Unmanned aerial vehicle system and method for traceability of atmospheric pollution source

Abstract

The invention relates to the technical field of air pollution source tracing, and discloses an unmanned aerial vehicle system and method for air pollution source tracing. The method comprises the following steps: collecting gas concentration distribution data of different spatial positions of a target area, and synchronously obtaining a meteorological parameter data set; determining a plurality of pollution diffusion stability factors based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter data set, and constructing an atmospheric pollution traceability spatial model; generating a pollution source search strategy of the unmanned aerial vehicle cluster based on the model, and dynamically adjusting the strategy according to gas concentration distribution data collected in real time; each unmanned aerial vehicle independently generates a local traceability decision by combining the environment characteristic data of the unmanned aerial vehicle and the adjacent unmanned aerial vehicle, and a collaborative decision result is obtained after consistency adjustment of a collaborative decision mechanism; the pollution source tracing response amount is determined according to the pollution source search strategy and the collaborative decision result, the pollution source positioning instruction is executed when the response amount exceeds the preset threshold value, and the comprehensiveness, accuracy and efficiency of pollution source tracing can be improved.

Description

Technical Field

[0001] This invention relates to the field of atmospheric pollution source tracing technology, specifically to an unmanned aerial vehicle (UAV) system and method for tracing atmospheric pollution sources. Background Technology

[0002] In atmospheric environmental protection, rapid and accurate location of pollution sources is a prerequisite for pollution control. Currently, tracing atmospheric pollution sources mainly relies on technologies such as ground monitoring stations, mobile monitoring vehicles, and satellite remote sensing. Although ground monitoring stations can achieve long-term, fixed-point monitoring, their coverage is limited by their fixed locations, making it difficult to comprehensively capture the spatial distribution differences of gas concentrations within a region. This is especially true in areas with complex terrain or sparse monitoring stations, where monitoring blind spots are prone to occur, making it impossible to obtain complete gas concentration distribution data in a timely manner. While mobile monitoring vehicles can move flexibly within a certain area, their limited accessibility makes them difficult to navigate complex terrains such as mountainous and forested areas, resulting in significant limitations in monitoring range and flexibility. Satellite remote sensing technology can achieve large-scale macroscopic monitoring, but its monitoring accuracy is limited by satellite resolution, making it difficult to acquire high-precision gas concentration distribution data over small areas. Furthermore, it is greatly affected by weather conditions; in situations with thick clouds or rainfall, monitoring cannot be carried out normally, failing to meet the needs for real-time and accurate source tracing. Existing source tracing methods often rely on simplistic approaches when addressing the impact of meteorological factors on pollution diffusion. They typically consider only a few meteorological parameters such as wind speed and direction, failing to comprehensively and systematically analyze the correlation between meteorological conditions and gas concentration distribution. This results in an inability to accurately determine the stability of pollution diffusion, leading to low accuracy in the constructed source tracing models and hindering effective guidance for pollution source searches. Furthermore, the search process often employs fixed search paths, lacking the ability to dynamically adjust based on real-time monitoring data, resulting in low search efficiency and an inability to quickly pinpoint pollution source locations. In addition, existing methods lack effective collaborative decision-making mechanisms for multi-device collaborative source tracing. Data sharing and collaborative decision-making among monitoring devices are difficult, preventing the effective integration of local monitoring data into global source tracing information. This further impacts the accuracy and efficiency of source tracing, failing to meet the urgent need for rapid and accurate source tracing in current precise air pollution control efforts. Summary of the Invention

[0003] The purpose of this invention is to provide a drone method for tracing the sources of air pollution, in order to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides a drone method for tracing the sources of air pollution, the method comprising: Collect gas concentration distribution data at different spatial locations in the target area and simultaneously acquire meteorological parameter datasets; Based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter dataset, multiple pollution diffusion stability factors are determined. A spatial model for tracing atmospheric pollution sources was constructed using the aforementioned multiple pollution diffusion stability factors. A pollution source search strategy for the UAV swarm is generated based on the aforementioned atmospheric pollution source tracing spatial model; The pollution source search strategy is dynamically adjusted based on real-time collected gas concentration distribution data; Each UAV independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of its neighboring UAVs; The local tracing decisions are adjusted for consistency through a collaborative decision-making mechanism to obtain collaborative decision-making results; The pollution source tracing response quantity is determined based on the pollution source search strategy and the collaborative decision-making results. When the pollution source tracing response exceeds the preset response threshold, a pollution source location command is executed.

[0005] Preferably, determining multiple pollution diffusion stability factors based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter dataset specifically includes: Analyze the spatial gradient variation characteristics of the gas concentration distribution data; Extract the wind speed distribution characteristics and wind direction stability parameters from the meteorological parameter dataset; The pollution diffusion stability factor is calculated based on the spatial gradient variation characteristics, wind speed distribution characteristics, and wind direction stability parameters.

[0006] Preferably, constructing an atmospheric pollution source tracing spatial model using the multiple pollution diffusion stability factors specifically includes: A three-dimensional pollution source tracing spatial coordinate system is constructed, where the X-axis represents the gas concentration gradient value, the Y-axis represents the wind speed influence factor, and the Z-axis represents the wind direction stability. The pollution diffusion stability factor is mapped to the three-dimensional pollution source tracing spatial coordinate system to form a source tracing spatial lattice. A pollution source probability distribution surface is generated based on the aforementioned source tracing spatial lattice.

[0007] Preferably, the pollution source search strategy for generating a drone swarm based on the atmospheric pollution source tracing spatial model specifically includes: Extract the coordinate set of high-probability pollution areas from the aforementioned spatial model for tracing atmospheric pollution sources; The search responsibility area of ​​the drone swarm is divided according to the coordinate set of the high-probability contamination area; Generate dynamic search path plans for each area of ​​responsibility.

[0008] Preferably, dynamically adjusting the pollution source search strategy based on real-time collected gas concentration distribution data specifically includes: Monitor the rate of change of current gas concentration distribution data relative to historical gas concentration distribution data; When the concentration change rate exceeds a preset change threshold, the pollution diffusion stability factor is recalculated. The atmospheric pollution source tracing spatial model is updated based on the recalculated pollution diffusion stability factor; The pollution source search strategy is revised based on the updated spatial model for tracing atmospheric pollution sources.

[0009] Preferably, each UAV independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of its neighboring UAVs, specifically including: The local environmental feature data and the environmental feature dataset of neighboring UAVs are fused to generate a local environmental feature vector; Calculate the probability distribution of pollution source directions based on the local environmental feature vectors; The direction with the highest probability is selected as the local tracing decision for the current drone.

[0010] Preferably, the consistency adjustment of the local tracing decision through a collaborative decision-making mechanism specifically includes: Collect local source tracing decisions from nearby drones to form a decision set; Calculate the deviation between the current UAV decision and the decision set; When the decision deviation exceeds a preset deviation threshold, a decision correction vector is generated; The aforementioned decision correction vector is applied to adjust the local source tracing decision of the current UAV.

[0011] Preferably, determining the pollution source tracing response quantity based on the pollution source search strategy and the collaborative decision-making result specifically includes: Extract the target area priority coefficient from the pollution source search strategy; Calculate the decision confidence level of the collaborative decision-making results; The pollution source tracing response quantity is generated based on the target area priority coefficient and the decision confidence level.

[0012] Preferably, when the pollution source tracing response exceeds a preset response threshold, executing the pollution source location command specifically includes: Activate the high-precision gas composition analyzer to collect data on the characteristic components of pollutants; A pollution source fingerprint map is generated based on the pollution characteristic component data; Mark the current spatial location as the coordinates of a suspected pollution source.

[0013] Preferably, the present invention also includes an unmanned aerial vehicle (UAV) system for tracing air pollution sources, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described UAV method for tracing air pollution sources.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This UAV method for tracing the sources of air pollution collects gas concentration distribution data at different spatial locations in the target area and simultaneously acquires meteorological parameter datasets. It can comprehensively grasp the spatial distribution characteristics of gas concentration in the target area and the meteorological conditions affecting pollution diffusion. Compared with traditional ground monitoring stations that can only acquire fixed-point data and mobile monitoring vehicles that are limited by terrain, UAVs can fly flexibly in different spatial locations, are not constrained by complex terrain and road conditions, and acquire more comprehensive and complete data. It can effectively avoid the existence of monitoring blind spots and provide more detailed basic data for subsequent source tracing work.

[0015] Based on the spatial distribution characteristics of gas concentration distribution data and meteorological parameter datasets, multiple pollution diffusion stability factors are identified. This approach overcomes the limitations of existing methods that only consider a few meteorological parameters. By analyzing the relationship between meteorological conditions and pollution diffusion from multiple dimensions, it can more accurately reflect the actual situation of pollution diffusion and provide strong support for constructing a precise spatial model for tracing atmospheric pollution sources. The spatial model for tracing atmospheric pollution sources, constructed using multiple pollution diffusion stability factors, better reflects actual pollution diffusion patterns. Compared to existing simple source tracing models, its accuracy is significantly improved, providing more effective scientific guidance for pollution source searching. A pollution source search strategy based on an atmospheric pollution source tracing spatial model generates a drone swarm. This strategy fully leverages the advantages of drone swarms; compared to single-drone searches, swarm operations can cover a wider area. Furthermore, through reasonable search strategy planning, redundant searches can be avoided, improving search efficiency. The pollution source search strategy is dynamically adjusted based on real-time collected gas concentration distribution data, making the search process more flexible and targeted. It can promptly adjust the search direction and range according to real-time changes in pollution concentration, avoiding the inefficiency caused by using fixed search paths and focusing more quickly on the pollution source area. Each drone independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of neighboring drones. This fully utilizes the real-time monitoring data of each drone to quickly form source tracing judgments for local areas, avoiding delays caused by centralized data transmission and processing. Through a collaborative decision-making mechanism, the local source tracing decisions are consistently adjusted, integrating the local decisions of each drone into a globally unified collaborative decision result. This effectively solves the problems of poor data sharing and difficulty in unifying decisions in existing multi-device collaboration, making source tracing decisions more comprehensive and accurate, and avoiding decision biases caused by the limitations of local data from a single drone. The pollution source tracing response quantity is determined based on the pollution source search strategy and collaborative decision-making results. When the response quantity exceeds the preset response threshold, the pollution source location command is executed. This ensures that the location is only performed after sufficient and reliable tracing information is obtained, avoiding the execution of the location command too early or too late, thus guaranteeing the accuracy of pollution source location. Compared with existing methods that lack clear response quantity judgment and are difficult to grasp the timing of location, this method can effectively improve the reliability of location, accurately lock the target for subsequent pollution control work, and help carry out air pollution control work efficiently. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the working principle of the UAV method for tracing atmospheric pollution sources as described in this invention. Figure 2 Flowchart illustrating the working principle for determining the pollution diffusion stability factor; Figure 3 A flowchart illustrating the working principle of a drone swarm pollution source search strategy. Figure 4 A flowchart illustrating the working principle of dynamically adjusting pollution source search strategies. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 This invention provides an unmanned aerial vehicle (UAV) system and method for tracing the sources of air pollution, the method comprising: Gas concentration distribution data at different spatial locations within the target area are collected, and meteorological parameter datasets are acquired simultaneously. Multiple pollution diffusion stability factors are determined based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter datasets. These factors are used to construct an atmospheric pollution source tracing spatial model, thereby generating a pollution source search strategy for the UAV swarm. The search strategy is dynamically adjusted based on real-time collected gas concentration distribution data. Each UAV independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of neighboring UAVs. A collaborative decision-making mechanism is used to consistently adjust the local source tracing decisions, obtaining collaborative decision results. The pollution source tracing response quantity is determined based on the pollution source search strategy and the collaborative decision results. When the pollution source tracing response quantity exceeds a preset response threshold, a pollution source location command is executed.

[0019] Example 1: See Figure 2 A swarm of drones, equipped with high-precision gas sensors and miniature weather stations, was deployed over the target area. These drones flew along a pre-planned grid pattern, covering a three-dimensional space from the ground to hundreds of meters in the air. Each drone synchronously collected data on the concentration of various gases at its location at a set frequency, including concentrations of characteristic pollutants such as sulfur dioxide, nitrogen oxides, and volatile organic compounds, while also recording wind speed, wind direction, temperature, humidity, and air pressure at that point. All data, accompanied by high-precision timestamps and spatial coordinates, was transmitted in real-time to the swarm's data processing center via a wireless data transmission module.

[0020] After the data acquisition phase, the system begins to analyze the spatial gradient variation characteristics of the gas concentration distribution data. The data processing center performs spatial interpolation on the acquired concentration data to generate a three-dimensional concentration distribution model of the target area. By analyzing this model, the system identifies the spatial variation patterns of concentration, including the degree of concentration change with distance and the distribution pattern of high-concentration areas. For example, in a certain downwind area, there may be a trend of concentration gradually increasing with distance, indicating a possible transport path of pollutants. The system calculates the concentration difference between adjacent sampling points and analyzes the spatial distribution patterns of these differences, thereby quantitatively describing the spatial non-uniformity and direction of change of the concentration field.

[0021] Simultaneously, a deep analysis of the meteorological parameter dataset is performed, with the system extracting wind speed distribution characteristics from the collected raw meteorological data. This includes not only instantaneous wind speed values ​​at each location, but more importantly, analyzing the spatial distribution pattern of wind speed, such as the variation of wind speed with increasing altitude and the differences in wind speed across different horizontal regions. The system calculates statistical indicators such as average wind speed and standard deviation of wind speed within a specific altitude layer to characterize the overall intensity and fluctuation of the wind field. At the same time, the system focuses on extracting wind direction stability parameters, calculating the frequency and amplitude of wind direction changes by analyzing wind direction data over continuous time series. Under relatively stable wind conditions, pollutant transport paths are more predictable; however, frequent changes in wind direction increase the uncertainty of pollutant diffusion direction.

[0022] Based on the spatial gradient variation characteristics, wind speed distribution characteristics, and wind direction stability parameters obtained from the above analysis, the system enters the calculation stage of the pollution diffusion stability factor. This calculation process is not a simple arithmetic operation, but a comprehensive assessment involving multiple coupled factors. The system performs correlation analysis between the spatial gradient variation characteristics and the wind speed distribution characteristics to assess the degree of influence of the wind field on the spatial distribution pattern of pollutants. Simultaneously, the moderating effect of the wind direction stability parameter on this relationship is considered. When the wind direction is stable, the relationship between wind speed and concentration gradient is clearer; however, when the wind direction is unstable, even with higher wind speeds, its transport effect on pollutants may be weakened due to the variable direction.

[0023] The system precisely calculates a pollution diffusion stability factor value for each sampling point or analysis unit. This factor value is not merely a numerical value, but rather a concrete reflection of the predictability and stability of pollutant diffusion behavior under the atmospheric environmental conditions at that sampling point or analysis unit. Specifically, a high factor value in a region usually indicates relatively stable wind direction, suitable wind speed, and a significant pollutant concentration gradient. Such regions often possess ideal conditions for source tracing, helping researchers track and locate pollution sources. Conversely, a low factor value in a region may be due to turbulent wind direction, excessively low or high wind speeds, or a relatively uniform distribution of pollutant concentrations. In such cases, locating pollution sources becomes more complex and difficult, increasing the difficulty and uncertainty of source tracing. Through this systematic calculation process, a set of pollution diffusion stability factors covering the target area is ultimately obtained. These factors form the basis for subsequently constructing a spatial model for atmospheric pollution source tracing. The entire implementation process embodies the technical characteristics of multi-source data fusion, spatiotemporal feature analysis, and comprehensive environmental parameter assessment, providing key input parameters for intelligent source tracing decisions by UAV swarms.

[0024] Example 2: See Figure 3The system first constructs a three-dimensional pollution source tracing spatial coordinate system. The design of this coordinate system fully considers the key factors affecting pollutant diffusion. The X-axis represents the gas concentration gradient value. This parameter is obtained by calculating the concentration difference between adjacent points in space, reflecting the uneven distribution of pollutants in space and their possible migration directions. The Y-axis represents the wind speed influence factor, which is obtained by normalizing measured wind speed data and is used to quantify the contribution of the wind field to the pollutant transport capacity. The Z-axis represents wind direction stability. Through statistical analysis of time-series wind direction data, the variation of wind direction over a certain period is calculated; the more stable the wind direction, the higher this value. This coordinate system constructs a feature space specifically for pollution source tracing analysis, where each dimension is directly related to the key influencing factors of pollutant diffusion behavior.

[0025] Mapping pollution diffusion stability factors to a three-dimensional coordinate system to construct a source-tracing spatial lattice requires sophisticated data processing techniques. This process involves the precise measurement and recording of three key parameters at each sampling point: gas concentration gradient, wind speed influence factor, and wind direction stability. These parameters determine the specific location of each sampling point in the coordinate system. In the three-dimensional coordinate system, each sampling point is precisely located to a coordinate position, thus forming a lattice. Each node in this lattice not only encompasses spatial coordinate information but also contains the specific value of the pollution diffusion stability factor at that point. These values ​​are a quantitative representation of the pollution diffusion characteristics at that point, which is of great significance for understanding and analyzing the pollution diffusion process. In this way, each node in the lattice becomes an information-rich data point, collectively constituting a digital representation of the atmospheric diffusion conditions in the target area. This representation method not only reveals the stability characteristics of pollution diffusion but also helps researchers understand the process and mechanism of pollution diffusion more intuitively. To connect the discrete lattice nodes into a continuous distribution surface, the system employs a spatial interpolation algorithm. This algorithm can deduce the values ​​of unknown nodes based on the values ​​of known nodes in a point matrix, thus transforming a discrete point matrix into a continuous surface. This continuous surface provides a complete characterization of the pollution diffusion conditions across the entire region, offering more comprehensive and accurate information on pollution diffusion.

[0026] By mapping pollution diffusion stability factors to a three-dimensional coordinate system and constructing a source-tracing spatial lattice, researchers can delve deeper into the processes and mechanisms of pollution diffusion. This digital representation not only improves the accuracy and efficiency of pollution diffusion research but also provides crucial data support for pollution control and management. Generating a pollution source probability distribution surface based on the source-tracing spatial lattice is the next important step. The system employs a spatial probability estimation algorithm to calculate the probability of a pollution source existing at each location in the three-dimensional coordinate system based on the stability factor values ​​and spatial distribution patterns of each node in the lattice. The probability calculation comprehensively considers multiple factors: in high-concentration gradient regions, if accompanied by appropriate wind speed and high wind direction stability, a higher probability of pollution source existence is assigned; while in regions with low concentration gradients or turbulent wind directions, the probability value decreases accordingly. The final generated pollution source probability distribution surface is a continuous probability field, where the height value of each point on the surface represents the probability of a pollution source existing under the corresponding atmospheric conditions at that location.

[0027] In extracting the coordinate set of high-probability pollution areas from the spatial model for tracing air pollution sources, the system operates according to a pre-set probability threshold. Specifically, the system meticulously scans the entire probability distribution surface to identify all areas whose probability values ​​exceed the preset threshold. Once these areas are identified, the system records their spatial coordinate range. Typically, these high-probability areas exhibit certain clustering characteristics in space, possibly due to one or more potential pollution emission sources. To better analyze and manage these high-probability areas, the system generates a spatial coordinate set for each identified area. This set includes not only the boundary coordinates of the area but also its probability distribution characteristics. In this way, the system can more accurately identify and track potential pollution sources, providing strong support for subsequent pollution control and environmental protection efforts.

[0028] Dividing the search responsibility area of ​​a drone swarm based on the coordinate set of high-probability contaminated areas requires comprehensive consideration of multiple factors. The system first analyzes the spatial distribution pattern of each high-probability area, merging spatially similar areas into larger search blocks. Then, based on the size of the drone swarm and the performance parameters of each drone, the entire search area is divided into several responsibility areas. During the division process, the size of each responsibility area should be proportional to its probability value; that is, areas with higher probabilities are allocated more refined search resources. Simultaneously, appropriate overlap zones are ensured between responsibility areas to avoid the emergence of search blind spots.

[0029] Generating dynamic search path plans for each responsibility area is the final crucial step. For each defined responsibility area, the system designs an adaptive search path based on the characteristics of the probability distribution surface within that area. In areas with high probability values, path points are set more densely, and the flight altitude is appropriately lowered to obtain more accurate environmental data; in areas with low probability values, a relatively sparse path point layout is used to improve search efficiency. The path planning also fully considers real-time weather conditions, especially changes in wind direction and speed, ensuring that the UAV is always in the optimal data acquisition position during the search process. The generated search path has dynamic adjustment characteristics and can automatically optimize the flight trajectory based on real-time acquired environmental data.

[0030] Example 3: See Figure 4 The implementation process begins with continuous monitoring of gas concentration distribution data. While flying according to a predetermined search strategy, the drone swarm continuously collects various gas concentration data at its location. This data is aggregated in real-time to a central processing system via a wireless transmission network. The system establishes a time-series database, recording the concentration values ​​at different time points for each spatial coordinate. The system calculates the rate of change of the current gas concentration distribution data relative to historical data. This calculation is performed independently for each sampling point, comparing the concentration differences at the same spatial location at different time points. The calculation of the rate of change considers not only the absolute difference but also the time interval factor, reflecting the rate and trend of concentration change.

[0031] The system sets a preset change threshold as a trigger condition, which is determined comprehensively based on the atmospheric characteristics, pollutant types, and monitoring accuracy requirements of the specific monitoring scenario. When the monitored concentration change rate exceeds this preset threshold, it indicates that the atmospheric environment or pollution emission status has changed significantly, and the existing pollution diffusion stability factor may no longer accurately reflect the current atmospheric conditions. At this time, the system initiates a recalculation process, re-executing the pollution diffusion stability factor calculation based on the latest concentration distribution data and real-time meteorological parameters. Although this recalculation process is consistent with the initial calculation in algorithmic logic, it uses a completely updated dataset, including all environmental parameters collected within the latest time window. The new calculation fully considers the latest changes in concentration distribution in both spatial and temporal dimensions, especially those areas showing significant gradient changes and points with abnormal concentration values. The calculation process also integrates meteorological conditions that may change simultaneously, including instantaneous fluctuations in wind speed and real-time changes in wind direction. Changes in these meteorological elements directly affect the diffusion path and rate of pollutants.

[0032] When processing the updated dataset, the system performs spatiotemporal alignment on the concentration distribution data to ensure comparability of data collected at different time points within a spatial reference frame. For meteorological parameters, the system pays particular attention to the temporal variation characteristics of wind speed and direction, analyzing their statistical distribution patterns and variability. In recalculating the pollution diffusion stability factor, the system employs a sliding time window analysis method, assigning higher weight to recent data to better reflect the dynamic trends of environmental conditions.

[0033] Based on the recalculated pollution diffusion stability factor, the system comprehensively updates the atmospheric pollution source tracing spatial model. This update process involves not only recalibrating the three-dimensional coordinate system but also dynamically adjusting the scale range of the coordinate axes to adapt to the possible variations in environmental parameters in the new dataset. The system remaps the updated pollution diffusion stability factor onto the adjusted three-dimensional coordinate system, forming a new source tracing spatial point distribution. Based on this updated point matrix, the system uses an improved interpolation algorithm to regenerate the pollution source probability distribution surface. The morphological characteristics and numerical distribution of the new surface reflect the latest status of the pollution source probability distribution under current atmospheric conditions.

[0034] The updated model may exhibit a high-probability region distribution pattern that is significantly different from the previous model. These changes may manifest in several ways: the spatial location of high-probability regions may shift, reflecting changes in the location of pollution sources or diffusion conditions; the distribution range of probability values ​​may change, indicating trends in pollution intensity; and the relative importance of multiple high-probability regions may be reordered, signifying a redistribution of pollution source contributions. The system will record these changes in detail and analyze the underlying environmental factors.

[0035] The entire model update process employs an incremental update mechanism, rapidly responding to environmental changes while maintaining model continuity. The system compares the differences between the old and new models, and when the differences exceed a certain threshold, a more in-depth model reconstruction process is triggered. This dynamic update mechanism ensures that the atmospheric pollution source tracing spatial model always remains highly consistent with the actual situation, providing accurate and reliable spatial reference for the search decisions of UAV swarms.

[0036] The updated model not only reflects the current environmental situation but also provides a data foundation for predicting the future development of pollution diffusion by analyzing model change trends. The system establishes a historical database of model updates, recording the timestamp, change characteristics, and environmental impact factors of each update. This historical data provides important data support for analyzing the spatiotemporal variation patterns of pollution sources. Based on the updated spatial model for tracing atmospheric pollution sources, the system modifies the pollution source search strategy of the UAV swarm. This modification process includes several aspects: re-extracting the coordinate set of high-probability pollution areas, which may change in location or range due to model updates; re-dividing the search responsibility areas of the UAV swarm and adjusting the responsibility area of ​​each UAV according to the new distribution of high-probability areas; and regenerating the dynamic search path plan for each responsibility area, optimizing the flight path and sampling point density according to the new probability distribution characteristics.

[0037] The entire dynamic adjustment process quantifies the concentration change rate and determines the threshold using the following formula: in: Indicates the rate of change in concentration. This indicates the current measured gas concentration. This represents a historical reference value for gas concentration. Indicates the time interval between two measurements. When The value exceeds the preset threshold At that time, the system triggers a recalculation and model update process.

[0038] This implementation method is characterized by establishing a closed-loop feedback mechanism from data monitoring to strategy adjustment. The system not only senses environmental changes but also automatically adjusts its monitoring strategies based on these changes, ensuring the drone swarm remains in an optimal search state. The implementation process considers the dynamic characteristics of atmospheric pollutant diffusion, effectively addressing the challenges of tracing pollution sources caused by changes in meteorological conditions or fluctuations in pollution emissions through continuous environmental perception and strategy optimization.

[0039] Example 4: Each drone continuously collects local environmental feature data of its location, including gas concentration readings, temperature, humidity, air pressure, and meteorological parameters such as wind speed and direction. Simultaneously, each drone receives environmental feature datasets from neighboring drones in real time via a wireless communication network. The proximity range is typically set based on communication distance and environmental complexity to ensure timely and effective information exchange. The drone fuses its own collected local environmental feature data with the received data from neighboring drones to generate a comprehensive local environmental feature vector. This fusion process uses a weighted average method, with weights assigned based on the time freshness of the data collection and spatial proximity. The most recently collected data and data from closer drones are given higher weights.

[0040] Based on the generated local environmental feature vectors, the UAV calculates the probability distribution of pollution source directions. This calculation process considers the combined influence of multiple environmental factors, including concentration gradient direction, prevailing wind direction, and temperature distribution characteristics. The system calculates the conditional probability of a pollution source existing for each possible direction angle, forming a complete probability distribution map. The probability calculation employs a pattern recognition method based on historical data and environmental features. The direction with the highest probability is selected as the local source tracing decision for the current UAV. This selection process considers not only the magnitude of the probability value but also the concentration of the probability distribution. When a certain direction shows a significantly higher probability value than other directions, the system determines it as the optimal search direction. The decision output includes the direction angle value and the corresponding probability confidence level, providing a reference for subsequent coordinated adjustments. Refer to Table 1 for local environmental feature data collected by the UAV at a certain moment.

[0041] Table 1: Local environmental characteristics data of UAVs.

[0042] A collaborative decision-making mechanism is used to optimize the consistency of local source tracing decisions, ensuring the coordination and accuracy of each UAV in the source tracing task. Each UAV first collects the local source tracing decisions of neighboring UAVs and integrates these decisions into a comprehensive decision set. When recording each decision, the direction angle and confidence level of each decision are recorded in detail for subsequent analysis and calculation. The deviation of the current UAV's decision from other decisions in the decision set is calculated. This calculation process considers not only the differences in direction angles but also the weighting effect of confidence levels to ensure the comprehensiveness and reliability of the calculation results. The deviation calculation uses a circular statistical method, which fully considers the periodicity of angle data and avoids calculation errors caused by angle periodicity. When the calculated decision deviation exceeds a preset deviation threshold, the system automatically generates a decision correction vector. The direction and magnitude of the correction vector are determined based on the overall tendency of the decision set and the deviation of the current decision. If the current UAV's decision deviates significantly from the group decision, the correction vector will guide the decision direction to adjust towards the group consensus. The setting of the correction intensity is positively correlated with the degree of deviation. This ensures both the appropriateness and effectiveness of the adjustment, and avoids decision distortion due to excessive correction.

[0043] Adjusting the local source tracing decisions of a drone using decision correction vectors is a comprehensive process. The adjusted decisions retain the unique environmental perception characteristics of individual drones while incorporating consensus information from the group decision-making process, achieving effective integration of individual and group aspects. This adjustment process is achieved through vector synthesis, ensuring a smooth transition and natural adjustment of the decision direction and avoiding instability caused by sudden decision changes. The adjusted decisions not only guide the next action of the current drone but also participate in the collaborative decision-making process of other drones, forming a dynamic and complementary decision-making network, further enhancing the collaborative source tracing efficiency of the entire drone swarm.

[0044] Example 5: The implementation process begins with the extraction of priority coefficients for target areas in the pollution source search strategy. These coefficients, determined in the previously established search strategy, are generated based on the probability distribution characteristics in the atmospheric pollution source tracing spatial model. Each defined search area is assigned a priority coefficient, reflecting the relative probability of a pollution source existing in that area. The assignment of priority coefficients considers multiple influencing factors, including the peak height of the area on the probability distribution surface, the spatial continuity and size of the area, and the relative position of the area to the prevailing wind direction. Areas with higher coefficient values ​​typically correspond to prominent peak areas on the probability surface or potential pollution source clusters pointed to by multiple environmental indicators. The system assesses the confidence level of the collaborative decision-making results generated by the UAV swarm. The calculation of decision confidence is based on the group decision-making characteristics formed during the collaborative decision-making process. The system analyzes the consistency of the decision directions of each UAV in the decision set, calculating the central tendency and dispersion of these directional data. When the decision directions of most UAVs in the swarm tend to be consistent, and their reported confidence values ​​are high, the system assigns a higher overall confidence level to the collaborative decision-making results. Conversely, if decision-making is decentralized or individual confidence levels are generally low, the overall confidence level will be lowered accordingly. Confidence assessment also considers environmental data quality factors, including the stability of sensor readings, the temporal synchronization of data acquisition, and the integrity of spatial coverage.

[0045] Based on the target area priority coefficient and decision confidence level, the system generates a pollution source tracing response quantity. This response quantity is a comprehensive quantitative indicator that organically combines the priority coefficient and decision confidence level through an algorithm. In the algorithm design, both the priority coefficient and decision confidence level have corresponding weight allocations, reflecting their respective importance in the final decision. The numerical range of the response quantity is standardized to place it within a unified dimensional system, facilitating comparison with a preset response threshold. The calculation of the response quantity is performed in real time, dynamically updating as environmental data and new decisions are continuously input. The system sets a preset response threshold as an action trigger condition. The determination of this threshold needs to comprehensively consider the specific needs of the actual application scenario, including the required level of accuracy in pollution source location, the acceptability of false alarms, and the accuracy standards of environmental monitoring. When the calculated pollution source tracing response quantity exceeds this preset threshold, it indicates that the system has accumulated sufficient evidence to support the judgment of the existence of the pollution source, and thus the execution program of the pollution source location command is initiated.

[0046] The execution of pollution source location commands involves several precise operational steps. First, the onboard high-precision gas composition analysis device is activated. These devices typically employ advanced technologies such as spectral analysis or mass spectrometry to precisely identify the components of pollutants in the atmosphere. The device collects pollution characteristic component data at the current location, obtaining the concentration levels and composition ratios of various pollutants. Based on this detailed composition data, the system generates a pollution source fingerprint map. The map construction uses feature extraction and pattern recognition technologies to transform complex composition data into a standardized set of feature vectors. These feature vectors uniquely characterize the emission characteristics of a specific pollution source. The system marks the current spatial location as the coordinates of a suspected pollution source. Coordinate recording uses a high-precision positioning system to accurately record latitude, longitude, altitude, and timestamp information. This coordinate data, along with simultaneously acquired environmental parameters, gas composition data, and the fingerprint map, is stored and transmitted to form a complete pollution source identification record.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used only 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 process, method, article, or apparatus.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A drone method for tracing the sources of air pollution, characterized in that, The method includes: Collect gas concentration distribution data at different spatial locations in the target area and simultaneously acquire meteorological parameter datasets; Based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter dataset, multiple pollution diffusion stability factors are determined. A spatial model for tracing atmospheric pollution sources was constructed using the aforementioned multiple pollution diffusion stability factors. A pollution source search strategy for the UAV swarm is generated based on the aforementioned atmospheric pollution source tracing spatial model; The pollution source search strategy is dynamically adjusted based on real-time collected gas concentration distribution data; Each UAV independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of its neighboring UAVs; The local tracing decisions are adjusted for consistency through a collaborative decision-making mechanism to obtain collaborative decision-making results; The pollution source tracing response quantity is determined based on the pollution source search strategy and the collaborative decision-making results. When the pollution source tracing response exceeds the preset response threshold, a pollution source location command is executed.

2. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, Based on the spatial distribution characteristics of the gas concentration distribution data and the meteorological parameter dataset, multiple pollution diffusion stability factors are determined, including: Analyze the spatial gradient variation characteristics of the gas concentration distribution data; Extract the wind speed distribution characteristics and wind direction stability parameters from the meteorological parameter dataset; The pollution diffusion stability factor is calculated based on the spatial gradient variation characteristics, wind speed distribution characteristics, and wind direction stability parameters.

3. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, The construction of an atmospheric pollution source tracing spatial model using the aforementioned multiple pollution diffusion stability factors specifically includes: A three-dimensional pollution source tracing spatial coordinate system is constructed, where the X-axis represents the gas concentration gradient value, the Y-axis represents the wind speed influence factor, and the Z-axis represents the wind direction stability. The pollution diffusion stability factor is mapped to the three-dimensional pollution source tracing spatial coordinate system to form a source tracing spatial lattice. A pollution source probability distribution surface is generated based on the aforementioned source tracing spatial lattice.

4. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, The pollution source search strategy for generating UAV swarms based on the aforementioned atmospheric pollution source tracing spatial model specifically includes: Extract the coordinate set of high-probability pollution areas from the aforementioned spatial model for tracing atmospheric pollution sources; The search responsibility area of ​​the drone swarm is divided according to the coordinate set of the high-probability contamination area; Generate dynamic search path plans for each area of ​​responsibility.

5. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, The specific steps of dynamically adjusting the pollution source search strategy based on real-time collected gas concentration distribution data include: Monitor the rate of change of current gas concentration distribution data relative to historical gas concentration distribution data; When the concentration change rate exceeds a preset change threshold, the pollution diffusion stability factor is recalculated. The atmospheric pollution source tracing spatial model is updated based on the recalculated pollution diffusion stability factor; The pollution source search strategy is revised based on the updated spatial model for tracing atmospheric pollution sources.

6. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, Each drone independently generates local source tracing decisions based on its own collected local environmental feature data and the environmental feature datasets of neighboring drones. Specifically, this includes: The local environmental feature data and the environmental feature dataset of neighboring UAVs are fused to generate a local environmental feature vector; Calculate the probability distribution of pollution source directions based on the local environmental feature vectors; The direction with the highest probability is selected as the local tracing decision for the current drone.

7. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, The consistency adjustment of the local tracing decision through the collaborative decision-making mechanism specifically includes: Collect local source tracing decisions from nearby drones to form a decision set; Calculate the deviation between the current UAV decision and the decision set; When the decision deviation exceeds a preset deviation threshold, a decision correction vector is generated; The aforementioned decision correction vector is applied to adjust the local source tracing decision of the current UAV.

8. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, Determining the pollution source tracing response quantity based on the pollution source search strategy and the collaborative decision-making results specifically includes: Extract the target area priority coefficient from the pollution source search strategy; Calculate the decision confidence level of the collaborative decision-making results; The pollution source tracing response quantity is generated based on the target area priority coefficient and the decision confidence level.

9. The UAV method for tracing air pollution sources as described in claim 1, characterized in that, When the pollution source tracing response exceeds a preset response threshold, the execution of the pollution source location command specifically includes: Activate the high-precision gas composition analyzer to collect data on the characteristic components of pollutants; A pollution source fingerprint map is generated based on the pollution characteristic component data; Mark the current spatial location as the coordinates of a suspected pollution source.

10. A drone system for tracing air pollution sources, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the UAV method for tracing atmospheric pollution sources as described in any one of claims 1 to 9.

Images