A smart city data processing method and system based on Beidou positioning

Through the smart city data processing system based on Beidou positioning, pollution trends in the next seven days are predicted and a comprehensive governance plan is generated, which solves the problem of high-cost governance and improves the effectiveness and efficiency of pollution control.

CN118780445BActive Publication Date: 2024-11-05LIAONING BEACON TECH CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411262478.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-10
Publication Date
2024-11-05
Estimated Expiration
2044-09-10

AI Technical Summary

Technical Problem

In smart city management, when the pollutant concentration exceeds the warning threshold, large-scale and high-cost governance measures are usually required, resulting in high governance costs, waste of resources, and expanding economic losses.

Method used

Adopt a smart city data processing system based on Beidou positioning, obtain pollution source coordinates through monitoring point settings modules, collect pollution data from the environmental data collection module, and predict pollution trends in the next seven days, and the comprehensive governance plan generation module uses deep Q network to generate comprehensive Governance plan, balance governance effects, resource consumption and costs.

Benefits of technology

By predicting the trend of pollutant concentration changes in the next seven days, a comprehensive governance plan can be generated, which can more effectively balance the governance goals, improve the governance effect, and reduce costs compared with a single governance plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118780445B_ABST
    Figure CN118780445B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of smart city technology, and discloses a smart city data processing method and system based on Beidou positioning. The smart city data processing method based on Beidou positioning includes a monitoring point setting module, which is used to determine the specific location of environmental pollution monitoring; an environmental data collection module, which is used to collect urban environmental pollution data; a pollution prediction module, which is used to predict the pollutant concentration within seven days and evaluate whether a pollution warning is issued within seven days; a comprehensive governance plan generation module, which is used to provide a comprehensive governance plan integrating pollution source control, pollutant treatment, and environmental restoration; the present invention predicts the changing trend of pollutant concentration within the next seven days, fully considers the governance effect, resource consumption and required expenditure of each governance plan, and makes a comprehensive governance plan including pollution source control, pollutant treatment, and environmental restoration. Compared with a single governance plan, the system can balance multiple governance goals and improve the governance effect.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of smart city technology, and more specifically, to a smart city data processing method and system based on Beidou positioning. Background Art

[0002] In smart city management, environmental pollution control is a key link in ensuring the health of residents and the sustainable development of the city. Environmental parameters such as air quality, water pollution, and soil pollution in areas with major pollution sources in the city are monitored. When the pollutant concentration in a certain area exceeds the warning threshold, corresponding control measures are taken, including pollution source control, pollutant treatment, and environmental restoration. However, when the pollutant concentration exceeds the warning threshold, it usually means that the pollution is already serious and large-scale, high-cost control measures need to be taken, which will lead to high control costs, waste of resources, and increased economic losses. Summary of the invention

[0003] The purpose of the present invention is to provide a smart city data processing method, system and storage medium based on Beidou positioning in order to solve the above problems.

[0004] The present invention provides a smart city data processing system based on Beidou positioning, comprising:

[0005] The monitoring point setting module uses the Beidou positioning system to obtain the precise coordinates of the main pollution sources, calculates the concentration distribution of pollutants at different distances through the diffusion model, and thus determines the potential key monitoring areas, and then sets monitoring sites in the key monitoring areas;

[0006] The environmental data collection module is used to collect urban environmental pollution data. The collection of urban environmental pollution data includes:

[0007] The environmental pollution parameters collected at all monitoring locations once a day include: the average air pollutant concentration, the average water pollutant concentration, the average soil pollutant concentration, and the average relevant meteorological parameters of the day;

[0008] The environmental pollution parameters collected at each monitoring location are encoded to obtain an environmental pollution parameter vector, and the environmental pollution parameter vectors of all monitoring locations on the same day are concatenated to obtain an environmental pollution vector;

[0009] The pollution prediction module inputs the data vector set into the prediction neural network, and the prediction neural network outputs the result indicating whether a pollution warning will occur in different monitoring locations within the next seven days. If the result is yes, it determines whether the time interval between the current date and the date of the previous implementation of the control measures is greater than the set time threshold. If the judgment result is yes, the comprehensive control plan generation module is started;

[0010] The data vector set includes environmental pollution vectors from day t-29 to day t; day t is defined as the current day;

[0011] A comprehensive governance solution generation module is used to input a data vector set into a trained decision neural network, and the decision neural network outputs an action vector;

[0012] Decode the action vector and output a comprehensive treatment plan to implement it and treat the pollution at the monitoring site.

[0013] Furthermore, if the number of days between the tth day and the monitoring start date is less than 30, an interpolation operation is performed on the missing data vector, and the interpolated data vector is a 0 vector.

[0014] Furthermore, an action vector is represented as:

[0015] ;

[0016] in, They represent the application degree of the first and h pollution source control measures at monitoring sites 1 and N, respectively;

[0017] It indicates the application degree of the first and nth pollutant treatment measures at monitoring sites 1 and N, respectively;

[0018] Represents the application degree of the first and mth environmental restoration measures at monitoring sites 1 and N, respectively.

[0019] Furthermore, the decision neural network uses a reinforcement learning model based on a deep Q network. The deep Q network includes an input layer, a hidden layer, and an output layer. The hidden layer uses a neural network to fit the Q value function, where the neural network includes a policy network and a target network.

[0020] Furthermore, the Q-value function Q(s, a) represents the reward that can be obtained after selecting action a in state s. The reward value is calculated by the reward function, where state s is the data vector at the current moment , action a is the action vector.

[0021] Furthermore, a reward function is set. The reward function needs to comprehensively consider the pollution control effect, resource consumption and cost. The control effect is the reduction ratio of pollutant concentration, resource consumption is the amount of resources after the control measures are used, and the cost is the total cost of implementing the control measures. It includes the immediate reward function and the final reward function. The immediate reward function is expressed as:

[0022] ;

[0023] in, represents the immediate reward at time t, represents the consumption of the selected resource at time t, represents the cost and cost of the governance measures selected at time t, α and β are weight parameters, and their default values ​​are 0.001 and 0.005 respectively;

[0024] The final reward needs to be calculated after the prevention and control measures have been implemented for a period of time and the prevention and control effects are evaluated, which can be expressed as:

[0025] ;

[0026] Among them, δ, ε, ζ are weight parameters, and their default values ​​are 100, 0.001, and 0.002, respectively. T represents the period from the implementation of governance measures to the evaluation of governance effects. is the final reward for period T, They represent the proportion of pollutant concentration reduction, resource consumption and total cost in period T respectively.

[0027] Furthermore, the specific steps of updating the policy network and the target network include:

[0028] Step 101, define the target Q value y of the Q value, y is expressed as:

[0029] ;

[0030] Among them, r is the reward value obtained after executing action a in the current state s, is the discount factor, the default value is 0.9, is the Q value of the target network, is the next state to be transferred after executing action a. For the status Next, select the action;

[0031] Step 102, use the policy network to calculate the current Q value:

[0032] ;

[0033] in, is the Q value of the policy network, is the current state, is the current action, represents the policy network;

[0034] Step 103, using the mean square error as the loss function, calculate the error between the current Q value and the target Q value:

[0035] ;

[0036] in, is the number of mini-batch samples, the default value is 64, is the target Q value of the i-th sample, is the current Q value of the i-th sample;

[0037] Step 104, minimizing the loss function through the back propagation algorithm and the Adam optimizer, and updating the parameters of the policy network;

[0038] Step 105, define the target network update interval H, the default value of H is 100, and the target network parameters are updated every H training steps; updating the target network parameters is to copy the parameters of the current policy network to the target network.

[0039] Furthermore, the ε-greedy strategy is used to select actions, including:

[0040] Step 201, ε represents the probability of selecting a random action, and its initial value is 0.9. 1-ε represents the probability of selecting the optimal action. During the training process, ε will decay linearly, specifically:

[0041] ;

[0042] in: is 0.9, is 0.1, T is the total number of training steps, the default value is 1000, tg step time The value of

[0043] Step 202, generating a random number r between 0 and 1;

[0044] Step 203, if , then choose a random action if Then select the action with the highest current Q value.

[0045] The present invention provides a smart city data processing method based on Beidou positioning, and the following steps are performed based on the aforementioned smart city data processing system based on Beidou positioning:

[0046] Step 301, starting from the first day of monitoring, collecting environmental pollution data once a day at different monitoring locations;

[0047] Step 302, encoding daily environmental pollution data into an overall environmental pollution vector;

[0048] Step 303, combine the environmental pollution vectors of a total of 30 days into a data vector set, input the data vector set into the prediction neural network, and output the prediction neural network to indicate whether pollution warning will occur in different monitoring locations in the next seven days. If the result is yes, determine whether the time interval between the current date and the date of the last implementation of the control measures is greater than the set time threshold. If yes, execute step 304; otherwise, execute step 301;

[0049] Step 304, inputting the data vector set into the trained decision neural network, and the decision neural network outputs the action vector;

[0050] Step 305 , decode the action vector, output a comprehensive governance plan for implementation, conduct pollution control in the urban area, and return to step 301 .

[0051] The present invention provides an application device that runs the aforementioned Beidou positioning-based smart city data processing system through a processor.

[0052] The beneficial effects of the present invention are as follows: the present invention predicts the changing trend of pollutant concentrations in the next seven days, fully considers the treatment effects, resource consumption and required expenses of various treatment plans, and makes a comprehensive treatment plan including pollution source control, pollutant treatment and environmental restoration. Compared with a single treatment plan, the system can balance multiple treatment objectives and improve the treatment effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a system module diagram of the present invention;

[0054] Figure 2 It is a flow chart of a smart city data processing method based on Beidou positioning of the present invention.

[0055] In the figure: monitoring point setting module 101, environmental data collection module 102, pollution prediction module 103, comprehensive treatment plan generation module 104. DETAILED DESCRIPTION

[0056] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0057] In a smart city data processing system based on Beidou positioning of the present invention, Figure 1 As shown, including:

[0058] The monitoring point setting module uses the Beidou positioning system to obtain the precise coordinates of the main pollution sources, calculates the concentration distribution of pollutants at different distances through the diffusion model, and thus determines the potential key monitoring areas, and then sets monitoring locations in the key monitoring areas; the calculation formula of the air pollution diffusion model is:

[0059]

[0060] in: is the distance from the source The concentration at ), It is the source of strength ( ), is the average wind speed ( ), and are the horizontal and vertical diffusion parameters, is the effective source height ( ); The calculation formula of the water pollution diffusion model is:

[0061]

[0062] in: is the pollutant concentration at time t and position (x, y, z) , , , are the diffusion parameters of water in the x, y, and z directions, respectively; u, v, and w are the flow velocities of water in the x, y, and z directions, respectively. , The degradation rate constant of the pollutant , represents the source term; the calculation formula of the soil pollution diffusion model is:

[0063]

[0064] in: is the pollutant concentration at time t and position (x, y, z) , is the diffusion parameter in soil, Indicates the flow rate of soil moisture , The degradation rate constant of pollutants in soil , represents the source term;

[0065] Based on the concentration distribution obtained from the simulation, areas with higher concentrations are selected as key monitoring areas, and monitoring points are arranged in the center of the area to implement precise monitoring.

[0066] The environmental data collection module is used to collect urban environmental pollution data. The collection of urban environmental pollution data includes:

[0067] Data are collected once a day at all monitoring locations. The environmental pollution parameters collected at each monitoring location each time include: the average air pollutant concentration of the day, the average water pollutant concentration of the day, the average soil pollutant concentration of the day, and the average relevant meteorological parameters of the day; among them, the air pollutant concentration includes the average PM2.5, PM10, nitrogen dioxide, sulfur dioxide, and carbon monoxide concentrations of the monitoring location on the day; the water pollutant concentration includes the average chemical oxygen demand, biochemical oxygen demand, total phosphorus, total nitrogen, heavy metals, and suspended solids concentrations of the monitoring location on the day; the soil pollutant concentration includes the average heavy metal, polycyclic aromatic hydrocarbon, and organochlorine pesticide concentrations of the monitoring location on the day; and the relevant meteorological parameters include the average temperature, humidity, wind speed, and precipitation of the monitoring location on the day;

[0068] The environmental pollution parameters collected at each monitoring location are encoded to obtain an environmental pollution parameter vector, and the environmental pollution parameter vectors of all monitoring locations on the same day are concatenated to obtain an environmental pollution vector;

[0069] The pollution prediction module inputs the data vector set into the prediction neural network, and the prediction neural network outputs the result indicating whether a pollution warning will occur in different monitoring locations within the next seven days. If the result is yes, it determines whether the time interval between the current date and the date of the previous implementation of the control measures is greater than the set time threshold. If the judgment result is yes, the comprehensive control plan generation module is started; the default value of the time threshold is 3 days;

[0070] Data vector set It is expressed as:

[0071]

[0072] in, , , They represent the environmental pollution vectors of the t-29th, t-28th and tth days respectively; the tth day is defined as the current day; if the number of days between the tth day and the start date of monitoring is less than 30, the missing data vectors are interpolated, and the interpolated data vectors are 0 vectors;

[0073] The prediction neural network consists of a hidden layer and an output layer, and the hidden layer consists of multiple layers of fully connected neurons.

[0074] Assume there are L hidden layers, each containing neurons, , then the calculation formula of the lth hidden layer is:

[0075]

[0076] in, represents the output of the lth layer, L≥l≥1, , is the weight matrix of the lth layer, is the bias vector of the lth layer, and f is the ReLU activation function;

[0077] The output layer outputs the last layer of the hidden layer Converted into prediction results, the formula for calculating the probability of pollution exceeding the threshold at the output layer is as follows:

[0078]

[0079] in, It indicates the probability that the pollution at the monitoring location exceeds the threshold on the next d day, and its value is between 0 and 1. , is the d-th weight matrix of the output layer, represents the output of the Lth hidden layer, is the dth bias vector of the output layer, is the Sigmoid activation function;

[0080] The probability value Convert to binary output :

[0081]

[0082] in, It indicates that the pollution will exceed the threshold on the dth day in the future. It indicates that the pollution will not exceed the threshold on the dth day in the future.

[0083] A comprehensive governance solution generation module is used to input a data vector set into a trained decision neural network, and the decision neural network outputs an action vector;

[0084] Decode the action vector and output a comprehensive treatment plan to implement pollution control at the monitoring site;

[0085] An action vector is represented as:

[0086]

[0087] in, They represent the application degree of the first and h pollution source control measures at monitoring sites 1 and N, respectively;

[0088] It indicates the application degree of the first and nth pollutant treatment measures at monitoring sites 1 and N, respectively;

[0089] Represents the application degree of the first and mth environmental restoration measures at monitoring sites 1 and N, respectively.

[0090] Among them, the degree of application of each measure refers to the intensity or coverage of the corresponding treatment measures taken for specific environmental pollution problems in different monitoring locations; pollution source control measures are aimed at directly reducing or eliminating the generation and emission of pollutants, and controlling pollution from the source, such as: limiting pollution source emissions, vehicle restrictions, etc.; pollutant treatment measures mainly intervene before pollutants come into contact with the environment, that is, they are treated by technical means before entering the environmental system, such as: industrial wastewater pretreatment, waste gas purification, etc. Environmental remediation measures are mainly aimed at the already polluted environment, and various technical means are used to repair and purify the polluted medium (such as soil, water body), such as: soil elution, treatment of polluted water bodies, etc.

[0091] In one embodiment of the present invention, the calculation formula of the decision neural network is as follows:

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098] in, and Represents the data vector set The first and tth environmental pollution data vectors, , , , represents the trainable weight parameters, , , , represents the trainable bias parameter, represents the dot product, represents the tth first intermediate feature, represents the t-th second intermediate feature, represents the tth third intermediate feature, and denote the t-1th and tth output features respectively, represents the output vector, the value of the i-th component of the output vector represents the Q value of the i-th action, tanh represents the tanh function, Represents the Sigmoid function.

[0099] The action output by the decision neural network is the action with the largest Q value.

[0100] The decision neural network uses a reinforcement learning model based on a deep Q network. The deep Q network includes an input layer, a hidden layer, and an output layer. The hidden layer uses a neural network to fit the Q value function, where the neural network includes a policy network and a target network.

[0101] The Q-value function Q(s, a) represents the reward that can be obtained after selecting action a in state s. The reward value is calculated by the reward function, where state s is the data vector at the current moment. , action a is the action vector.

[0102] Set up a reward function. The reward function needs to comprehensively consider the pollution control effect, resource consumption and cost. The control effect is the reduction ratio of pollutant concentration, resource consumption is the amount of resources after the control measures are used, and the cost is the total cost of implementing the control measures. It includes an immediate reward function and a final reward function. The immediate reward function is expressed as:

[0103]

[0104] in, represents the immediate reward at time t, represents the consumption of the selected resource at time t, represents the cost and cost of the governance measures selected at time t, α and β are weight parameters, and their default values ​​are 0.001 and 0.005 respectively;

[0105] The final reward needs to be calculated after the prevention and control measures have been implemented for a period of time and the prevention and control effects are evaluated, which can be expressed as:

[0106]

[0107] Among them, δ, ε, ζ are weight parameters, and their default values ​​are 100, 0.001, and 0.002, respectively. T represents the period from the implementation of governance measures to the evaluation of governance effects. is the final reward for period T, They represent the proportion of pollutant concentration reduction, resource consumption and total cost in period T respectively.

[0108] The specific steps to update the policy network and target network are:

[0109] Step 101, define the target Q value y of the Q value, y is expressed as:

[0110]

[0111] Among them, r is the reward value obtained after executing action a in the current state s, is the discount factor, the default value is 0.9, is the Q value of the target network, is the next state to be transferred after executing action a. For the status Next, select the action;

[0112] Step 102, use the policy network to calculate the current Q value:

[0113]

[0114] in, is the Q value of the policy network, is the current state, is the current action, represents the policy network;

[0115] Step 103, using mean square error (MSE) as the loss function, calculate the error between the current Q value and the target Q value:

[0116]

[0117] in, is the number of mini-batch samples, the default value is 64, is the target Q value of the i-th sample, is the current Q value of the i-th sample;

[0118] Step 104, minimize the loss function through the back propagation algorithm and Adam optimizer, and update the parameters of the policy network, specifically,

[0119]

[0120] in, are the parameters of the policy network before updating, represents the parameters of the updated policy network, is the learning rate, the default value is 0.01, is the gradient of the loss function with respect to the policy network parameters;

[0121] Step 105, define the target network update interval H, the default value of H is 100, and the target network parameters are updated every H training steps; updating the target network parameters is to copy the parameters of the current policy network to the target network.

[0122] When selecting actions, the ε-greedy strategy is used to select actions, including:

[0123] Step 201, ε represents the probability of selecting a random action, and its initial value is 0.9. 1-ε represents the probability of selecting the optimal action. During the training process, ε will decay linearly, specifically:

[0124]

[0125] in: is 0.9, is 0.1, T is the total number of training steps, the default value is 1000, tg step time The value of

[0126] Step 202, generating a random number r between 0 and 1;

[0127] Step 203, if , then choose a random action if Then select the action with the highest current Q value.

[0128] In one embodiment of the present invention, a smart city data processing method based on Beidou positioning is provided. Figure 2 As shown, the following steps are included:

[0129] Step 301, starting from the first day of monitoring, collecting environmental pollution data once a day at different monitoring locations;

[0130] Step 302, encoding daily environmental pollution data into an overall environmental pollution vector;

[0131] Step 303, combine the environmental pollution vectors of a total of 30 days into a data vector set, input the data vector set into the prediction neural network, and output the prediction neural network to indicate whether pollution warning will occur in different monitoring locations in the next seven days. If the result is yes, determine whether the time interval between the current date and the date of the last implementation of the control measures is greater than the set time threshold. If yes, execute step 304; otherwise, execute step 301;

[0132] Step 304, inputting the data vector set into the trained decision neural network, and the decision neural network outputs the action vector;

[0133] Step 305 , decode the action vector, output a comprehensive governance plan for implementation, conduct pollution control in the urban area, and return to step 301 .

[0134] At least one embodiment of the present invention provides a storage medium storing non-temporary computer-readable instructions. When the non-temporary computer-readable instructions are executed by a computer, the steps of the aforementioned smart city data processing method based on Beidou positioning can be executed.

[0135] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A smart city data processing system based on Beidou positioning, characterized in that: include: The monitoring point setting module uses the Beidou positioning system to obtain the precise coordinates of the main pollution sources, calculates the concentration distribution of pollutants at different distances through the diffusion model, and thus determines the potential key monitoring areas, and then sets monitoring sites in the key monitoring areas; The environmental data collection module is used to collect urban environmental pollution data. The collection of urban environmental pollution data includes: The environmental pollution parameters collected at all monitoring locations once a day include: the average air pollutant concentration, the average water pollutant concentration, the average soil pollutant concentration, and the average relevant meteorological parameters of the day; The environmental pollution parameters collected at each monitoring location are encoded to obtain an environmental pollution parameter vector, and the environmental pollution parameter vectors of all monitoring locations on the same day are concatenated to obtain an environmental pollution vector; The pollution prediction module inputs the data vector set into the prediction neural network, and the prediction neural network outputs the result indicating whether a pollution warning will occur in different monitoring locations within the next seven days. If the result is yes, it determines whether the time interval between the current date and the date of the previous implementation of the control measures is greater than the set time threshold. If the judgment result is yes, the comprehensive control plan generation module is started; The data vector set includes environmental pollution vectors from day t-29 to day t; day t is defined as the current day; A comprehensive governance solution generation module is used to input a data vector set into a trained decision neural network, and the decision neural network outputs an action vector; The decision neural network uses a reinforcement learning model based on a deep Q network. The deep Q network includes an input layer, a hidden layer, and an output layer. The hidden layer uses a neural network to fit the Q value function. The neural network includes a policy network and a target network. An action vector is represented as: ; in, They represent the application degree of the first and h pollution source control measures at monitoring sites 1 and N, respectively; It indicates the application degree of the first and nth pollutant treatment measures at monitoring sites 1 and N, respectively; They represent the application degree of the 1st and mth environmental restoration measures at monitoring sites 1 and N, respectively; Decode the action vector and output a comprehensive treatment plan to implement it and treat the pollution at the monitoring site.

2. According to claim 1, a smart city data processing system based on Beidou positioning is characterized in that: If the number of days between the tth day and the start date of monitoring is less than 30, the missing data vector is interpolated and the interpolated data vector is a 0 vector.

3. The smart city data processing system based on Beidou positioning according to claim 1 is characterized in that: The Q-value function Q(s, a) represents the reward that can be obtained after selecting action a in state s. The reward value is calculated by the reward function, where state s is the data vector at the current moment. , action a is the action vector.

4. The smart city data processing system based on Beidou positioning according to claim 3 is characterized in that: Set up a reward function. The reward function needs to comprehensively consider the pollution control effect, resource consumption and cost. The control effect is the reduction ratio of pollutant concentration, resource consumption is the amount of resources after the control measures are used, and the cost is the total cost of implementing the control measures. It includes an immediate reward function and a final reward function. The immediate reward function is expressed as: ; in, represents the immediate reward at time t, represents the consumption of the selected resource at time t, represents the cost and cost of the governance measures selected at time t, α and β are weight parameters, and their default values ​​are 0.001 and 0.005 respectively; The final reward needs to be calculated after the prevention and control measures have been implemented for a period of time and the prevention and control effects are evaluated, which can be expressed as: ; Among them, δ, ε, ζ are weight parameters, and their default values ​​are 100, 0.001, and 0.002, respectively. T represents the period from the implementation of governance measures to the evaluation of governance effects. is the final reward for period T, They represent the proportion of pollutant concentration reduction, resource consumption and total cost in period T respectively.

5. The smart city data processing system based on Beidou positioning according to claim 1 is characterized in that: The specific steps to update the policy network and target network include: Step 101, define the target Q value y of the Q value, y is expressed as: ; Among them, r is the reward value obtained after executing action a in the current state s, is the discount factor, the default value is 0.9, is the Q value of the target network, is the next state to be transferred after executing action a. For the status Next, select the action; Step 102, use the policy network to calculate the current Q value: ; in, is the Q value of the policy network, is the current state, is the current action, represents the policy network; Step 103, using the mean square error as the loss function, calculate the error between the current Q value and the target Q value: ; in, is the number of mini-batch samples, the default value is 64, is the target Q value of the i-th sample, is the current Q value of the i-th sample; Step 104, minimizing the loss function through the back propagation algorithm and the Adam optimizer, and updating the parameters of the policy network; Step 105, define the target network update interval H, the default value of H is 100, and the target network parameters are updated every H training steps; updating the target network parameters is to copy the parameters of the current policy network to the target network.

6. The smart city data processing system based on Beidou positioning according to claim 3 is characterized in that: When selecting actions, the ε-greedy strategy is used to select actions, including: Step 201, ε represents the probability of selecting a random action, and its initial value is 0.

9. 1-ε represents the probability of selecting the optimal action. During the training process, ε will decay linearly, specifically: ; in: is 0.9, is 0.1, T is the total number of training steps, the default value is 1000, tg step time The value of Step 202, generating a random number r between 0 and 1; Step 203, if , then choose a random action if Then select the action with the highest current Q value.

7. A smart city data processing method based on Beidou positioning, characterized in that: Based on a smart city data processing system based on Beidou positioning as described in any one of claims 1-6, the following steps are performed: Step 301, starting from the first day of monitoring, collecting environmental pollution data once a day at different monitoring locations; Step 302, encoding daily environmental pollution data into an overall environmental pollution vector; Step 303, combine the environmental pollution vectors of a total of 30 days into a data vector set, input the data vector set into the prediction neural network, and output the prediction neural network to indicate whether pollution warning will occur in different monitoring locations in the next seven days. If the result is yes, determine whether the time interval between the current date and the date of the last implementation of the control measures is greater than the set time threshold. If yes, execute step 304; otherwise, execute step 301; Step 304, inputting the data vector set into the trained decision neural network, and the decision neural network outputs the action vector; Step 305 , decode the action vector, output a comprehensive governance plan for implementation, conduct pollution control in the urban area, and return to step 301 .

8. A computer device, characterized in that: A smart city data processing system based on Beidou positioning as described in any one of claims 1-6 is run by a processor.

Citation Information

Patent Citations

  • Building environment sudden pollution source identification method combining visual detection and concentration sensing

    CN113627246A

  • Atmospheric pollution prediction method and device based on data mining

    CN114970922A

  • Intelligent management method for watershed water environment

    CN115965496A

  • Method for tracking and locating contamination sources in water distribution systems based on consumer complaints

    US20200208786A1