Urban planning method based on NOx emission detection

Abstract

The invention relates to an urban planning method based on NOx emission detection, and the method comprises the steps: S1, obtaining an urban road network and road network data, obtaining a basic space unit according to the road network, and obtaining the information of the basic space unit according to the number of the road networks; s2, acquiring an explanation variable set and a target variable set based on urban built-up environmental factors and NOx emission, and collecting policy influence to acquire policy information; s3, constructing a mixed effect model representing the relationship between the explaining variable set and the target variable set, and performing training; and S4, carrying out correlation analysis between the city built environment and the NOx emission amount on the sample city by using the mixed effect model, and generating a city planning scheme according to an analysis result. Compared with the prior art, the method has the advantages that the prediction precision of the NOx emission is improved, and the low-NOx emission city planning scheme is generated based on the prediction result.

Description

Technical Field

[0001] The present invention relates to the field of urban environment and traffic emission impact assessment, in particular to a method based on NO x An urban planning approach to emissions monitoring. Background Art

[0002] Heavy-duty diesel trucks (HDDTs) are a key component of the highway transportation system. x ) emissions pose a major challenge to air quality and public health. In China, HDDTs account for only 3% of the total number of vehicles, but contribute 74% of motor vehicle NO x Emissions. Evidence shows that built environment factors and related policy restrictions have a certain impact on vehicle emissions, so urban planning can be used to reduce NO x emission.

[0003] However, existing NOx regulations for heavy diesel trucks are x Research on the impact of emissions mainly focuses on vehicle internal factors or a subset of factors, lacks a comprehensive analysis that takes into account the complex interactions of traffic conditions, road infrastructure, socioeconomic attributes and land use factors, and ignores the interaction between variables; and the differentiated effects of policy interventions have not been fully understood, limiting the ability to design and implement effective emission reduction strategies. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a novel method based on NO x An urban planning approach to emissions monitoring.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A NO-based x An urban planning method for emission detection that comprehensively considers urban built environment factors and relevant policies on vehicle NO x The impact of emissions and the generation of urban planning solutions based on the results to reduce urban NO x emission.

[0007] The method comprises the following steps:

[0008] S1. Obtain the urban road network and road network data, obtain the basic spatial unit according to the road network, and obtain the basic spatial unit information according to the road network number;

[0009] S2. Calculate the urban built environment factor data as the explanatory variable set based on the basic spatial unit information; investigate the vehicle restriction policy, classify the policy impact from the time and space dimensions to obtain policy information; calculate the vehicle NO based on the vehicle data x Emissions data as the target variable set;

[0010] S3, constructing a mixed effect model that represents the relationship between the explanatory variable set and the target variable set and training it;

[0011] S4. Use mixed effect model to analyze urban built environment and NO in sample cities. x Analyze the correlation between emissions and generate urban planning solutions based on the analysis results.

[0012] As a preferred technical solution, S1 specifically includes:

[0013] Acquire the urban road network and road network data, wherein the road network data includes geometric information and attribute information of all roads in the road network;

[0014] Performing spatial overlay analysis on the road network, segmenting the road network according to administrative boundaries to obtain basic spatial units, and assigning unique identifiers to the basic spatial units to establish spatial indexes;

[0015] The basic spatial unit information of the basic spatial unit is extracted based on the identifier and the road network data, wherein the basic spatial unit information includes the area of ​​the basic spatial unit and the lengths of different types of roads in the spatial unit.

[0016] As a preferred technical solution, the explanatory variable set includes sample city land use mixture rate data, traffic facility density data, industrial structure distribution density data and traffic status attribute data; the method for obtaining the explanatory variable set includes the following steps:

[0017] The land use mixing rate of the basic spatial unit is calculated based on the identifier, and the calculation formula is as follows:

[0018] p μ,m =S μ,m / S μ

[0019]

[0020] Among them, p μ,m Indicates the proportion of land use type m in the basic spatial unit with identifier μ; S μ,m represents the area of ​​land of type m; M represents the total number of land types; H μ represents the land use mixture ratio of the basic spatial unit μ;

[0021] The traffic facility density data is calculated based on the identifier and basic space unit information, and the calculation formula is as follows:

[0022] Density μ,r = l μ,r / S μ

[0023] Density μ,road =Number μ,road / S μ

[0024] Among them, Density μ,r Represents the density of type r roads of the basic spatial unit with identifier μ; Density μ,road Represents the road intersection density of basic spatial unit i; Number μ,road The number of road intersections representing the basic spatial unit μ;

[0025] The industrial structure distribution density is calculated based on the identifier, and the calculation formula is as follows:

[0026] Density μ,industty =Number μ,industry / S μ

[0027] Among them, Density μ,industry Indicates the density of different types of industries in the basic spatial unit with spatial identifier μ; Number μ,industry is the number of different types of industries in the basic spatial unit μ;

[0028] Traffic state attributes are acquired based on the identifier, where the traffic state attributes include a congestion level, an average vehicle speed, and a vehicle speed standard deviation.

[0029] As a preferred technical solution, the method for obtaining policy information includes the following steps:

[0030] The survey covers the period of no driving restrictions, the period of driving restrictions for vehicles with specific emission standards, and the morning and evening peak periods of comprehensive driving restrictions, capturing the changes in policies over time.

[0031] Different restricted areas with clear boundaries are divided according to geographical location, functional attributes and policy implementation, and each restricted area is given a spatial identification;

[0032] Integrate time and space to analyze policy impacts and classify similar policy provisions to obtain policy information.

[0033] As a preferred technical solution, the method for obtaining the target variable set includes the following steps:

[0034] Acquire vehicle data and preprocess the data to obtain valid data, wherein the preprocessing includes removing duplicate, missing and abnormal data;

[0035] Calculate the vehicle NO based on the valid data x The emission is calculated as follows:

[0036]

[0037] in, represents the modified emission factor; μ represents the space identifier of the basic space; b represents the emission standard; w represents the vehicle load level; represents the traffic speed at a time interval of τ on a road of type r; α, β, γ, δ, ε, θ, is the relevant parameter; Fcorr Fuel and Fcorr Base is the fuel correction factor; L μ is the vehicle running distance.

[0038] As a preferred technical solution, the mixed effect model is:

[0039] y i =f(X i )+Z i b i +∈ i ,

[0040] b i ~N(0,D),∈ i ~N(0,R i ),i=1,…,n,

[0041] Among them, the total number of samples in cluster i is is the response variable within cluster i; is the explanatory variable; random effects covariate matrix; is the random effect vector within cluster i, which follows a multivariate normal distribution; is the error vector, which obeys the multivariate normal distribution; f(X i ) is the fixed effect function.

[0042] As a preferred technical solution, the training includes the following specific steps: matching and fusing the explanatory variable set, target variable set and policy information based on timestamp and spatial coordinates to obtain a data set; dividing the data set according to a specific ratio to obtain a training set, a validation set and a test set; estimating model parameters based on the training set; using the validation set and the test set to evaluate the estimated parameters of the obtained model, if the evaluation passes, the training is completed, otherwise the parameters are re-estimated.

[0043] As a preferred technical solution, the parameter estimation method includes the following steps: replacing the fixed effect f(X) with a machine learning method based on gradient boosting i ); design random effect grouping variables based on time dimension, space dimension and policy information; calculate the covariance matrix of each random effect grouping variable, and update the random effect vector according to the covariance matrix.

[0044] As a preferred technical solution, designing random effect grouping variables based on the time dimension means dividing time periods according to the level of activity of human social activities, obtaining and using the data of urban built environment factors and NO in the corresponding time period. x Emission data, design random effect grouping variables;

[0045] Designing random effect grouping variables based on spatial dimension means dividing the regions according to urban planning and obtaining and using the urban built environment factor data and NO in the corresponding regions. x Emission data, design random effect grouping variables;

[0046] Designing random effect grouping variables based on policy factors means dividing regions according to the type of policy impact, obtaining and using the urban built environment factor data and NO in the corresponding region. x Emissions data, design random effects grouping variables.

[0047] As a preferred technical solution, SHAP value is used to compare urban built environment factors and NO x Emissions correlation analysis includes the following steps:

[0048] The total SHAP value for each built environment factor is calculated using the following formula:

[0049]

[0050] Among them, P is the set of all urban built environment factors q, H is the number of trees, L h is the number of leaves in the tree h, is the number of samples that reach node j in tree h, is the value of urban built environment factor q at node j; is the value of the previous node of the urban built environment factor q;

[0051] The average absolute value of the total SHAP value of each city’s built environment factors was calculated as the correlation analysis indicator.

[0052] Compared with the prior art, the feature importance evaluation and nonlinear relationship interpretation framework based on SHAP analysis in this paper comprehensively considers the complex interaction effects of traffic conditions, road infrastructure, socio-economic attributes and land use factors, and can identify the influence of NO in more detail. x Key built environment factors of NO emissions, from time, space and policy factors x The nonlinear relationship analysis of NOx emissions significantly improved the accuracy of NOx emission prediction and assessment; quantified the impact of these factors on NOx emissions under different policy scenarios. x The specific contribution of HDDTs to NO x The relationship between emissions and urban built environment factors and policy impacts is the key to building a low NO x Provide guidance to the emitting cities. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flow chart of the method of the present invention;

[0054] Figure 2 Flowchart for designing mixed effects model for the present invention;

[0055] Figure 3 Schematic diagram of the interactive relationship between HDDTs speed and urban built environment factors according to an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0057] This embodiment provides a NO-based x The urban planning method for emission detection is shown in the flowchart of Shanghai as a sample city. Figure 1 As shown, it includes designing a mixed effect model to predict the impact of urban built environment on vehicle NOx emissions; using the mixed effect model to analyze the correlation between urban built environment and NOx emissions in sample cities, and generating urban planning plans based on the analysis results.

[0058] The flow chart for designing a mixed effects model is as follows: Figure 2 As shown, the following steps are included:

[0059] S1. Obtain the basic spatial unit information of Shanghai:

[0060] S11. Obtain the overall road network data and vector boundaries of Shanghai from OpenStreetMap, and obtain road network data, wherein the road network data includes geometric information and attribute information of all roads in the road network.

[0061] S12. Use GIS software to perform spatial overlay analysis on the road network, segment it according to the street or township level to obtain basic spatial units, and assign unique identifiers to the basic spatial units to establish spatial indexes.

[0062] S13, extracting basic spatial unit information of the basic spatial unit based on the identifier and the road network data, the basic spatial unit information including the area S of each basic spatial unit i μ and the length l of different road types r μ,r .

[0063] S2. Obtain explanatory variable set, policy information and target variable set:

[0064] S21. Calculating the built environment factor data of Shanghai based on the basic spatial unit information as an explanatory variable set includes the following steps:

[0065] S211. Use remote sensing image data to classify land use in the study area, obtain the distribution of industrial land, public management and service land, commercial land, etc., and calculate the land use mixing rate of the basic spatial unit according to the following formula:

[0066] p μ,m =S μ,m / S μ

[0067]

[0068] Among them, p μ,m Indicates the proportion of land use type m in the basic spatial unit with identifier μ; S μ,m represents the area of ​​land of type m; M represents the total number of land types; H μ represents the land use mixture ratio of the basic spatial unit μ;

[0069] S212. Extract different types of roads from OpenStreetMap data, use GIS tools to calculate the total length of each road type, and obtain the road intersection density. The calculation formula is as follows:

[0070] Density μ,r = l μ,r / S μ

[0071] Density μ,road =Number μ,road / Sμ

[0072] Among them, Density μ,r Represents the density of type r roads of the basic spatial unit with identifier μ; Density μ,road Represents the road intersection density of basic spatial unit i; Number μ,road The number of road intersections representing the basic spatial unit μ;

[0073] S213. Based on the industrial park data obtained from the Amap API and combined with the industry type information of each park, calculate the number of enterprises in the primary, secondary and tertiary industries in each basic spatial unit, and calculate the industrial structure distribution density. The calculation formula is as follows:

[0074] Density μ,industry =Number μ,industry / S μ

[0075] Among them, Density μ,industry Indicates the density of different types of industries in the basic spatial unit with spatial identifier μ; Number μ,industry is the number of different types of industries in the basic spatial unit μ;

[0076] S214. Use the data from the seventh national census and the economic data published by the local government of the study area to obtain the population and GDP per capita of each basic spatial unit. Use the GPS data of heavy trucks in combination with the geographic information system to analyze the average speed, speed standard deviation and congestion level of the road.

[0077] S22. Investigate vehicle restriction policies and classify policy impacts from the time and space dimensions to obtain policy information.

[0078] S221. In this embodiment, Shanghai's policy regulations for HDDTs are divided into four time periods according to the time and space dimensions, including T1 (1:00 a.m. to 6:00 a.m., unlimited travel), T2 (6:00 a.m. to 7:00 a.m. and 8:00 p.m. to 1:00 a.m. the next day, restrictions for China 3 standards), T3 (7:00 a.m. to 10:00 a.m. and 5:00 p.m. to 8:00 p.m., morning and evening peak hours with comprehensive restrictions) and T4 (10:00 a.m. to 5:00 p.m., unlimited travel);

[0079] S222, divided into three areas according to the spatial dimension, including Z1 (central urban area of ​​the study area), Z2 (area between the inner and outer rings) and Z3 (area outside the outer ring);

[0080] S223. Policy restriction units are formed by combining the time and space dimensions, including PRU1 (no time and space policy restrictions), PRU2 (policy restrictions for diesel vehicles of National III standard), PRU3 (policy restrictions for local / non-local HDDTs) and PRU4 (the strictest time and space policy restrictions).

[0081] S23, calculating vehicle NO based on vehicle data x Emissions data as the target variable set:

[0082] S231, collect GPS trajectory data and basic vehicle information of heavy diesel trucks, including vehicle ID, latitude and longitude, timestamp, vehicle type and other attributes;

[0083] S232, perform preliminary cleaning on the collected data and calculate the average value, standard deviation, minimum value and maximum value, remove abnormal values ​​and incomplete records, and obtain valid data;

[0084] S233, calculating vehicle NO based on the valid data x The emission is calculated as follows:

[0085]

[0086] in, represents the modified emission factor; μ represents the space identifier of the basic space; b represents the emission standard; w represents the vehicle load level; represents the traffic speed at a time interval of τ on a road of type r; α, β, γ, δ, ε, θ, is the relevant parameter; Fcorr Fuel and Fcorr Base is the fuel correction factor; L μ is the vehicle running distance.

[0087] S3. Construct a mixed effect model that represents the relationship between the explanatory variable set and the target variable set and train it. Specifically:

[0088] S31. Construct a mixed effect model with the following structure:

[0089] y i =f(X i )+Z i b i +∈ i ,

[0090] b i ~N(0,D),∈ i ~N(0,R i ),i=1,…,n,

[0091] Among them, the total number of samples in cluster i is is the response variable within cluster i; is the explanatory variable; random effects covariate matrix; is the random effect vector within cluster i, which follows a multivariate normal distribution; is the error vector, which obeys the multivariate normal distribution; f(X i ) is the fixed effect function;

[0092] And, the random effects covariate matrix It can be constructed along multiple dimensions, including the temporal dimension (considering changes at a specific time), the spatial dimension (capturing the impact of a specific location), and the impact of policy restrictions. The flexibility in defining Z allows different clustering schemes to be established based on these dimensions alone or in combination.

[0093] S32, matching and fusing the explanatory variable set, the target variable set and the policy information based on the timestamp and the spatial coordinates to obtain a data set;

[0094] S33, dividing the data set into a training set, a validation set, and a test set in a ratio of 6:2:2;

[0095] S34. Train the mixed effects model based on the training set:

[0096] S341, based on the training set, the machine learning method of gradient boosting is used to replace the fixed effect f(X i ), thereby improving model training speed and prediction efficiency;

[0097] S342. Design random effect grouping variables based on time dimension, space dimension and policy information, and calculate the covariance matrix of each random effect grouping variable Where Z i is the random effects covariate matrix, is the covariance matrix of the random effects estimated in the previous iteration, represents the variance contribution of the random effects, is the residual variance estimated in the previous iteration, I n is the identity matrix, corresponding to the uncertainty of the residual;

[0098] S343. Update the random effect vector based on the covariance matrix Estimation of model parameters is performed, where is the transpose of the random effects covariate matrix, which is used to adjust the random effects in the model. is the inverse matrix of the covariance matrix of the random effect, which represents the uncertainty of the random effect; It is the residual after subtracting the fixed effects, which is used to update the estimate of random effects after eliminating the fixed effects;

[0099] S35. Use the validation set and the test set to evaluate the estimated parameters of the obtained model. If the evaluation passes, the training is completed and the mixed effect model is obtained. Otherwise, the parameters are re-estimated.

[0100] In this experiment, the root mean square error, mean absolute error and coefficient of determination are selected to compare the performance of the selected model with the established baseline method, including the following steps:

[0101] Set the candidate adjustment parameters to learning_rate∈{0.1,0.05,0.01}, max_depth∈{-1,1,5,10}, the number of boosting rounds to n_estimators∈{50,…,1000}, the minimum number of samples required to form a leaf to min_data_in_leaf∈{10,…,100}; the number of boosting iterations∈{1,…,1000};

[0102] A 5-fold cross validation was used on the training data for each run, and the combination of tuning parameters was selected from the comprehensive grid to minimize the average mean squared error on the validation data;

[0103] The root mean square error, mean absolute error and determination coefficient of each model are calculated respectively. The specific values ​​are shown in Table 1. It can be seen that the model selected by the present invention has a smaller degree of dispersion of the deviation distribution and a better fitting effect, and its error with the true value is small, and the performance is excellent.

[0104] Table 1

[0105]

[0106]

[0107] The mixed effect model was used to analyze the urban built environment factors and NO through SHAP value. x For the emission correlation analysis, first, the total SHAP value of each built environment factor is calculated as follows:

[0108]

[0109] Among them, P is the set of all urban built environment factors q, H is the number of trees, L h is the number of leaves in the tree h, is the number of samples that reach node j in tree h, is the value of urban built environment factor q at node j; is the value of the previous node of the urban built environment factor q;

[0110] Then, the average absolute value of the total SHAP value of each city’s built environment factors was calculated as the correlation analysis index;

[0111] Draw the SHAP dependency graph based on the calculation results. Figure 3 As shown;

[0112] In this embodiment, industrial land density and sidewalk density are taken as examples to analyze Figures (3a) and (3b). From (3a), it can be found that there is a significant interaction between industrial land density and speed. At low speed, the SHAP value corresponding to high industrial land density (represented by red and purple points) is higher, that is, NOx emissions are higher. However, as the speed increases, the SHAP value of the high industrial land density area decreases faster, indicating that in these areas, higher speeds can more effectively reduce NOx emissions. From Figure (3b), it can be found that in areas with high sidewalk density, when the speed is within the range of 10-60 km / h and is lower than 42 km / h, the SHAP value is positive, indicating that NO x When the speed exceeds 42 km / h, the SHAP value is negative, indicating that the emission is reduced, while when the speed is within the range of 60-70 km / h, the SHAP value is the lowest, indicating that the NOx emission is the lowest at this time.

[0113] Based on the above results, the following urban planning scheme can be formed, including:

[0114] At the traffic management level: Implement targeted speed management policies in industrially intensive areas, such as limiting vehicle speeds to 60-70km / h to optimize environmental effects.

[0115] At the land use planning level: Considering the impact of industrial land layout on traffic emissions, it may be necessary to design a more reasonable road system around industrial areas.

[0116] At the environmental policy formulation level: In the outer ring areas of the city, the restrictions on relevant policies should be strengthened to avoid NO x There is a risk of exceeding the emission standard. In the inner ring area of ​​the city, the traffic restriction policy should be strictly implemented and the speed should be kept within the prescribed speed limit to ensure low NO x emission.

[0117] At the infrastructure planning level: add environmental monitoring stations or implement stricter emission control measures in high emission risk areas.

[0118] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A NO-based x Urban planning approach to emissions monitoring, It is characterized in that The method includes: S1. Obtain the urban road network and road network data, obtain the basic spatial unit according to the road network, and obtain the basic spatial unit information according to the road network number; S2. Calculate the urban built environment factor data as the explanatory variable set based on the basic spatial unit information; investigate the vehicle restriction policy, classify the policy impact from the time and space dimensions to obtain policy information; calculate the vehicle NO based on the vehicle data x Emissions data as the target variable set; S3, constructing a mixed effect model that represents the relationship between the explanatory variable set and the target variable set and training it; S4. Use mixed effect model to analyze urban built environment and NO in sample cities. x Analyze the correlation between emissions and generate urban planning solutions based on the analysis results.

2. A method based on NO according to claim 1. x Urban planning approach to emissions monitoring, It is characterized in that The S1 specifically includes: Acquire the urban road network and road network data, wherein the road network data includes geometric information and attribute information of all roads in the road network; Performing spatial overlay analysis on the road network, segmenting the road network according to administrative boundaries to obtain basic spatial units, and assigning unique identifiers to the basic spatial units to establish spatial indexes; The basic spatial unit information of the basic spatial unit is extracted based on the identifier and the road network data, wherein the basic spatial unit information includes the area of ​​the basic spatial unit and the lengths of different types of roads in the spatial unit.

3. A method based on NO according to claim 2 x Urban planning approach to emissions monitoring, It is characterized in that The explanatory variable set includes sample urban land use mixed rate data, traffic facility density data, industrial structure distribution density data and traffic status attribute data; The method to obtain the set of explanatory variables includes the following steps: The land use mixing rate of the basic spatial unit is calculated based on the identifier, and the calculation formula is as follows: p μ,m =S μ,m / S μ Among them, p μ,m Indicates the proportion of land use type m in the basic spatial unit with identifier μ; S μ,m represents the area of ​​land of type m; M represents the total number of land types; H μ represents the land use mixture ratio of the basic spatial unit μ; The traffic facility density data is calculated based on the identifier and basic space unit information, and the calculation formula is as follows: Density μ,r =l μ,r / S μ Density μ,road =Number μ,road / S μ Among them, Density μ,r Represents the density of type r roads of the basic spatial unit with identifier μ; Density μ,road Represents the road intersection density of basic spatial unit i; Number μ,road The number of road intersections representing the basic spatial unit μ; The industrial structure distribution density is calculated based on the identifier, and the calculation formula is as follows: Density μ,industry =Number μ,industry / S μ Among them, Density μ,industry Indicates the density of different types of industries in the basic spatial unit with spatial identifier μ; Number μ,industry is the number of different types of industries in the basic spatial unit μ; Traffic state attributes are acquired based on the identifier, where the traffic state attributes include a congestion level, an average vehicle speed, and a vehicle speed standard deviation.

4. A method according to claim 1 based on NO x Urban planning approach to emissions monitoring, It is characterized in that The method for obtaining policy information includes the following steps: The survey covers the period of no driving restrictions, the period of driving restrictions for vehicles with specific emission standards, and the morning and evening peak periods of comprehensive driving restrictions, capturing the changes in policies over time. Different restricted areas with clear boundaries are divided according to geographical location, functional attributes and policy implementation, and each restricted area is given a spatial identification; Integrate time and space dimensions to analyze policy impacts and classify similar policy provisions to obtain policy information.

5. A method according to claim 1 based on NO x Urban planning approach to emissions monitoring, It is characterized in that The method for obtaining the target variable set comprises the following steps: Acquire vehicle data and preprocess the data to obtain valid data, wherein the preprocessing includes removing duplicate, missing and abnormal data; Calculate the vehicle NO based on the valid data x The emission is calculated as follows: in, represents the modified emission factor; μ represents the space identifier of the basic space; b represents the emission standard; w represents the vehicle load level; represents the traffic speed at a time interval of τ on a road of type r; α, β, γ, δ, ε, θ, is the relevant parameter; Fcorr Fuel and Fcorr Base is the fuel correction factor; L μ is the vehicle running distance.

6. A method according to claim 1 based on NO x Urban planning approach to emissions monitoring, It is characterized in that The mixed effects model is: y i =f(X i )+Z i b i +∈ i , b i ~N(0,D),∈ i ~N(0,R i ),i=1,…,n, Among them, the total number of samples in cluster i is is the response variable within cluster i; is the explanatory variable; random effects covariate matrix; is the random effect vector within cluster i, which follows a multivariate normal distribution; is the error vector, which obeys the multivariate normal distribution; f(X i ) is the fixed effect function.

7. A method according to claim 1 based on NO x Urban planning approach to emissions monitoring, It is characterized in that The training includes the following specific steps: Based on timestamp and spatial coordinates, the explanatory variable set, target variable set and policy information are matched and fused to obtain a data set; the data set is divided into a training set, a validation set and a test set according to a specific ratio; model parameters are estimated based on the training set; the estimated parameters of the obtained model are evaluated using the validation set and the test set. If the evaluation passes, the training is completed, otherwise the parameters are re-estimated.

8. A method according to claim 7 based on NO x Urban planning approach to emissions monitoring, It is characterized in that The parameter estimation method comprises the following steps: The machine learning method based on gradient boosting replaces the fixed effect f(X i ); Design random effect grouping variables based on time dimension, space dimension and policy information; The covariance matrix of each random effect grouping variable is calculated, and the random effect vector is updated according to the covariance matrix.

9. A method according to claim 8 based on NO x Urban planning approach to emissions monitoring, It is characterized in that Designing random effect grouping variables based on the time dimension means dividing time periods according to the level of activity of human social activities, obtaining and using the data of urban built environment factors and NO in the corresponding time period. x Emission data, design random effect grouping variables; Designing random effect grouping variables based on spatial dimension means dividing the regions according to urban planning and obtaining and using the urban built environment factor data and NO in the corresponding regions. x Emission data, design random effect grouping variables; Designing random effect grouping variables based on policy factors means dividing regions according to the type of policy impact, obtaining and using the urban built environment factor data and NO in the corresponding region. x Emissions data, design random effects grouping variables.

10. A method according to claim 1 based on NO x Urban planning approach to emissions monitoring, It is characterized in that Urban built environment factors and NO x Emissions correlation analysis includes the following steps: The total SHAP value for each built environment factor is calculated using the following formula: Among them, P is the set of all urban built environment factors q, H is the number of trees, L h is the number of leaves in the tree h, is the number of samples that reach node j in tree h, is the value of urban built environment factor q at node j; is the value of the previous node of the urban built environment factor q; The average absolute value of the total SHAP value of each city’s built environment factors was calculated as the correlation analysis indicator.

Images