Mobile VOCs rapid detection and positioning system

Abstract

The invention relates to the field of environment monitoring, and discloses a mobile VOCs rapid detection and positioning system, which comprises a data acquisition module, a data fusion module, a concentration gradient analysis module, a Bayesian filtering module and a particle filtering optimization module, the data acquisition module is used for acquiring VOCs concentration data and multi-source positioning data; the data fusion module outputs high-precision position information for the positioning data; the concentration gradient analysis module is used for calculating a concentration gradient based on the VOCs concentration data and meteorological information and simulating concentration field distribution; the Bayesian filtering module dynamically updates the probability distribution of the pollution source position in combination with the concentration field distribution and the fusion position information; and the particle filtering optimization module optimizes the Bayesian filtering result. The system is suitable for various complex environments, can realize rapid detection of VOCs and high-precision positioning of pollution sources, and has good real-time performance, precision and environmental adaptability.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and in particular to a mobile VOCs rapid detection and positioning system. Background Art

[0002] In the existing technology, mobile volatile organic compounds (VOCs) detection and positioning systems are widely used in environmental monitoring, industrial park emission monitoring and other fields. However, the current VOCs detection system still faces many challenges in achieving fast and accurate pollution source positioning in complex environments. Most traditional systems rely on a single type of positioning method, such as using only GPS for positioning, which is difficult to ensure accuracy in obstructed areas or indoor environments. Even if some systems introduce wireless signal positioning or SLAM technology, the lack of fusion processing of multi-source data, the limitations of a single data source still lead to unstable and inaccurate positioning results, which is difficult to meet the precise positioning requirements in complex environments.

[0003] In addition, the concentration analysis methods for locating pollution sources in the existing technology are relatively simple, usually relying on direct concentration measurement, and not fully considering the impact of meteorological factors (such as wind speed and wind direction) on the diffusion of pollutants, resulting in a lack of accuracy in the concentration distribution model and difficulty in accurately indicating the direction of the pollution source. At the same time, most traditional detection systems rely on fixed sampling frequencies and positioning modes, and cannot be adaptively adjusted according to environmental changes, resulting in insufficient system response when concentrations fluctuate violently, or excessive power consumption when concentrations change slowly, thereby affecting detection efficiency and system endurance.

[0004] Therefore, the defects of the existing technology are mainly reflected in: single positioning method, insufficient accuracy of concentration distribution model, and lack of adaptive sampling and positioning control mechanism. These deficiencies limit the reliability and flexibility of the system in complex environments in practical applications. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a mobile VOCs rapid detection and positioning system, which can realize a high-precision, fast-response VOCs detection and positioning system in a complex environment. Through multi-source data fusion, concentration gradient analysis and adaptive control, it can achieve accurate positioning of the pollution source and optimization of system power consumption.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a mobile VOCs rapid detection and positioning system, including a data acquisition module, a data fusion module, a concentration gradient analysis module, a Bayesian filter module and a particle filter optimization module;

[0007] The data acquisition module is used to collect VOCs concentration data and positioning data, and the positioning data includes GPS data, wireless signal data and SLAM data;

[0008] The data fusion module is connected to the data acquisition module, and is used to receive the positioning data, fuse the positioning data through Kalman filtering and weighted average algorithm, and output fused high-precision position information;

[0009] The concentration gradient analysis module is connected to the data acquisition module, and is used to receive VOCs concentration data, and determine the concentration field distribution and concentration gradient direction based on concentration gradient calculation and diffusion model simulation;

[0010] The Bayesian filtering module is used to dynamically estimate the possible location of the pollution source based on the concentration field distribution output by the concentration gradient analysis module and the location information output by the data fusion module;

[0011] The particle filter optimization module is used to gradually approach the location of the pollution source using a particle filter algorithm based on the output of the Bayesian filter module.

[0012] Preferably, the data acquisition module includes a GPS receiver, a wireless signal receiver, a SLAM system and a VOCs sensor; wherein:

[0013] The GPS receiver is used to obtain the geographical coordinate information of the system in an open area and output positioning data;

[0014] The wireless signal receiver is used to receive signal strengths from at least three wireless signal sources at known locations, and calculate the device location by measuring the relationship between signal strength and distance to the signal source;

[0015] The SLAM system includes a laser radar or a camera, which is used to construct an environmental map in real time in an unknown environment and determine the relative position information of the device, and output SLAM positioning coordinates;

[0016] The VOCs sensor is used to collect volatile organic compound concentration data in the environment in real time and record it synchronously with the positioning data to support the calculation of concentration gradient and the construction of concentration field distribution.

[0017] Preferably, the data fusion module includes a Kalman filter and a weighted average fusion unit; wherein:

[0018] The Kalman filter performs data fusion based on the location information obtained by the GPS receiver, the wireless signal receiver and the SLAM system;

[0019] The weighted average fusion unit is used to perform weighted average fusion on the final position based on the high-precision position information output by the Kalman filter in combination with the precision weight of each data source.

[0020] Preferably, the precision weight of each data source is calculated based on the error standard deviation of each data source, and the calculation formula is:

[0021]

[0022] Among them, w i represents the weight of data source i, σ i Represents the standard deviation of the error of data source i.

[0023] Preferably, the concentration gradient analysis module is used to calculate the concentration gradient based on the concentration data and position information collected by the VOCs sensor to infer the direction of the pollution source; the calculation of the concentration gradient includes the following steps:

[0024] Obtain VOCs concentration data C(x,y,t) collected by mobile devices at different locations;

[0025] Calculating concentration gradients The concentration gradient is defined as:

[0026]

[0027] in, and Respectively represent the rate of change of VOCs concentration in the x-axis and y-axis directions;

[0028] Identify the possible direction of the pollution source based on the direction of the concentration gradient. Specifically, the direction of the concentration gradient is the direction in which the VOCs concentration increases fastest, and perform mobile detection along this direction.

[0029] Preferably, the concentration gradient analysis module further includes a diffusion equation calculation unit for simulating the concentration field distribution to improve the accuracy of the concentration field and consider the influence of wind speed and wind direction; the calculation of the diffusion equation calculation unit is as follows:

[0030] The diffusion of pollutant concentration over time is modeled based on the advection-diffusion equation:

[0031]

[0032] Where D is the diffusion coefficient, is the Laplace operator of concentration, represents the wind speed vector, S(x,y) represents the intensity of the pollution source;

[0033] The diffusion equation calculation unit is based on the wind speed vector The direction and intensity of the dynamic adjustment of the advection term in the diffusion model to reflect the impact of wind speed and direction on pollutant dispersion;

[0034] The advection-diffusion equation is solved by finite difference through numerical solution method to obtain the spatial distribution and variation trend of pollutant concentration in the region, thus generating a concentration field distribution model;

[0035] The concentration field distribution model is combined with the concentration gradient direction to provide accurate concentration field information for the subsequent positioning module, which is used for further estimation and positioning optimization of the pollution source location.

[0036] Preferably, the Bayesian filtering module is used to dynamically estimate the possible location of the pollution source based on the output of the concentration gradient analysis module and the high-precision location information of the data fusion module, and recursively update the probability distribution of the pollution source location; the steps of the Bayesian filtering module are as follows:

[0037] State initialization: Set the initial position probability distribution p(x0) of the pollution source, where x0 represents the initial position state;

[0038] State prediction: At each time step k, the prior probability distribution of the pollution source location is predicted through the state transition model. The state prediction formula is:

[0039] p(x k |z 1:k-1 )=∫p(x k |x k-1 )p(x k-1 |z 1:k-1 )dx k-1

[0040] Among them, p(x k |x k-1 ) represents the state transition probability distribution of the pollution source location, p(x k-1 |z 1:k-1 ) is the posterior distribution of the previous step, z 1:k-1 Represents all observation data from the initial moment to the current moment;

[0041] Observation update: Based on the observed concentration gradient and concentration field distribution data, the probability distribution of the pollution source location is updated. The observation update formula is:

[0042]

[0043] Among them, p(z k |x k ) means at position x k The observation probability at the moment depends on the current observation concentration data z k , p(z k |z 1:k-1 ) represents a normalization constant;

[0044] Observation probability calculation: The observation probability p(zk |x k ) Calculate based on the concentration field model and the direction of the concentration gradient to ensure that the closer the concentration value is to the high concentration area, the higher the observation probability value, thereby strengthening the speculation of the pollution source direction;

[0045] Recursive iteration: The Bayesian filter module dynamically estimates the location of the pollution source through the iteration of state prediction and observation update at each time step, and generates the posterior probability distribution p(x k |z 1:k );

[0046] Output pollution source location distribution: The Bayesian filter module outputs the posterior probability distribution of the pollution source location at the current moment, and provides the distribution to the particle filter optimization module for further particle filter optimization and precise positioning of the pollution source location.

[0047] Preferably, the particle filter optimization module is used to accurately estimate the location of the pollution source through a particle filter algorithm based on the posterior probability distribution of the pollution source location output by the Bayesian filter module; the steps of the particle filter optimization module are as follows:

[0048] Particle initialization: Generate N particles in the system detection area, each particle represents a possible pollution source location hypothesis Where i = 1, 2, ..., N;

[0049] Weight update: At each time step k, the weight of each particle is updated based on the posterior probability distribution of the pollution source location output by the Bayesian filter module and the current observation data. To update, the weight update formula is:

[0050]

[0051] in, Indicated in particles The current concentration data z is observed at the position k probability;

[0052] Observation probability calculation: Observation probability Based on the concentration value and concentration gradient direction at the current particle position, the concentration field model generated by the concentration gradient analysis module is used for estimation, and a higher weight is given to particles located in high concentration areas;

[0053] Weight normalization: weight of each particle Normalization is performed so that the sum of the weights of all particles is 1. The normalization formula is:

[0054]

[0055] Resampling: Resampling is performed according to the normalized weights, retaining high-weight particles and removing low-weight particles, thereby concentrating the particles in the high-probability area of ​​the posterior probability distribution and improving the accuracy of the pollution source location estimation;

[0056] State update and iteration: Repeat the above weight update and resampling steps. Through multiple rounds of iterations, the particles gradually approach the actual location of the pollution source, achieving an accurate estimation of the location of the pollution source.

[0057] Output pollution source location estimation: The particle filter optimization module outputs the pollution source location estimation result at the current moment, and uses the result for subsequent positioning display and data recording.

[0058] Preferably, the system further comprises an adaptive control module for adaptively adjusting the sampling frequency and positioning mode of the system according to environmental conditions and concentration changes.

[0059] The present invention also provides a mobile VOCs rapid detection and positioning method, comprising the following steps:

[0060] Collect VOCs concentration data and positioning data, the positioning data including GPS data, wireless signal data and SLAM data;

[0061] Perform Kalman filtering and weighted average fusion on the positioning data to output high-precision fused position information;

[0062] Calculate the concentration gradient based on VOCs concentration data and meteorological information, and simulate the concentration field distribution through the diffusion model;

[0063] Update the probability distribution of pollution source locations based on concentration field distribution and fused location information;

[0064] Particle filtering is performed on the Bayesian filtering results to gradually approach the actual pollution source location;

[0065] Dynamically adjust the system's sampling frequency and positioning mode.

[0066] The present invention provides a mobile VOCs rapid detection and positioning system. It has the following beneficial effects:

[0067] 1. The present invention realizes high-precision positioning in complex environments through multi-source data fusion of GPS, wireless signals and SLAM systems. The combination of Kalman filtering and weighted average algorithm effectively reduces the error caused by a single data source, so that the system can provide stable location information in open, blocked and indoor environments, meeting the requirements of mobile detection for positioning accuracy.

[0068] 2. The present invention can effectively identify the possible direction of pollution sources by calculating the gradient of VOCs concentration and simulating the diffusion trend of pollutants in combination with meteorological information. The combination of concentration gradient analysis and diffusion model ensures the accuracy of the concentration field, enables the system to accurately track concentration changes, and supports directional positioning of pollution sources.

[0069] 3. The present invention dynamically updates the probability distribution of the pollution source location through the Bayesian filtering algorithm, so that the system can continuously optimize the positioning results by combining the prior position and the new observation data. Bayesian filtering is particularly suitable for complex scenes with environmental changes and can enhance the stability of the system in locating pollution sources in dynamic environments.

[0070] 4. By optimizing the Bayesian filtering results through particle filtering, the present invention can gradually approach the actual pollution source location under nonlinear and uncertain conditions. The resampling mechanism of particle filtering makes the system gradually focus on the high probability area, thereby improving the positioning accuracy and realizing the accurate tracking of the pollution source location.

[0071] 5. The present invention integrates a variety of positioning and detection technologies, enabling the system to adapt to a variety of complex environments such as open, blocked and indoor. By flexibly selecting data sources and automatically switching positioning modes, the system can adjust its working mode in time when the environment changes, and has good environmental adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a system schematic diagram of the present invention;

[0073] Figure 2 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0075] Please refer to the attached Figure 1 , an embodiment of the present invention provides a mobile VOCs rapid detection and positioning system, which is intended to quickly detect and accurately locate volatile organic compounds (VOCs) through mobile devices. The system integrates multiple sensors and data processing algorithms, and can obtain VOCs concentration information in real time under complex environments and locate the location of pollution sources. The system includes a data acquisition module, a data fusion module, a concentration gradient analysis module, a Bayesian filter module, a particle filter optimization module, and an adaptive control module; the various modules of the mobile VOCs rapid detection and positioning system of the present invention will be described in detail below.

[0076] Data acquisition module

[0077] In an embodiment of the present invention, the data acquisition module is used to obtain positioning data and VOCs concentration data, providing basic data support for subsequent concentration gradient analysis, data fusion and pollution source positioning. The main task of the data acquisition module is to obtain the current location of the mobile device and the VOCs concentration information in the surrounding environment in real time. The module integrates multiple data sources, including GPS receivers, wireless signal receivers, SLAM systems and VOCs sensors, to ensure that reliable positioning and concentration information can be provided in different environments.

[0078] In this embodiment, the data acquisition module includes a GPS receiver, a wireless signal receiver, a SLAM system and a VOCs sensor. The specific functions and implementation methods of each component are as follows:

[0079] 1. GPS receiver

[0080] In this embodiment, the GPS receiver is used to provide absolute positioning data in an open area. The GPS receiver calculates the geographic coordinates (x) of the current device by receiving satellite signals. GPS ,y GPS ). When the device is in an unobstructed outdoor area, the GPS signal quality is high and can provide a stable absolute position reference for the system. However, in areas where the signal is blocked (such as around high-rise buildings or indoors), the GPS signal quality may be reduced, so this module also integrates other positioning methods to improve positioning accuracy and adaptability.

[0081] 2. Wireless signal receiver

[0082] In this embodiment, the wireless signal receiver is used to obtain positioning data indoors or in an obstructed area where the GPS signal is unstable. The wireless signal receiver receives signals from at least three wireless signal sources (such as Wi-Fi or Bluetooth beacons) with known locations, and estimates the distance d to each signal source by measuring the signal strength. i , and then determine the relative position of the device through triangulation. The implementation steps of this positioning method are as follows:

[0083] First, the device receives wireless signals from three known location signal sources and records the receiving strength RRS of each signal. i ;

[0084] According to the attenuation model of distance and signal strength, the signal strength RRS i Convert to distance d i ;

[0085] Based on the triangulation method, the relative position of the device (xWiFi ,y WiFi ).

[0086] 3.SLAM system

[0087] In this embodiment, the SLAM (Simultaneous Localization and Mapping) system is used to perform self-positioning in unknown or complex environments (such as industrial areas and indoor areas). The SLAM system integrates a laser radar or camera to build an environmental map of the current area and determine the relative position of the device in the map by scanning feature points in the surrounding environment in real time. The SLAM system is not only suitable for areas where GPS signals are unstable, but can also dynamically update the environmental map as the device moves, providing relative position coordinates (x SLAM ,y SLAM ). The positioning accuracy in the SLAM system depends on the recognition effect of environmental feature points, so the accuracy is higher in scenes with rich feature points.

[0088] 4. VOCs sensor

[0089] In this embodiment, the VOCs sensor is used to monitor the concentration of volatile organic compounds in the environment in real time. The VOCs sensor is installed on a mobile device and can continuously collect the VOCs concentration C(x, y, t) at the location of the device, and record the correspondence between the concentration data and the location information for subsequent concentration gradient analysis and concentration field modeling. The VOCs concentration data collected by the sensor provides basic support for the simulation of the concentration field and the speculation of the pollution source direction.

[0090] The data acquisition module in this embodiment can adaptively acquire positioning and concentration data in different environments. Specifically, the GPS receiver provides basic location information in an open environment, the wireless signal receiver provides supplementary positioning data in indoor areas where the signal is blocked, and the SLAM system builds an environmental map in an unknown and complex environment to locate the device in real time. Through the combined application of multiple sensors, the data acquisition module can acquire high-precision positioning data and concentration data in real time, providing sufficient data support for subsequent data processing and positioning analysis.

[0091] Data fusion module

[0092] In the embodiment of the present invention, the main function of the data fusion module is to fuse the positioning data from multiple sources to obtain a more accurate and stable device location. The data acquisition module provides positioning data from the GPS receiver, wireless signal receiver and SLAM system, and the accuracy, applicability and error characteristics of each of these data are different. The data fusion module fuses each positioning data through Kalman filtering and weighted average algorithm, effectively reducing the error caused by a single data source, and providing high-precision location information support for subsequent concentration gradient analysis and pollution source positioning.

[0093] In this embodiment, the data fusion module includes a Kalman filter and a weighted average fusion unit. The specific implementation is as follows:

[0094] 1. Kalman filter

[0095] In this embodiment, the Kalman filter is used to process the positioning data from the GPS, wireless signals and SLAM system, and obtains relatively stable position information by filtering and smoothing the errors of each data source.

[0096] State definition: Define the state vector of the system as Where (x, y) is the location coordinate of the device. The speed of the device.

[0097] State transfer equation: The system state is predicted by the state transfer equation of Kalman filtering. The state transfer equation is as follows:

[0098] X k =A·X k-1 +B·U k +W k

[0099] in:

[0100] X k is the current state vector;

[0101] A is the state transfer matrix, defined as:

[0102]

[0103] Where Δt is the time interval;

[0104] B is the control input matrix, defined as:

[0105]

[0106] U k is the control input;

[0107] W k is the process noise.

[0108] Observation equation: The Kalman filter converts the collected positioning data into observation values ​​through the observation equation, and updates the estimation results in combination with the predicted state. The observation equation is as follows:

[0109] Z k =H·X k +V k

[0110] in:

[0111] Z k is the observed value;

[0112] H is the observation matrix, which is dynamically adjusted according to the positioning accuracy of different data sources;

[0113] V k is the observation noise, which reflects the measurement errors of each data source.

[0114] Prediction step: The Kalman filter first predicts the current state through the state transfer equation. The prediction steps include:

[0115] X k|k-1 =A·X k-1|k-1

[0116] P k|k-1 =A·P k-1|k-1 ·A T +Q

[0117] Among them, P k|k-1 is the prediction error covariance matrix, and Q is the process noise covariance matrix.

[0118] Update step: After obtaining the observation value, the Kalman filter updates the state. The update steps include:

[0119] K k =P k|k-1 ·H T ·(H·P k|k-1 ·H T +R) -1

[0120] X k|k =X k|k-1 +K k ·(Z k -H·X k|k-1 )

[0121] P k|k =(IK k ·H)·P k|k-1

[0122] in:

[0123] Kk is the Kalman gain;

[0124] R is the observation noise covariance matrix;

[0125] I is the identity matrix.

[0126] The prediction and update process of the Kalman filter effectively smooths the positioning errors of each data source, providing high-quality position information input for subsequent weighted average fusion.

[0127] 2. Weighted average fusion unit

[0128] In this embodiment, the weighted average fusion unit further performs weighted average on the positioning data from GPS, wireless signals and SLAM system on the basis of Kalman filtering to obtain the final high-precision position estimation. The weighted average algorithm assigns different weights to each data source, and the weight value is based on the error standard deviation σ of each data source. i The specific formula is as follows:

[0129]

[0130] in:

[0131] w i represents the weight of data source i;

[0132] σ i Represents the standard deviation of the error of data source i.

[0133] The final fused device position (x, y) is calculated by weighted average of each data source, and the formula is as follows:

[0134] (x,y)=w GPS ·(x GPS ,y GPS )+w WiFi ·(x WiFi ,y WiFi )+w SLAM ·(x SLAM ,y SLAM )

[0135] in:

[0136] (x GPS ,y GPS )、(x WiFi ,y WiFi )、(x SLAM ,y SLAM ) are the location information of the GPS receiver, the wireless signal receiver and the SLAM system respectively;

[0137] w GPS 、x WiFi and wSLAM is the weight of each data source.

[0138] Through the Kalman filter and smooth weighted average algorithm, the data fusion module effectively integrates information from multiple data sources and provides the system with more accurate location information.

[0139] Concentration Gradient Analysis Module

[0140] In the embodiment of the present invention, the main function of the concentration gradient analysis module is to perform gradient analysis on the concentration data collected by the VOCs sensor to determine the direction and trend of change in the pollutant concentration, thereby inferring the approximate direction of the pollution source. The data acquisition module provides VOCs concentration information at different locations. Through the concentration gradient analysis module, the system can calculate the rate of change of concentration with spatial position and generate a concentration distribution model. At the same time, the module also combines the advection diffusion model to simulate the diffusion process of pollutants in the environment, especially considering the influence of wind speed and wind direction.

[0141] In this embodiment, the concentration gradient analysis module includes a concentration gradient calculation unit and a diffusion model simulation unit, and the specific implementation method is as follows:

[0142] 1. Concentration gradient calculation unit

[0143] In this embodiment, the concentration gradient calculation unit is used to calculate the spatial gradient of the concentration based on the concentration data collected by the VOCs sensor to infer the direction of the pollution source.

[0144] Concentration data acquisition: First, the system collects VOCs concentration data C(x, y, t) at different locations through the data acquisition module, and records the concentration value at the corresponding position (x, y) and time t.

[0145] Concentration gradient calculation: Based on the concentration data at each location, the system calculates the concentration gradient of the area The concentration gradient is defined as:

[0146]

[0147] in:

[0148] represents the rate of change of concentration in the x direction;

[0149] It represents the rate of change of concentration in the y direction.

[0150] Gradient direction judgment: The direction of the concentration gradient is the direction in which the concentration changes most rapidly, so it can be used as an estimate of the direction pointing to the pollution source. In actual operation, the size and direction of the concentration gradient can be used as a reference for judging the direction of equipment movement, that is, the concentration may gradually increase in the direction pointed by the concentration gradient, and it is inferred that this direction is the possible location of the pollution source.

[0151] 2. Diffusion model simulation unit

[0152] In this embodiment, the diffusion model simulation unit is used to consider the influence of the external environment (such as wind speed and wind direction) on the diffusion of pollutants, and simulate the concentration distribution of VOCs in the environment through the advection diffusion model to obtain a more accurate concentration field distribution model.

[0153] Advection-diffusion equation: This module simulates the diffusion process of pollutants based on the advection-diffusion equation. The mathematical expression of the advection-diffusion equation is:

[0154]

[0155] in:

[0156] C represents the pollutant concentration;

[0157] D is the diffusion coefficient, which reflects the rate at which pollutants diffuse in the air;

[0158] is the Laplace operator of concentration, representing the diffusion term;

[0159] Represents the wind speed vector, whose direction and magnitude affect the advection diffusion direction of the concentration;

[0160] S(x,y) is the source term, which represents the intensity and location of the pollution source.

[0161] Influence of wind speed and direction: To further improve the accuracy of the concentration distribution model, the system adjusts the advection term in the model according to the wind speed and direction data of the current environment. Dynamic adjustment. Wind speed vector The direction and size of the wind determine the speed and direction of pollutant propagation in the environment. For example, when the wind speed is high, the diffusion speed of pollutants in the concentration field will increase, and the direction of the concentration gradient may also adjust with the change of wind speed.

[0162] Numerical solution: Since the advection-diffusion equation is a partial differential equation, the system uses the finite difference method to numerically solve it. After discretizing the regional space and time, the difference method is used to obtain the numerical solution of the concentration at each discrete point, thereby obtaining the concentration distribution model of the pollutant in the entire area.

[0163] Through concentration gradient calculation and diffusion model simulation, the concentration gradient analysis module can provide directional information on pollutant concentration changes, providing a reliable basis for concentration field distribution for locating pollution sources.

[0164] Bayesian Filter Module

[0165] In the embodiment of the present invention, the main function of the Bayesian filter module is to dynamically estimate the possible location of the pollution source based on the concentration field distribution output by the concentration gradient analysis module and the high-precision location information provided by the data fusion module. Since the location of the pollution source may change under different environmental conditions, it is necessary to iteratively update the probability distribution of the pollution source location through Bayesian filtering. The Bayesian filter module uses the recursive relationship between the prior probability and the observed data to predict and update the pollution source location at each time step, providing a stable pollution source location estimation basis for the subsequent particle filter optimization module.

[0166] In this embodiment, the Bayesian filter module includes a state prediction unit and an observation update unit, and the specific implementation method is as follows:

[0167] 1. State prediction unit

[0168] In this embodiment, the state prediction unit is used to predict the prior probability distribution of the pollution source position at the current moment based on the pollution source position estimation at the previous moment and the state transition model.

[0169] State initialization: When the system starts running, the initial position probability distribution p(x0) of the pollution source is set, where x0 represents the initial state of the pollution source. The initial position distribution can be set based on known information (such as the possible area of ​​the pollution source) or assumed to be uniformly distributed.

[0170] State transition model: The state prediction unit calculates the prior distribution based on the transfer law of pollution sources between different time steps. The state transition equation is expressed as:

[0171] p(x k |z 1:k-1 )=∫p(x k |x k-1 )p(x k-1 |z 1:k-1 )dx k-1

[0172] in:

[0173] x k Indicates the location status of the pollution source at the current moment;

[0174] p(x k |x k-1 ) is the state transition probability, describing the change of the pollution source location between time k-1 and k;

[0175] p(x k-1 |z 1:k-1 ) is the posterior distribution of the previous moment;

[0176] z 1:k-1 Represents all observation data from the initial moment to the current moment.

[0177] The state prediction unit calculates the prior probability distribution p(x k |z 1:k-1 ), providing a basis for subsequent processing by the observation update unit.

[0178] 2. Observation update unit

[0179] In this embodiment, the observation updating unit updates the probability distribution of the pollution source location based on the observation data at the current moment. By combining the current observation data, the observation updating unit updates the prior distribution to the posterior distribution, further improving the accuracy of the pollution source location estimation.

[0180] Observation model: The observation update unit uses the observation model p(z k |x k ) represents the relationship between the pollution source location and the observed data, where z k Represents the observation data at the current moment. The observation model is determined based on the observation results of concentration gradient and concentration field distribution. Near the pollution source, the observed concentration data is usually higher, so the observation model will assign corresponding probability weights according to the concentration value.

[0181] 3. Recursive Iteration

[0182] The Bayesian filter module dynamically updates the probability distribution of the pollution source location through the recursive process of state prediction and observation update. At each time step, the system continuously adjusts the posterior distribution of the pollution source location through the prediction of the state prediction unit and the update of the observation update unit, generating a gradually convergent pollution source location estimate.

[0183] Through the iterative process of state prediction and observation update, the Bayesian filtering module can maintain a dynamic estimate of the location of pollution sources in a variety of environments, enhancing the accuracy of the system's pollution source positioning under uncertain conditions.

[0184] Particle filter optimization module

[0185] In the embodiment of the present invention, the main function of the particle filter optimization module is to further optimize the pollution source location through a non-parametric particle filter algorithm based on the posterior probability distribution of the pollution source location output by the Bayesian filter module. Particle filtering is a filtering algorithm suitable for nonlinear and non-Gaussian environments. It samples and estimates the possibility of the pollution source location through multiple particles, and gradually converges to the true pollution source location through iterative updates of particle weights and resampling. The role of this module is to further improve the accuracy and stability of the pollution source location estimation and provide optimized results for the final positioning.

[0186] In this embodiment, the particle filter optimization module includes four units: particle initialization, weight update, resampling and state update. The specific implementation method is as follows:

[0187] 1. Particle initialization

[0188] In this embodiment, the particle filter optimization module first generates N particles in the possible pollution source area, each particle represents a possible pollution source location hypothesis, and uses represents the position of the i-th particle, i = 1, 2, ..., N. The initial distribution of particles can be set to a uniform distribution or based on the known preliminary pollution source location information. By sampling a large number of particles in the area, the particle filter optimization module can cover all possible pollution source locations and provide comprehensive spatial information for subsequent estimation.

[0189] 2. Weight Update

[0190] In this embodiment, the particle filter optimization module assigns a weight to each particle according to the posterior probability distribution of the Bayesian filter module and the observation data at the current moment. The weight reflects the degree of match between the particle position and the actual pollution source position, and the calculation formula is:

[0191]

[0192] in:

[0193] is the weight of the i-th particle at the current moment;

[0194] Represents the particle position under the observation model With the observed data z k The matching probability of

[0195] is the weight of the particle at the previous moment.

[0196] In calculation When the system is based on the concentration field data generated by the concentration gradient analysis module, if the particle is located in a high concentration area, its observation probability is high and its weight will increase accordingly. In this way, the particle weight update process will gradually increase the weight of particles located in the possible pollution source area.

[0197] 3. Weight Normalization

[0198] In this embodiment, in order to ensure that the sum of the weights of all particles is 1, the particle filter optimization module normalizes the weights. The normalization formula is:

[0199]

[0200] After normalization, the weight of a particle can be used as the relative likelihood of the location of the pollution source it represents.

[0201] 4. Resampling

[0202] In this embodiment, the particle filter optimization module resamples the particles according to the normalized weights to retain high-weight particles and remove low-weight particles. The resampling process is to avoid particle degradation, that is, to gradually reduce the number of low-weight particles during iteration, so that the particles are concentrated near the pollution source. The steps of resampling are as follows:

[0203] Perform probability sampling according to the weight of particles and select particles with high weight;

[0204] Redistribute the positions of low-weight particles to concentrate them near high-weight particles, thereby increasing the particle density around the pollution source.

[0205] 5. Status Update and Iteration

[0206] In this embodiment, the particle filter optimization module repeats the iterative process of weight updating and resampling at each time step. Through continuous iteration, the particle filter gradually converges to the actual location of the pollution source, that is, the location of the high-weight particles will be closer and closer to the actual location of the pollution source.

[0207] Adaptive Control Module

[0208] In the embodiment of the present invention, the main function of the adaptive control module is to automatically adjust the sampling frequency and positioning mode of the system according to the dynamic changes of environmental conditions and VOCs concentration data, so as to optimize the real-time performance and power consumption performance of the system. The environmental data and concentration information provided by the data acquisition module and the concentration gradient analysis module provide basic data support for adaptive control. The adaptive control module monitors the positioning signal quality and concentration change rate in real time to determine whether the current detection environment has changed, and adjusts the sampling frequency and positioning mode according to different environmental conditions to ensure the flexibility and efficiency of the system in different scenarios.

[0209] In this embodiment, the adaptive control module includes an environment determination unit, a positioning mode switching unit, a sampling frequency adjustment unit and a dynamic feedback unit, and the specific implementation method is as follows:

[0210] 1. Environmental identification unit

[0211] In this embodiment, the environment determination unit is used to determine the type of the current detection environment according to the quality of the positioning signal, so as to decide whether to switch the positioning mode. The unit mainly determines whether the current environment is an open area, a blocked area or an indoor environment based on the GPS signal strength, the wireless signal quality and the feature point matching rate of the SLAM system.

[0212] GPS signal identification: When the GPS signal strength is high and stable, the system determines that the current environment is an open area;

[0213] Wireless signal strength determination: When the GPS signal is weak or unstable, but the wireless signal strength is high, the system determines that the current environment is a blocked area;

[0214] SLAM feature point identification: When both GPS and wireless signals are weak, but the SLAM feature point matching rate is high, the system determines that the current environment is indoor or complex.

[0215] 2. Positioning mode switching unit

[0216] In this embodiment, the positioning mode switching unit automatically adjusts the positioning mode according to the environmental discrimination result to ensure the positioning accuracy and applicability of the system in different environments.

[0217] Open areas: GPS positioning data is preferred in open areas because it provides an absolute position reference;

[0218] Blocked area: In the blocked area, wireless signal ranging data is used first, and relative positioning is performed by receiving strength information of multiple signal sources;

[0219] Indoor or complex environments: SLAM systems are used preferentially in indoor or complex environments. Positioning is performed by matching environmental feature points to ensure that the system can maintain high positioning accuracy in a variety of environments.

[0220] 3. Sampling frequency adjustment unit

[0221] In this embodiment, the sampling frequency adjustment unit is used to dynamically adjust the sampling frequency according to the change rate of the VOCs concentration to balance the real-time performance and power consumption of the system.

[0222] Concentration change rate calculation: The system calculates the change rate of VOCs concentration at each sampling moment To determine the severity of the concentration change, the calculation formula is as follows:

[0223]

[0224] Among them, C t and C t-1 are the concentration values ​​at the current moment and the previous moment respectively, and Δt is the sampling time interval.

[0225] Adaptive adjustment of sampling frequency: When the concentration change rate exceeds the preset threshold, the system automatically increases the sampling frequency to more accurately capture the concentration change trend; when the concentration change rate is low, the sampling frequency is reduced to save power. The adaptive adjustment formula of the sampling frequency f is as follows:

[0226]

[0227] Among them, f base is the basic sampling frequency, and k is the proportional coefficient, which is used to adjust the relationship between the sampling frequency and the concentration change rate.

[0228] 4. Dynamic feedback unit

[0229] In this embodiment, the dynamic feedback unit is used to feed back the adjustment result of the adaptive control module to the data acquisition module and the data fusion module, so as to ensure that the acquisition and processing process of the entire system can adapt to environmental changes in real time.

[0230] When the environment discrimination unit identifies environmental changes, the dynamic feedback unit immediately sends a positioning mode switching instruction to the data acquisition module to adjust the method of acquiring positioning data;

[0231] When the sampling frequency adjustment unit detects a significant change in the concentration change rate, the dynamic feedback unit sends a sampling frequency adjustment signal to the data acquisition module to ensure that the sampling frequency adapts to the concentration change requirements in real time.

[0232] Through the coordinated control of environmental discrimination, positioning mode switching, sampling frequency adjustment and dynamic feedback, the adaptive control module ensures the flexibility and efficiency of the system in complex environments, and effectively improves the overall performance of the VOCs rapid detection and positioning system.

[0233] Through multi-source data fusion and adaptive control, the present invention can operate stably in complex and changeable environments, automatically adjust the positioning mode and sampling frequency to optimize power consumption and real-time performance. Concentration gradient analysis is combined with diffusion model, Bayesian filtering and particle filtering to gradually optimize positioning accuracy, so that the system can efficiently identify the location of pollution sources, providing reliable technical support for environmental monitoring and emergency response.

[0234] Please refer to the attached Figure 2 The present invention also provides a mobile VOCs rapid detection and positioning method, comprising the following steps:

[0235] Step S1: First, the system obtains VOCs concentration data and positioning data through the data acquisition module. Positioning data includes GPS data, wireless signal data and SLAM data to ensure that stable location information can be provided under different environmental conditions. These data provide basic support for subsequent concentration analysis and positioning processing.

[0236] Step S2: The system performs Kalman filtering and weighted average fusion processing on the collected positioning data. The Kalman filter smoothes the data to reduce the error of a single data source, while the weighted average combines the accuracy of each data source, assigns different weights to the data source, and outputs accurate fused location information.

[0237] Step S3: Based on VOCs concentration data and meteorological information, the system calculates the concentration gradient and determines the direction of concentration change. Subsequently, the concentration field distribution is simulated through the diffusion model to predict the diffusion trend of pollutants in the environment and provide concentration field distribution information for pollution source location.

[0238] Step S4: Based on the concentration field distribution and the fused location information, the system uses Bayesian filtering to update the probability distribution of the pollution source location, and gradually optimizes the pollution source location estimation in combination with the new observation data to reflect the possible location area of ​​the pollution source.

[0239] Step S5: Particle filter optimization is performed on the Bayesian filter result. Particle filter makes the system gradually approach the actual pollution source position by weight adjustment and resampling of multiple particle positions, further improving the positioning accuracy.

[0240] Step S6: The adaptive control module dynamically adjusts the sampling frequency and positioning mode of the system according to the changes in environmental conditions and concentration data. This adjustment enables the system to maintain the best balance between real-time performance and power consumption in different scenarios, ensuring detection efficiency and reliability.

[0241] Through the above steps, this method realizes the VOCs concentration detection of mobile devices and the precise positioning of pollution sources. It has strong adaptability, high accuracy, and can operate stably under various environmental conditions.

[0242] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A mobile VOCs rapid detection and positioning system, characterized in that: It includes data acquisition module, data fusion module, concentration gradient analysis module, Bayesian filtering module and particle filtering optimization module; The data acquisition module is used to collect VOCs concentration data and positioning data, and the positioning data includes GPS data, wireless signal data and SLAM data; The data fusion module is connected to the data acquisition module, and is used to receive the positioning data, fuse the positioning data through Kalman filtering and weighted average algorithm, and output fused high-precision position information; The concentration gradient analysis module is connected to the data acquisition module, and is used to receive VOCs concentration data, and determine the concentration field distribution and concentration gradient direction based on concentration gradient calculation and diffusion model simulation; The Bayesian filtering module is used to dynamically estimate the possible location of the pollution source based on the concentration field distribution output by the concentration gradient analysis module and the location information output by the data fusion module; The particle filter optimization module is used to gradually approach the location of the pollution source using a particle filter algorithm based on the output of the Bayesian filter module.

2. A mobile VOCs rapid detection and positioning system according to claim 1, characterized in that: The data acquisition module includes a GPS receiver, a wireless signal receiver, a SLAM system and a VOCs sensor; wherein: The GPS receiver is used to obtain the geographical coordinate information of the system in an open area and output positioning data; The wireless signal receiver is used to receive signal strengths from at least three wireless signal sources at known locations, and calculate the device location by measuring the relationship between signal strength and distance to the signal source; The SLAM system includes a laser radar or a camera, which is used to construct an environmental map in real time in an unknown environment and determine the relative position information of the device, and output SLAM positioning coordinates; The VOCs sensor is used to collect volatile organic compound concentration data in the environment in real time and record it synchronously with the positioning data to support the calculation of concentration gradient and the construction of concentration field distribution.

3. A mobile VOCs rapid detection and positioning system according to claim 1, characterized in that: The data fusion module includes a Kalman filter and a weighted average fusion unit; wherein: The Kalman filter performs data fusion based on the location information obtained by the GPS receiver, the wireless signal receiver and the SLAM system; The weighted average fusion unit is used to perform weighted average fusion on the final position based on the high-precision position information output by the Kalman filter in combination with the precision weight of each data source.

4. A mobile VOCs rapid detection and positioning system according to claim 3, characterized in that: The precision weight of each data source is calculated based on the error standard deviation of each data source, and the calculation formula is: Among them, w i represents the weight of data source i, σ i Represents the standard deviation of the error of data source i.

5. The mobile VOCs rapid detection and positioning system according to claim 1 is characterized in that: The concentration gradient analysis module is used to calculate the concentration gradient based on the concentration data and position information collected by the VOCs sensor to infer the direction of the pollution source; the calculation of the concentration gradient includes the following steps: Obtain VOCs concentration data C(x,y,t) collected by mobile devices at different locations; Calculating concentration gradients The concentration gradient is defined as: in, and Respectively represent the rate of change of VOCs concentration in the x-axis and y-axis directions; Identify the possible direction of the pollution source based on the direction of the concentration gradient. Specifically, the direction of the concentration gradient is the direction in which the VOCs concentration increases fastest, and perform mobile detection along this direction.

6. A mobile VOCs rapid detection and positioning system according to claim 5, characterized in that: The concentration gradient analysis module also includes a diffusion equation calculation unit for simulating the concentration field distribution to improve the accuracy of the concentration field and consider the influence of wind speed and wind direction; the calculation of the diffusion equation calculation unit is as follows: The diffusion of pollutant concentration over time is modeled based on the advection-diffusion equation: Where D is the diffusion coefficient, is the Laplace operator of concentration, represents the wind speed vector, S(x,y) represents the intensity of the pollution source; The diffusion equation calculation unit is based on the wind speed vector The direction and intensity of the dynamic adjustment of the advection term in the diffusion model to reflect the impact of wind speed and direction on pollutant dispersion; The advection-diffusion equation is solved by finite difference through numerical solution method to obtain the spatial distribution and variation trend of pollutant concentration in the region, thus generating a concentration field distribution model; The concentration field distribution model is combined with the concentration gradient direction to provide accurate concentration field information for the subsequent positioning module, which is used for further estimation and positioning optimization of the pollution source location.

7. The mobile VOCs rapid detection and positioning system according to claim 1, characterized in that: The Bayesian filtering module is used to dynamically estimate the possible location of the pollution source based on the output of the concentration gradient analysis module and the high-precision location information of the data fusion module, and recursively update the probability distribution of the pollution source location; the steps of the Bayesian filtering module are as follows: State initialization: Set the initial position probability distribution p(x0) of the pollution source, where x0 represents the initial position state; State prediction: At each time step k, the prior probability distribution of the pollution source location is predicted through the state transition model. The state prediction formula is: p(x k |z 1：k-1 )=∫p(x k |x k-1 )p(x k-1 |z 1：k-1 )dx k-1 Among them, p(x k |x k-1 ) represents the state transition probability distribution of the pollution source location, p(x k-1 |z 1：k-1 ) is the posterior distribution of the previous step, z 1：k-1 Represents all observation data from the initial moment to the current moment; Observation update: Based on the observed concentration gradient and concentration field distribution data, the probability distribution of the pollution source location is updated. The observation update formula is: Among them, p(z k |x k ) means at position x k The observation probability at the moment depends on the current observation concentration data z k , p(z k |z 1：k-1 ) represents a normalization constant; Observation probability calculation: The observation probability p(z k |x k ) Calculate based on the concentration field model and the concentration gradient direction to ensure that the closer the concentration value is to the high concentration area, the higher the observation probability value, thereby strengthening the speculation of the pollution source direction: Recursive iteration: The Bayesian filter module dynamically estimates the location of the pollution source through the iteration of state prediction and observation update at each time step, and generates the posterior probability distribution p(x k |z 1：k ); Output pollution source location distribution: The Bayesian filter module outputs the posterior probability distribution of the pollution source location at the current moment, and provides the distribution to the particle filter optimization module for further particle filter optimization and precise positioning of the pollution source location.

8. A mobile VOCs rapid detection and positioning system according to claim 1, characterized in that: The particle filter optimization module is used to accurately estimate the location of the pollution source through a particle filter algorithm based on the posterior probability distribution of the pollution source location output by the Bayesian filter module; the steps of the particle filter optimization module are as follows: Particle initialization: Generate N particles in the system detection area, each particle represents a possible pollution source location hypothesis Where i = 1, 2, ..., N; Weight update: At each time step k, the weight of each particle is updated based on the posterior probability distribution of the pollution source location output by the Bayesian filter module and the current observation data. To update, the weight update formula is: in, Indicated in particles The current concentration data z is observed at the position k probability; Observation probability calculation: Observation probability Based on the concentration value and concentration gradient direction at the current particle position, the concentration field model generated by the concentration gradient analysis module is used for estimation, and a higher weight is given to particles located in high concentration areas; Weight normalization: weight of each particle Normalization is performed so that the sum of the weights of all particles is 1. The normalization formula is: Resampling: Resampling is performed according to the normalized weights, retaining high-weight particles and removing low-weight particles, thereby concentrating the particles in the high-probability area of ​​the posterior probability distribution and improving the accuracy of the pollution source location estimation; State update and iteration: Repeat the above weight update and resampling steps. Through multiple rounds of iterations, the particles gradually approach the actual location of the pollution source, achieving an accurate estimation of the location of the pollution source. Output pollution source location estimation: The particle filter optimization module outputs the pollution source location estimation result at the current moment, and uses the result for subsequent positioning display and data recording.

9. The mobile VOCs rapid detection and positioning system according to claim 1, characterized in that: The system also includes an adaptive control module for adaptively adjusting the sampling frequency and positioning mode of the system according to environmental conditions and concentration changes.

10. A mobile VOCs rapid detection and positioning method, applied to the mobile VOCs rapid detection and positioning system according to any one of claims 1 to 9, characterized in that: The following steps are involved: Collect VOCs concentration data and positioning data, the positioning data including GPS data, wireless signal data and SLAM data; Perform Kalman filtering and weighted average fusion on the positioning data to output high-precision fused position information; Calculate the concentration gradient based on VOCs concentration data and meteorological information, and simulate the concentration field distribution through the diffusion model; Update the probability distribution of pollution source locations based on concentration field distribution and fused location information; Particle filtering is performed on the Bayesian filtering results to gradually approach the actual pollution source location; Dynamically adjust the system's sampling frequency and positioning mode.

Images