching prompts or pick-up route planning.

[0120] Step 403: If the newly generated environmental semantic label of the terminal device does not match the existing environmental semantic label of the grid, or if the confidence of the newly generated environmental semantic label of the terminal device is less than the second preset confidence threshold, then reduce the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model.

[0121] For example, in a high-rise building construction scenario, suppose the tower crane operating area is divided into grid G. In the barometric fingerprint model, due to its open, high-altitude characteristics, it is labeled as a "wind vent" (e.g., initial confidence level 0.75). If, on a certain morning, a terminal device (e.g., a barometric pressure sensor installed in the tower crane cab) detects a sudden drop in air pressure of 1.5 hPa and a temperature drop of 0.8°C within 10 seconds, the newly generated environmental semantic label is "strong wind vent" (e.g., confidence level 0.85). Because the new and old label types are incompatible, the confidence level of the original "wind vent" label will be reduced from 0.75 to 0.65.

[0122] Alternatively, if the terminal experiences a brief sensor malfunction, generating a new label indicating a wind vent but with a confidence level of only 0.3 (e.g., below the second preset confidence threshold of 0.5), the original label's confidence level will also be reduced to 0.65. When the wind vent label's confidence level remains below the threshold, the barometric fingerprint model will automatically prompt adjustments to the tower crane's hoisting speed or a halt to high-altitude operations to avoid safety risks caused by wind disturbances. This mechanism, through dynamic confidence level adjustments, ensures that environmental semantic labels promptly reflect actual risk changes.

[0123] Step 404: If the confidence level of the environmental semantic label of any grid in the barometric fingerprint model is lower than the third preset confidence threshold within a specified time range, then delete the environmental semantic label of the corresponding grid in the barometric fingerprint model.

[0124] It should be noted that any two of the first, second, and third preset reliability thresholds may be equal or unequal.

[0125] As an optional implementation, the construction method provided in this application embodiment can also use deep learning models based on big data to improve the accuracy of the barometric fingerprint model.

[0126] Based on the same concept, this application also provides a system for constructing a barometric fingerprint model, which constructs a barometric fingerprint model based on the above-described method for constructing a barometric fingerprint model.

[0127] Specifically, the barometric fingerprint model construction system provided in this application can be an integrated construction system based on the aforementioned barometric fingerprint model construction method. This system acquires barometric pressure and temperature data with location information through sampling devices; further standardizes and corrects the data and constructs a sample information database; divides the data into grids according to the device's influence range and generates standard barometric pressure baselines for each grid using interpolation algorithms; matches the sampled data from the terminal device with environmental semantic templates, assigns semantic labels and confidence levels to the grids, and finally outputs a barometric fingerprint model that integrates the barometric pressure baselines and semantic labels, achieving fully automated management and dynamic updates from data acquisition and processing to model generation.

[0128] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a solid-state disk (SSD), etc.

[0129] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The above are merely preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A method for constructing a barometric fingerprint model, characterized in that, The construction method includes: Sampling is performed using sampling devices with known horizontal and vertical positioning to obtain sample information, which includes at least air pressure data and temperature data. Construct a sample information database, which includes multiple sample information items; Define the influence range of the sampling device, and perform grid division based on the influence range; Based on the sample information database, the air pressure baseline of each grid is fitted, and the air pressure baseline is the air pressure reference value of the grid at a specified temperature and specified altitude. Based on the horizontal positioning of the terminal device, sample information is sampled within any grid. The amount of sample information sampled by the terminal device meets a preset quantity. The sample information sampled by the terminal device is compared with a preset environmental semantic label template to obtain at least one environmental semantic label for the grid. A barometric fingerprint model is constructed, which includes the barometric baseline of each grid and the environmental semantic label of at least one grid.

2. The method for constructing a barometric fingerprint model according to claim 1, characterized in that, The construction method also includes: The confidence level of the air pressure baseline in any grid of the air pressure fingerprint model is assigned, and the confidence level of the air pressure baseline of each grid is inversely proportional to the distance between the nearest sampling device.

3. The method for constructing a barometric fingerprint model according to claim 2, characterized in that, The construction method also includes: Based on the sample information of any grid sampled by the terminal device, the confidence level of the air pressure baseline is adjusted. When the number of times the confidence level of the air pressure baseline of any grid is lower than a preset threshold is greater than or equal to a preset number, the air pressure baseline of the grid is updated.

4. The method for constructing a barometric fingerprint model according to claim 3, characterized in that, Adjusting the confidence level of the air pressure baseline based on sample information from arbitrary grid sampling by the terminal device includes: Based on the air pressure value sampled by the terminal device in any grid, the absolute difference between the air pressure value and the air pressure baseline of the corresponding grid is calculated. When the absolute difference is less than or equal to a preset error range, the confidence level of the air pressure baseline is increased; when the absolute difference is greater than the preset error range, the confidence level of the air pressure baseline is decreased. The preset error range is defined by the environmental semantic labels of the grid.

5. The method for constructing a barometric fingerprint model according to claim 4, characterized in that, When adjusting the confidence level of the air pressure baseline, the confidence level of the air pressure baseline decreases or increases according to a preset step size; When the number of times the confidence level of the air pressure baseline of any grid is lower than a preset threshold is greater than or equal to a preset number, the air pressure baseline of the grid is updated to: the weighted average between the initial air pressure baseline fitted based on the sample information database and the air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline is adjusted to a first height, which is obtained by calculating the average of the height corresponding to the initial air pressure baseline and the second height, which is the height at which the terminal device samples the air pressure data, and the confidence level of the air pressure baseline of the grid is increased to a value between the current air pressure baseline confidence level and the preset threshold for air pressure baseline confidence level; If the confidence level of the air pressure baseline still decreases after the air pressure baseline is updated, the air pressure baseline of the grid is updated to the average value of the partial air pressure data reported by the terminal device, and the height corresponding to the modified air pressure baseline is adjusted to the third height. The third height is the average height of the partial air pressure data sampled by the terminal device for calculating the average value. The partial air pressure data used by the terminal device for calculating the average value satisfies the following condition: the sampling time of the air pressure data is less than a preset time threshold from the current time.

6. The method for constructing a barometric fingerprint model according to claim 1, characterized in that, The construction method also includes: The sample information sampled by the terminal device is compared with a preset environmental semantic label template to obtain at least one environmental semantic label for the grid. Based on the sample information collected by the terminal device within a specified time range, physical tag data is calculated. The physical tag data includes one or more of the following types: air pressure change rate, absolute air pressure change value, temperature change rate, and temperature value. The calculated physical label data is compared with the physical label data in the environmental semantic label template. When the similarity is greater than a preset similarity threshold, the grid is labeled with the environmental semantic label corresponding to the environmental semantic label template. Based on the similarity, the environmental semantic label confidence score is assigned to the environmental semantic label of the grid. The environmental semantic label confidence score is proportional to the similarity.

7. The method for constructing a barometric fingerprint model according to claim 6, characterized in that, The construction method also includes: Based on the sample information collected by the terminal device and the horizontal positioning at the time of sampling, the air pressure baseline of the corresponding grid is adjusted, and the confidence level of the environmental semantic label is adjusted according to the adjusted air pressure baseline, including the following steps: Based on the set environmental semantic tags, perturbation items are removed from the sample information sampled by the terminal device to obtain barometric pressure correction data containing altitude information; Based on the air pressure correction data and the adjusted air pressure baseline, the vertical height positioning information when the terminal device samples the sample information is calculated. Based on the vertical height positioning information, the height information is removed from the air pressure correction data, and the corrected air pressure value when the terminal device is at the height corresponding to the air pressure baseline is calculated. Based on the corrected air pressure value, the physical label data is calculated, and the similarity between the calculated physical label data and the physical label data in the environmental semantic label template is calculated. Based on the similarity, the confidence level of the environmental semantic label for the grid is adjusted.

8. The method for constructing a barometric fingerprint model according to claim 6, characterized in that, The construction method also includes: The sample information collected by the terminal device in the latest time period is compared with the preset environmental semantic label template to obtain at least one environmental semantic label of the grid and the corresponding environmental semantic label confidence. The newly generated environmental semantic label of the terminal device is compared with the existing environmental semantic label of the grid. If the two labels are the same and the confidence of the newly generated environmental semantic label of the terminal device is greater than the first preset confidence threshold, the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model is increased. If the newly generated environmental semantic label of the terminal device does not match the existing environmental semantic label of the grid, or if the confidence of the newly generated environmental semantic label of the terminal device is less than the second preset confidence threshold, then the confidence of the environmental semantic label of the corresponding grid in the barometric fingerprint model is reduced. If the confidence level of the environmental semantic label of any grid in the barometric fingerprint model is lower than the third preset confidence threshold within a specified time range, then the environmental semantic label of the corresponding grid in the barometric fingerprint model is deleted. Among the first, second, and third preset reliability thresholds, any two preset reliability thresholds may be equal or unequal.

9. The method for constructing a barometric fingerprint model according to claim 1, characterized in that, Based on the sample information database, the horizontal air pressure is calculated using the following formula; In the formula, T This indicates the temperature sampled by the sampling device. h This indicates the vertical positioning height of the sampling device. This represents the air pressure value sampled by the sampling device. R d This represents the specific gas constant of dry air. The specified temperature represents the air pressure reference value. g =9.81 m / s 2 , T k Indicates Kelvin temperature. b Indicates the correction factor. p ( T ref The specified temperature represents the air pressure reference value. T ref The standard atmospheric pressure value below.

10. A system for constructing a barometric fingerprint model, characterized in that, The construction system constructs the barometric fingerprint model based on the barometric fingerprint model construction method according to any one of claims 1 to 9.

Images