Pollution source rapid positioning system and method based on underway monitoring data and traceability algorithm

Abstract

The invention discloses a pollution source rapid positioning system and method based on underway monitoring data and a traceability algorithm. The method comprises the steps of water quality underway monitoring data collection, data preprocessing, traceability algorithm model construction, traceability algorithm model optimization, pollution source positioning and result display. When the underway monitoring equipment carries out water quality tracing, a dynamic tracking mechanism is adopted, and the monitoring equipment can move autonomously by constructing a tracing algorithm, so as to adapt to the position change of a pollution source and carry out rapid positioning on the pollution source; energy consumption-precision balance of the underway monitoring equipment in the sailing process is considered, energy consumption penalty is introduced into a set reward function, invalid movement is avoided, and the traceability efficiency of the monitoring equipment is improved; and in combination with a historical monitoring data sequence, the change trend characteristics of water quality along with time are reflected, the accuracy of model training is improved, water quality traceability work can be accurately and quickly performed, and high efficiency and accuracy of water quality traceability are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile monitoring, and more specifically, to a rapid pollution source location system and method based on mobile monitoring data and a traceability algorithm. Background Art

[0002] With the acceleration of the industrialization process and the improvement of the urbanization level, the environmental pollution problem has become increasingly severe. Traditional pollution source monitoring methods often have problems such as long monitoring cycles, lagging data, inaccurate positioning, etc., and it is difficult to meet the urgent needs of current environmental protection work. In particular, although fixed monitoring stations can provide continuous pollution data, their coverage is limited, and it is difficult to meet the rapid response requirements for sudden pollution incidents. In addition, the construction cost of fixed monitoring stations is high and their flexibility is poor, and they cannot meet the monitoring requirements of large ranges and high spatio-temporal resolutions.

[0003] To make up for the deficiencies of fixed monitoring stations, mobile monitoring technology has emerged. Mobile monitoring can cover a large area in a short time by carrying high-precision monitoring equipment on a mobile platform (such as vehicles, drones, ships, etc.) and obtain real-time pollutant concentration distribution data. However, it is difficult to accurately locate pollution sources in complex environments when used alone, and it needs to be combined with an effective traceability algorithm. Existing traceability algorithms such as the Lagrangian particle diffusion model rely on a large amount of meteorological data and are computationally complex; they require sufficient sample data and are limited in application in the field of pollution source location.

[0004] The accuracy of the traceability algorithm highly depends on model assumptions and parameter selection. How to combine real-time monitoring data to achieve model construction and parameter optimization, and achieve rapid and accurate location of pollution sources, providing strong technical support for the emergency response to sudden pollution incidents and environmental pollution control is the technical challenge faced by the current rapid pollution source location using mobile monitoring data. Summary of the Invention

[0005] The purpose of the present invention is to provide a rapid pollution source location system and method based on mobile monitoring data and a traceability algorithm to solve the above problems existing in the prior art.

[0006] Specifically, this application is as follows:

[0007] Provide a rapid pollution source location system based on mobile monitoring data and a traceability algorithm, characterized in that it includes a water quality mobile monitoring data collection module, a data preprocessing module, a traceability algorithm model construction module, a traceability algorithm model optimization module, a pollution source location module, and a result display module;

[0008] The water quality mobile monitoring data collection module: activates the mobile monitoring device, collects the monitoring information of water quality environment parameters in real time through the dynamic position adjustment of the mobile monitoring device, obtains the mobile monitoring data, and transmits the mobile monitoring data to the data preprocessing module;

[0009] The data preprocessing module: preprocesses the collected mobile monitoring data to obtain the monitoring feature data of the mobile monitoring data, and transmits the monitoring feature data to the source tracing algorithm model construction module;

[0010] The source tracing algorithm model construction module: describes the states of different monitoring feature data for the preprocessed mobile monitoring data, designs a reward function in combination with the source tracing target, controls the subsequent navigation direction and monitoring frequency of the mobile monitoring device, and constructs a source tracing algorithm model for dynamic pollution source tracing;

[0011] The source tracing algorithm model optimization module: performs model training in combination with the historical monitoring data sequence, cuts the amplitude of the policy update, performs backpropagation according to the losses of the learning strategy and the value function, and optimizes the parameters in combination with the mobile monitoring policy network layer and the mobile monitoring value network layer;

[0012] The pollution source positioning module: processes the preprocessed water quality mobile monitoring data by using the source tracing algorithm model, determines the target navigation trajectory according to the navigation coordinates of the mobile monitoring device, controls the mobile monitoring device to cruise according to the target navigation trajectory, and determines the pollution source location;

[0013] The result display module: displays the data of the pollution source positioning module, marks the position of the mobile monitoring device on the map, visually displays the moving path of the mobile monitoring device, marks the pollution source position information, judges whether the source tracing task is successful, and calculates and outputs the success rate.

[0014] The description of the states of different monitoring feature data for the preprocessed mobile monitoring data includes:

[0015] Concatenates the position coordinates of the mobile monitoring device and the pollutant concentration into a state data set of the mobile monitoring device containing the position of the mobile monitoring device and the pollutant concentration, and the data set is (Δx, Δy, C p ); where, (Δx, Δy) represents the position coordinates of the mobile monitoring device, and C p represents the pollutant concentration at the current position and is calculated through the water quality diffusion equation.

[0016] The design of the reward function in combination with the source tracing target and the control of the subsequent navigation direction and monitoring frequency of the mobile monitoring device includes:

[0017] Positive rewards or negative punishments are given according to the reward function evaluated based on the distance of the mobile monitoring device from the pollution source, energy consumption cost, time cost, and information value, to encourage learning the correct source tracing strategy, determine the navigation direction of the mobile monitoring device, and limit the moving range and monitoring frequency of the mobile monitoring device;

[0018] The reward function consists of two parts: positive reward and negative punishment, and is expressed as:

[0019] R = ω1*(-d) + ω2*(-(P*T)) + ω3*(-T) + ω4*(C p *A)

[0020] where R is the reward function; ω1, ω2, ω3, and ω4 represent weights used to balance the importance of each reward term; d represents the Euclidean distance between the mobile monitoring device and the pollution source; P represents the power of the mobile monitoring device; T represents the moving time of the mobile monitoring device; C p represents the pollutant concentration, and A represents the accuracy coefficient of the mobile monitoring data.

[0021] The design of the reward function in combination with the source tracing objective to control the subsequent navigation direction and monitoring frequency of the mobile monitoring device further includes:

[0022] Set the termination conditions for the mobile monitoring device. The first termination condition is the distance condition, and set the first threshold condition for the Euclidean distance between the mobile monitoring device and the pollution source;

[0023] The second termination condition is the time condition, and set the second threshold condition for the moving time of the mobile monitoring device;

[0024] When the first threshold condition is reached or within the second threshold condition, return a data set {St, Re, Do, In} and store it. Among them, St represents the state data set of the mobile monitoring device, St = (Δx, Δy, C p ); Re represents the reward obtained after the mobile monitoring device executes an action; Do represents whether the termination condition is reached; In represents additional information.

[0025] The construction of the source tracing algorithm model for dynamic pollution source tracing further includes:

[0026] Set a feature extraction network layer to extract the feature information of the state description. The feature extraction network layer consists of two fully connected layers and a ReLU activation function. The policy network layer and the value network layer share a feature extraction network layer;

[0027] The mobile monitoring policy network layer extracts the feature information to obtain the mean of the navigation actions of the mobile monitoring device, adds Gaussian noise to the basis of the mean of the navigation actions, and outputs the probability distribution of the navigation actions to guide the selection of the navigation actions executed by the mobile monitoring device;

[0028] The cruise monitoring value network layer extracts the feature information, outputs the value estimation of the cruise monitoring device status according to the value function, conducts the value evaluation of the current status, and helps the cruise monitoring policy network layer update the policy.

[0029] The cruise monitoring policy network layer extracts the feature information to obtain the mean value of the cruise monitoring device's navigation actions, adds Gaussian noise based on the mean value of the navigation actions, and outputs the probability distribution of the navigation actions. The guidance for the cruise monitoring device to perform the selection of navigation actions includes:

[0030] The mean value of the cruise monitoring device's navigation actions is:

[0031] μ = ReLU(W * St + b)

[0032] where μ represents the mean value, W represents the weight vector, St represents the set of status data of the cruise monitoring device, b represents the bias vector, and ReLU() represents the activation function;

[0033] Adding Gaussian noise based on the mean value of the navigation actions is:

[0034] a = μ + σ * ε

[0035] where a represents the subsequent navigation action of the cruise monitoring device, σ represents the standard deviation, controlling the intensity of the noise, and ε represents the noise sampled from the standard normal distribution:

[0036] The probability distribution of the navigation actions is:

[0037] The cruise monitoring value network layer extracts the feature information, outputs the value estimation of the cruise monitoring device status according to the value function, conducts the value evaluation of the current status, and helps the cruise monitoring policy network layer update the policy, including:

[0038] The value function is expressed as:

[0039]

[0040] where V(St) represents the state value, which is the expected value of the future cumulative reward for the cruise monitoring device starting from the state St, R k is the immediate reward at time step k, γ represents the discount factor, 0 ≤ γ < 1, which is used to balance the importance of the current reward and the future reward, and St0 represents the initial state of the cruise monitoring device;

[0041] For each state St of the cruise monitoring device, calculate the target state value:

[0042] V t (St) = R + γ(1 - Do)V(St')

[0043] Among them, V t (St) represents the target state value of the mobile monitoring device, R represents the reward of the current state, V(St') represents the future state value, Do represents whether the termination condition is reached. When Do reaches the termination condition, the future reward is ignored, that is, V(St') = 0;

[0044] Calculate the value loss: Ls = (V(St) - V t (St)) 2 .

[0045] The update amplitude of the pruning strategy is backpropagated according to the losses of the learning strategy and the value function, and the parameter optimization is carried out by combining the mobile monitoring strategy network layer and the mobile monitoring value network layer, including:

[0046] The update amplitude of the pruning strategy is realized through the pruning objective function, and the pruning objective function is expressed as:

[0047]

[0048] Among them, r t (θ) represents the proportion of policy update, At is the advantage function, which represents the pros and cons of the current mobile monitoring device action relative to the average performance, and ∈ is the pruning range, which is used to limit the update amplitude of the policy, represents the expected value at time step k.

[0049] The pollution source location module further includes:

[0050] Through the preprocessed water quality mobile monitoring data, obtain the navigation position coordinates and pollutant concentration of the mobile monitoring device, determine the parameter indicators of the tracing algorithm model. The parameter indicators include the location information of the pollution source and the characteristic information that can reflect the change trend of water quality over time. Determine the target navigation trajectory according to the navigation coordinates of the mobile monitoring device, and control the mobile monitoring device to cruise according to the target navigation trajectory to determine the final pollution source location.

[0051] This application also provides a pollution source rapid location method based on mobile monitoring data and tracing algorithm, which is applied to the above-mentioned pollution source rapid location system based on mobile monitoring data and tracing algorithm, and specifically includes the following steps:

[0052] Step 1: Start the mobile monitoring device, and collect the monitoring information of water quality environment parameters in real time through the dynamic position adjustment of the mobile monitoring device to obtain mobile monitoring data;

[0053] Step 2: Preprocess the collected mobile monitoring data to obtain the monitoring characteristic data of the mobile monitoring data, and transmit the monitoring characteristic data to the tracing algorithm model construction module;

[0054] Step 3: Describe the status of different monitoring feature data for the preprocessed cruise monitoring data, design a reward function in combination with the traceability target, control the subsequent navigation direction and monitoring frequency of the cruise monitoring device, and construct a traceability algorithm model for dynamic pollution source traceability;

[0055] Step 4: Combine the historical monitoring data sequence for model training, trim the amplitude of the policy update, perform backpropagation according to the losses of the learning policy and value function, and optimize the parameters in combination with the cruise monitoring policy network layer and the cruise monitoring value network layer;

[0056] Step 5: Use the traceability algorithm model to process the preprocessed water quality cruise monitoring data, determine the target navigation trajectory according to the navigation coordinates of the cruise monitoring device, control the cruise monitoring device to cruise according to the target navigation trajectory, and determine the pollution source location;

[0057] Step 6: Mark the position of the cruise monitoring device on the map, visually display the moving path of the cruise monitoring device, mark the pollution source location information, judge whether the traceability task is successful, and calculate and output the success rate.

[0058] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects:

[0059] 1. When using a cruise monitoring device for water quality traceability, a dynamic tracking mechanism is adopted, and the monitoring device is enabled to move autonomously through constructing a traceability algorithm, adapt to the position change of the pollution source, and quickly locate the pollution source;

[0060] 2. Considering the energy consumption-accuracy trade-off of the cruise monitoring device during navigation, an energy consumption penalty is introduced in the set reward function to avoid ineffective movement and improve the traceability efficiency of the monitoring device;

[0061] 3. Combining the historical monitoring data sequence to reflect the changing trend characteristics of water quality over time, improving the accuracy of model training, being able to accurately and quickly carry out water quality traceability work, and realizing the efficiency and accuracy of water quality traceability; BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 is a schematic structural diagram of a pollution source rapid positioning system based on cruise monitoring data and traceability algorithm provided by an embodiment of the present invention;

[0063] Figure 2 is a schematic flowchart of a pollution source rapid positioning method based on cruise monitoring data and traceability algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0064] The present invention will be described in detail below with reference to the accompanying drawings.

[0065] Example 1

[0066] The present invention provides a rapid pollution source positioning system based on cruise monitoring data and a traceability algorithm, as Figure 1 shown, which includes a water quality cruise monitoring data collection module, a data preprocessing module, a traceability algorithm model construction module, a traceability algorithm model optimization module, a pollution source positioning module, and a result display module;

[0067] The water quality cruise monitoring data collection module: starts the cruise monitoring equipment, and through the dynamic position adjustment of the cruise monitoring equipment, it collects the monitoring information of water quality environment parameters in real time, obtains the cruise monitoring data, and transmits the cruise monitoring data to the data preprocessing module;

[0068] Specifically, the cruise monitoring equipment is mobile monitoring equipment such as unmanned boats and drones containing automatic collection devices, and specific limitations are not made here; the cruise monitoring equipment is deployed in the area to be monitored to ensure that the equipment can collect cruise monitoring data in real time; the cruise monitoring data includes: chemical oxygen demand (COD), ammonia nitrogen concentration, acidity and alkalinity, position coordinates of the mobile monitoring equipment, water flow velocity, wind direction, and temperature.

[0069] The data preprocessing module: preprocesses the collected cruise monitoring data to obtain the monitoring characteristic data of the cruise monitoring data, and transmits the monitoring characteristic data to the traceability algorithm model construction module;

[0070] Specifically, the preprocessing includes cleaning, denoising, and standardization of the cruise monitoring data; data cleaning can remove incomplete, inaccurate, repeated, or irrelevant data; data denoising can remove noise interference in the data through technical means; data standardization can uniformly process data from different sources and in different formats.

[0071] The traceability algorithm model construction module: describes the states of different monitoring characteristic data of the preprocessed cruise monitoring data, designs a reward function in combination with the traceability target, controls the subsequent sailing direction and monitoring frequency of the cruise monitoring equipment, and constructs a traceability algorithm model for dynamic pollution source traceability;

[0072] Specifically, the different monitoring characteristic data of the preprocessed cruise monitoring data include: chemical oxygen demand (COD), ammonia nitrogen concentration, and acidity and alkalinity reflecting the water quality characteristics collected; position coordinates of the cruise monitoring equipment reflecting the collection position characteristics; water flow velocity, wind direction, and temperature reflecting the collection environment characteristics; at the same time, adding a historical monitoring data sequence to reflect the change trend characteristics of water quality over time;

[0073] Extract a set of characteristic data based on the monitored characteristic data. Set the position coordinates of the mobile monitoring device as (Δx, Δy). Combine the water quality characteristic data and the environmental characteristic data to calculate the diffusion coefficient D of the pollutant and the flow velocity vector U, where the flow velocity components are U x ,U y , representing the flow velocity of the fluid in the x and y directions; Since different pollutants contribute differently to the pollutant concentration, obtain the relative molar mass M0 of the water quality based on different water quality characteristic data and environmental characteristic data. According to the water quality characteristic data of each monitoring point in the monitoring data and the water flow velocity, wind direction, and temperature in the environmental characteristic data, establish a diffusion model and use the water quality diffusion equation to calculate the pollutant concentration;

[0074] The state description of the preprocessed mobile monitoring data for different monitoring characteristic data includes:

[0075] Concatenate the position coordinates of the mobile monitoring device and the pollutant concentration into a state data set of the mobile monitoring device containing the position of the mobile monitoring device and the pollutant concentration. The data set is (Δx, Δy, C p ); where, (Δx, Δy) represents the position coordinates of the mobile monitoring device, and C p represents the pollutant concentration at the current position, which is calculated by the water quality diffusion equation.

[0076]

[0077] Among them, (x0, y0) is the position of the pollution source, U x ,U y are the flow velocity components, M0 represents the total mass of the pollutant released by the pollution source, that is, the total amount of the initially released pollutant, D represents the diffusion coefficient of the pollutant, and t represents time.

[0078] Design a reward function in combination with the tracing target to control the subsequent navigation direction and monitoring frequency of the mobile monitoring device, including: performing positive rewards or negative punishments according to the distance of the mobile monitoring device from the pollution source, energy consumption cost, time cost, and reward function evaluated by information value, inspiring to learn the correct tracing strategy, determining the navigation direction of the mobile monitoring device, and controlling the movement range and monitoring frequency of the mobile monitoring device;

[0079] Calculate the reward according to the real-time position information of the mobile monitoring device, and calculate the Euclidean distance between the mobile monitoring device and the pollution source:

[0080]

[0081] Among them, the position coordinates of the mobile monitoring device are (Δx, Δy), the position coordinates of the pollution source are (x0, y0), and d represents the straight-line distance between the two points;

[0082] The source tracing objective is to comprehensively consider the distance, energy consumption cost, time cost, and information value of pollution sources, guide the device to complete the monitoring task efficiently and at low cost, while maximizing the information value. The reward function consists of two parts: positive reward and negative penalty. Among them, the weights of the reward function are determined by simulation experiments or optimization algorithms to ensure the effectiveness of the reward function, which is expressed as:

[0083] R = ω1*(-d) + ω2*(-(P*T)) + ω3*(-T) + ω4*(C p *A)

[0084] Among them, R is the reward function; ω1, ω2, ω3, and ω4 represent weights used to balance the importance of each reward term; d represents the Euclidean distance between the mobile monitoring device and the pollution source; P represents the power of the mobile monitoring device; T represents the moving time of the mobile monitoring device; C p represents the pollutant concentration, and A represents the accuracy coefficient of the mobile monitoring data;

[0085] Specifically, the navigation coordinates of the mobile monitoring device are (Δx, Δy), and the range is limited to no more than 10 meters per movement, and an adjustable monitoring frequency is provided; if the action taken by the mobile monitoring device makes it closer to the pollution source, that is, the change trend of water quality indicators points to a specific area through sampling data, a positive reward can be given; conversely, if the action leads to moving away from the pollution source or making a wrong decision, such as choosing the wrong navigation direction and not obtaining more valuable information within a certain period of time, a negative reward is given; if the mobile monitoring device accurately identifies and locates the pollution source, a relatively large positive reward is given to encourage it to learn the correct source tracing strategy; the design of the reward encourages the monitoring device to approach the pollution source as soon as possible while avoiding unnecessary movements;

[0086] Determining the navigation direction of the mobile monitoring device through different monitoring feature data collected in real time by the mobile monitoring device, controlling the moving range and monitoring frequency of the mobile monitoring device also includes:

[0087] Setting the termination conditions of the mobile monitoring device. The first termination condition is the distance condition, setting the first threshold condition for the Euclidean distance between the mobile monitoring device and the pollution source;

[0088] The second termination condition is the time condition, setting the second threshold condition for the moving time of the mobile monitoring device;

[0089] When the first threshold condition is reached or the second threshold condition is met, return a data set {St, Re, Do, In} and store it. Among them, St represents the state data set of the mobile monitoring device, St = (Δx, Δy, C p); Re represents the reward obtained after the mobile monitoring device executes an action; Do represents whether the termination condition is reached; In represents additional information.

[0090] Obtaining the status and action monitoring data of the mobile monitoring device, describing the status of the preprocessed mobile monitoring data with different monitoring feature information, designing a reward function in combination with the traceability target, controlling the subsequent navigation direction and monitoring frequency of the mobile monitoring device, and constructing a traceability algorithm model for dynamic pollution source traceability further includes:

[0091] Setting a feature extraction network layer to extract the feature information of the status description. The feature extraction network layer consists of two fully connected layers and a ReLU activation function. The policy network layer and the value network layer share a feature extraction network layer;

[0092] The mobile monitoring policy network layer extracts the feature information to obtain the mean value of the navigation actions of the mobile monitoring device, adds Gaussian noise based on the mean value of the navigation actions, and outputs the probability distribution of the navigation actions to guide the selection of the navigation actions executed by the mobile monitoring device;

[0093] The mobile monitoring value network layer extracts the feature information, outputs the value estimation of the status of the mobile monitoring device according to the value function, conducts the value evaluation of the current status, and helps the mobile monitoring policy network layer update the policy;

[0094] Specifically, the feature extraction network layer extracts the feature information of the status description, consists of two fully connected layers and a ReLU activation function. The network structure: The first layer: a fully connected layer with a dimension of 64, the input dimension: the status dimension of the mobile monitoring device, the output dimension: 64; The ReLU activation function introduces non-linearity and alleviates the problem of gradient disappearance; The second layer: a fully connected layer with a dimension of 64, the input dimension: 64, the output dimension: 64; The ReLU activation function introduces non-linearity again; The third layer branches to the mobile monitoring policy network layer and the mobile monitoring value network layer; Among them, the input of the mobile monitoring policy network layer is: the output dimension of the feature extraction network layer: 64, and the probability distribution of the output action; The input of the mobile monitoring value network layer is: the output dimension of the feature extraction network layer: 64, and the output is a 1D scalar representing the value estimation of the status;

[0095] The mobile monitoring policy network layer extracts the feature information to obtain the mean value of the navigation actions of the mobile monitoring device, adds Gaussian noise based on the mean value of the navigation actions, and outputs the probability distribution of the navigation actions to guide the selection of the navigation actions executed by the mobile monitoring device includes:

[0096] The mean value of the navigation actions of the mobile monitoring device is:

[0097] μ = ReLU(W * St + b)

[0098] Among them, μ represents the mean, W represents the weight vector, St represents the set of status data of the moving monitoring device, b represents the bias vector, and ReLU() represents the activation function;

[0099] Adding Gaussian noise to the mean of the in-flight action is:

[0100] a = μ + σ * ε

[0101] Among them, a represents the subsequent in-flight action of the moving monitoring device, σ represents the standard deviation, which controls the intensity of the noise, and ε represents the noise sampled from the standard normal distribution:

[0102] The probability distribution of the in-flight action is:

[0103]

[0104] The moving monitoring value network layer extracts the feature information, outputs the value estimation of the moving monitoring device status according to the value function, and conducts the value evaluation of the current status to help the moving monitoring policy network layer update the policy, including:

[0105] The value function is expressed as:

[0106]

[0107] Among them, V(St) represents the state value, which is the expected value of the future cumulative reward starting from the state St of the moving monitoring device, and R k is the immediate reward at time step k, γ represents the discount factor, 0 ≤ γ < 1, which is used to balance the importance of the current reward and the future reward, and St0 represents the initial state of the moving monitoring device;

[0108] For each state St of the moving monitoring device, calculate the target state value:

[0109] V t (St) = R + γ(1 - Do)V(St')

[0110] Among them, V t (St) represents the target state value of the moving monitoring device, R represents the reward of the current state, V(St') represents the future state value, and Do represents whether the termination condition is reached. When Do reaches the termination condition, the future reward is ignored, that is, V(St') = 0;

[0111] Calculate the value loss:

[0112] Ls = (V(St) - V t (St)) 2

[0113] The cruise monitoring policy network layer and the cruise monitoring value network layer respectively output action probabilities and state values, which are used to solve the reinforcement learning problem of continuous or discrete action spaces. In a specific implementation, the dimension of the state space of the cruise monitoring device is set to 3, and the state is represented as St = (Δx, Δy, C p ), the dimension of the action space is 2, which is represented as the moving coordinate representation (Δx, Δy) of the cruise monitoring device, and hyperparameters including the learning rate, discount factor, and clipping range are set.

[0114] The traceability algorithm model optimization module: combines historical monitoring data sequences for model training, clips the amplitude of policy updates, performs backpropagation according to the losses of the learning policy and value function, and optimizes parameters by combining the cruise monitoring policy network layer and the cruise monitoring value network layer;

[0115] The amplitude of the clipped policy update is achieved through a clipping objective function, and the clipping objective function is expressed as:

[0116]

[0117] where r t (θ) represents the ratio of policy update, At is the advantage function, which represents the pros and cons of the current action of the cruise monitoring device relative to the average performance, ∈ is the clipping range, which is used to limit the amplitude of policy update, represents the expected value at time step k;

[0118] Specifically, the ratio of policy update r t (θ) is the ratio between the current policy and the old policy, and its role is to measure the degree of change in action selection of the current policy relative to the old policy in a certain state. If the value of r t (θ) is greater than 1, it means that the current policy is more inclined to select this action; if it is less than 1, it means that the probability of the current policy selecting this action is low. The change of the policy is evaluated by the ratio to ensure that each update will not be too large;

[0119] Calculate the gradient of the loss function with respect to the model parameters through the chain rule, and use the gradient descent method to update the parameters of the cruise monitoring policy network layer and the value network layer. The traceability algorithm is optimized by adjusting the hyperparameters of the learning rate, discount factor, and clipping range.

[0120] Example 2

[0121] As a specific example, during the model training process, the environment is initialized: the position of the pollution source is set to (50, 50). The flow velocity vector is set to (0.1, 0.0), indicating that the fluid flows at a speed of 0.1 in the x direction, and the cruise monitoring policy network layer and the value network layer are trained;

[0122] Number of training rounds: Set the total number of training rounds to 500. Initialization: At the start of each training round, reset the environment and obtain the initial state of the mobile monitoring device. Initialize an empty list to store trajectory data, and set up state lists, action lists, reward lists, next state lists, and termination flag lists.

[0123] Select an action for the mobile monitoring device. The action range is restricted to [-1, 1], which is scaled to [-10, 10] meters, and then execute the action of the mobile monitoring device, and obtain the next state, reward, and termination flag. Store trajectory data: Store the current state, action, reward, next state, and termination flag into the corresponding lists. Update the state: Assign the next state to the state and continue to interact with the environment.

[0124] Update the mobile monitoring policy network layer and value network layer using the trajectory data collected in the current round. The specific update logic includes calculating the advantage function, clipping the policy update amplitude, calculating the loss function, and performing backpropagation. Output the training progress: Output the training progress every 50 rounds. Adjust the number of training rounds according to the task complexity to ensure that the model converges sufficiently.

[0125] The pollution source location module: Use the traceability algorithm model to process the preprocessed water quality mobile monitoring data, determine the target navigation trajectory based on the navigation coordinates of the mobile monitoring device, and control the mobile monitoring device to cruise according to the target navigation trajectory to determine the pollution source location.

[0126] Through the preprocessed water quality mobile monitoring data, obtain the navigation position coordinates of the mobile monitoring device and the pollutant concentration, determine the parameter indicators of the traceability algorithm model. The parameter indicators include the location information of the pollution source and the characteristic information that can reflect the change trend of water quality over time. Determine the target navigation trajectory based on the navigation coordinates of the mobile monitoring device, and control the mobile monitoring device to cruise according to the target navigation trajectory to determine the final pollution source location information.

[0127] The result display module: Display the data of the pollution source location module, mark the position of the mobile monitoring device on the map, visually display the moving path of the mobile monitoring device, mark the pollution source location information, judge whether the traceability task is successful, and calculate and output the success rate.

[0128] Specifically, record the moving path of the mobile monitoring device, judge whether each traceability task is successful, that is, whether the mobile monitoring device is close to the pollution source, visually monitor the moving path of the device, and mark the pollution source location. Record the position of the mobile monitoring device and judge whether the traceability task is successful. If the distance between the monitoring device and the pollution source is less than 5, the task is considered successful. Calculate and output the success rate. The moving path of the mobile monitoring device is represented by a blue dot connection, and the pollution source location is marked by a red star.

[0129] Example 3

[0130] As Figure 2 shown, the present invention further provides a method for quickly locating pollution sources based on on-road monitoring data and a tracing algorithm, which is applied to the above-mentioned system for quickly locating pollution sources based on on-road monitoring data and a tracing algorithm, and specifically includes the following steps:

[0131] Step 1: Start the on-road monitoring device, and collect monitoring information of water quality environment parameters in real time through the dynamic position adjustment of the on-road monitoring device to obtain on-road monitoring data;

[0132] Step 2: Preprocess the collected on-road monitoring data to obtain monitoring characteristic data of the on-road monitoring data, and transmit the monitoring characteristic data to the tracing algorithm model construction module;

[0133] Step 3: Describe the states of different monitoring characteristic data for the preprocessed on-road monitoring data, design a reward function in combination with the tracing target, control the subsequent navigation direction and monitoring frequency of the on-road monitoring device, and construct a tracing algorithm model for dynamic pollution source tracing;

[0134] Step 4: Combine historical monitoring data sequences for model training, cut the amplitude of the strategy update, perform backpropagation according to the losses of the learning strategy and the value function, and optimize the parameters in combination with the on-road monitoring strategy network layer and the on-road monitoring value network layer;

[0135] Step 5: Use the tracing algorithm model to process the preprocessed water quality on-road monitoring data, determine the target navigation trajectory according to the navigation coordinates of the on-road monitoring device, control the on-road monitoring device to cruise according to the target navigation trajectory, and determine the pollution source location;

[0136] Step 6: Mark the position of the on-road monitoring device on the map, visually display the moving path of the on-road monitoring device, mark the pollution source location information, judge whether the tracing task is successful, and calculate and output the success rate.

[0137] In the specification provided here, a large number of specific details are described. However, it can be understood that the embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures, and technologies are not shown in detail so as not to obscure the understanding of this specification.

[0138] Similarly, it should be understood that, for the purpose of streamlining the present disclosure and aiding in the understanding of one or more of the various inventive aspects, in the foregoing description of the exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, the inventive aspects lie in less than all of the features of the single foregoing disclosed embodiment. Thus, the claims following the detailed description hereby expressly incorporate the detailed description, wherein each claim itself serves as a separate embodiment of the present invention.

[0139] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments and not others, combinations of features of different embodiments are meant to be within the scope of the present invention and form different embodiments. For example, in the following claims, any one of the claimed embodiments can be used in any combination.

Claims

1. A rapid pollution source location system based on mobile monitoring data and traceability algorithm, characterized in that, It includes a water quality mobile monitoring data collection module, a data preprocessing module, a source tracing algorithm model construction module, a source tracing algorithm model optimization module, a pollution source location module, and a result display module; The water quality mobile monitoring data collection module: starts the mobile monitoring device, collects the monitoring information of water quality environment parameters in real time through the dynamic position adjustment of the mobile monitoring device, obtains the mobile monitoring data, and transmits the mobile monitoring data to the data preprocessing module; The data preprocessing module: preprocesses the collected mobile monitoring data to obtain the monitoring feature data of the mobile monitoring data, and transmits the monitoring feature data to the source tracing algorithm model construction module; The source tracing algorithm model construction module: describes the states of different monitoring feature data for the preprocessed mobile monitoring data, designs a reward function in combination with the source tracing target, controls the subsequent navigation direction and monitoring frequency of the mobile monitoring device, and constructs a source tracing algorithm model for dynamic pollution source tracing; The source tracing algorithm model optimization module: conducts model training in combination with the historical monitoring data sequence, cuts the amplitude of the policy update, performs backpropagation according to the losses of the learning policy and the value function, and optimizes the parameters in combination with the mobile monitoring policy network layer and the mobile monitoring value network layer; The pollution source location module: processes the preprocessed water quality mobile monitoring data by using the source tracing algorithm model, determines the target navigation trajectory according to the navigation coordinates of the mobile monitoring device, controls the mobile monitoring device to cruise according to the target navigation trajectory, and determines the pollution source location; The result display module: displays the data of the pollution source location module, marks the position of the mobile monitoring device on the map, visually displays the moving path of the mobile monitoring device, marks the pollution source location information, judges whether the source tracing task is successful, and calculates and outputs the success rate.

2. The rapid pollution source positioning system based on mobile monitoring data and traceability algorithm according to claim 1, characterized in that, The description of the states of different monitoring feature data for the preprocessed mobile monitoring data includes: Concatenate the position coordinates of the mobile monitoring device and the pollutant concentration into a set of status data of the mobile monitoring device that includes the position of the mobile monitoring device and the pollutant concentration. The data set is (Δx, Δy, C p ); where (Δx, Δy) represents the position coordinates of the mobile monitoring device, and C p represents the pollutant concentration at the current position and is calculated through the water quality diffusion equation.

3. A rapid pollution source location system based on mobile monitoring data and traceability algorithm according to claim 1, characterized in that, The design of the reward function in combination with the source tracing target and the control of the subsequent navigation direction and monitoring frequency of the mobile monitoring device include: Conducts positive reward or negative punishment according to the reward function evaluated by the distance of the mobile monitoring device from the pollution source, energy consumption cost, time cost, and information value, stimulates learning the correct source tracing strategy, determines the navigation direction of the mobile monitoring device, and limits the moving range and monitoring frequency of the mobile monitoring device; The reward function consists of two parts, positive reward and negative punishment, and is expressed as: R = ω1*(-d) + ω2*(-(P*T)) + ω3*(-T) + ω4*(C p *A) Among them, R is the reward function; ω1, ω2, ω3, ω4 represent weights used to balance the importance of each reward item; d represents the Euclidean distance between the mobile monitoring device and the pollution source; P represents the power of the mobile monitoring device; T represents the moving time of the mobile monitoring device; C p represents the pollutant concentration, and A represents the accuracy coefficient of the mobile monitoring data.

4. The rapid pollution source positioning system based on mobile monitoring data and traceability algorithm according to claim 3, characterized in that, The design of the reward function in combination with the source tracing target and the control of the subsequent navigation direction and monitoring frequency of the mobile monitoring device further includes: Sets the termination conditions of the mobile monitoring device. The first termination condition is the distance condition, and sets the first threshold condition for the Euclidean distance between the mobile monitoring device and the pollution source; The second termination condition is the time condition, and sets the second threshold condition for the moving time of the mobile monitoring device; When the first threshold condition is reached or under the second threshold condition, return a data set {St, Re, Do, In} and store it, where St represents the status data set of the mobile monitoring device, St = (Δx, Δy, C p ); Re represents the reward obtained after the mobile monitoring device performs an action; Do represents whether the termination condition is reached; In represents additional information.

5. The rapid pollution source positioning system based on mobile monitoring data and traceability algorithm according to claim 1, characterized in that, The construction of the source tracing algorithm model for dynamic pollution source tracing further includes: Sets a feature extraction network layer to extract the feature information of the state description. The feature extraction network layer consists of two fully connected layers and a ReLU activation function, and the policy network layer and the value network layer share a feature extraction network layer; The navigation monitoring strategy network layer extracts the feature information to obtain the mean of the navigation actions of the navigation monitoring device, adds Gaussian noise based on the mean of the navigation actions, outputs the probability distribution of the navigation actions, and guides the selection of the navigation actions of the navigation monitoring device; The navigation monitoring value network layer extracts the feature information, outputs the value estimation of the state of the navigation monitoring device according to the value function, conducts the value evaluation of the current state, and helps the navigation monitoring strategy network layer update the strategy.

6. The rapid pollution source positioning system based on mobile monitoring data and traceability algorithm according to claim 5, characterized in that The navigation monitoring strategy network layer extracts the feature information to obtain the mean of the navigation actions of the navigation monitoring device, adds Gaussian noise based on the mean of the navigation actions, and the output of the probability distribution of the navigation actions to guide the selection of the navigation actions of the navigation monitoring device includes: The mean of the navigation actions of the navigation monitoring device is: μ = ReLU(W * St + b) where μ represents the mean, W represents the weight vector, St represents the set of state data of the navigation monitoring device, b represents the bias vector, and ReLU() represents the activation function; Adding Gaussian noise based on the mean of the navigation actions is: a = μ + σ * ε where a represents the subsequent navigation action of the navigation monitoring device, σ represents the standard deviation, controlling the intensity of the noise, and ε represents the noise sampled from the standard normal distribution: The probability distribution of the sailing actions is as follows:

7. A rapid pollution source location system based on mobile monitoring data and traceability algorithm according to claim 5, characterized in that The navigation monitoring value network layer extracts the feature information, outputs the value estimation of the state of the navigation monitoring device according to the value function, conducts the value evaluation of the current state, and helps the navigation monitoring strategy network layer update the strategy includes: The value function is expressed as: Among them, V(St) represents the state value, which is the expected value of the future cumulative reward starting from the state St of the underway monitoring device, and R k is the immediate reward at time step k. γ represents the discount factor, where 0 ≤ γ < 1, which is used to balance the importance of the current reward and the future reward. St0 represents the initial state of the underway monitoring device; For each state St of the navigation monitoring device, calculate the target state value: V t (St) = R + γ(1 - Do)V(St') Among them, V t (St) represents the target state value of the underway monitoring device, R represents the reward of the current state, V(St') represents the future state value, Do represents whether the termination condition is reached. When Do reaches the termination condition, the future reward is ignored, that is, V(St') = 0; Calculate the value loss: Ls = (V(St) - V t (St)) 2 .

8. The rapid pollution source positioning system based on moving monitoring data and traceability algorithm according to claim 1, characterized in that, The amplitude of the clipped policy update is backpropagated according to the losses of the learning policy and the value function calculation, and the parameter optimization is carried out by combining the navigation monitoring strategy network layer and the navigation monitoring value network layer includes: The amplitude of the clipped policy update is achieved through the clipped objective function, and the clipped objective function is expressed as: where r t (θ) represents the proportion of policy update. At is the advantage function, which represents the superiority or inferiority of the current action of the navigation monitoring device relative to the average performance. ∈ is the clipping range, which is used to limit the amplitude of policy update. represents the expected value at time step k.

9. The rapid pollution source positioning system based on mobile monitoring data and traceability algorithm according to claim 1, characterized in that The pollution source location module further includes: Through the preprocessed water quality navigation monitoring data, obtain the navigation position coordinates and pollutant concentration of the navigation monitoring device, determine the parameter indicators of the traceability algorithm model, the parameter indicators include the location information of the pollution source and the feature information that can reflect the change trend of water quality over time, determine the target navigation trajectory according to the navigation coordinates of the navigation monitoring device, and control the navigation monitoring device to cruise according to the target navigation trajectory to determine the final pollution source location.

10. A rapid pollution source location method based on vehicle-mounted monitoring data and traceability algorithm, characterized in that, Applied to a pollution source rapid location system based on navigation monitoring data and traceability algorithm according to any one of claims 1 to 9, specifically including the following steps: Step 1: Start the navigation monitoring device, and collect the monitoring information of the water quality environment parameters in real time through the dynamic position adjustment of the navigation monitoring device to obtain the navigation monitoring data; Step 2: Preprocess the collected navigation monitoring data to obtain the monitoring feature data of the navigation monitoring data, and transmit the monitoring feature data to the traceability algorithm model construction module; Step 3: Describe the states of the different monitoring feature data for the preprocessed cruise monitoring data, design a reward function in combination with the tracing target, control the subsequent sailing direction and monitoring frequency of the cruise monitoring equipment, and construct a tracing algorithm model for dynamic pollution source tracing; Step 4: Combine with the historical monitoring data sequence for model training, trim the amplitude of the policy update, perform backpropagation according to the losses of the learning policy and the value function, and optimize the parameters in combination with the cruise monitoring policy network layer and the cruise monitoring value network layer; Step 5: Use the tracing algorithm model to process the preprocessed water quality cruise monitoring data, determine the target sailing trajectory according to the sailing coordinates of the cruise monitoring equipment, control the cruise monitoring equipment to cruise according to the target sailing trajectory, and determine the pollution source location; Step 6: Mark the position of the cruise monitoring equipment on the map, visually display the moving path of the cruise monitoring equipment, mark the pollution source location information, judge whether the tracing task is successful, and calculate and output the success rate.

Images