A method for predicting carbon emissions from motor vehicles on a road network

Abstract

A method for predicting carbon emissions from motor vehicles on a road network belongs to the technical field of road carbon emissions prediction. In order to solve the problem of accurately predicting road carbon emissions, the present invention takes the urban road network as the research object, divides it into sections with intersections as the boundaries, collects road network data for each section, and constructs a road network data model; collects real-time dynamic traffic information data on section traffic flow, vehicle driving conditions, and vehicle types, constructs a carbon emission calculation model for motor vehicles on the urban road network, and calculates carbon emission data for motor vehicles on the urban road network; constructs a carbon emission grouping model for motor vehicles on the road network to obtain grouped carbon emission data for motor vehicles on the urban road network; based on the grouped carbon emission data for motor vehicles on the urban road network obtained in step S3, a long short-term memory neural network is used to construct a carbon emission prediction model for motor vehicles on the road network to predict carbon emissions for motor vehicles on the road network. The present invention makes accurate predictions.

Description

Technical Field

[0001] The present invention belongs to the technical field of road carbon emission prediction, and in particular relates to a method for predicting carbon emissions from motor vehicle travel on a road network. Background Art

[0002] With the continuous acceleration of urbanization and the rapid growth of the number of motor vehicles, road traffic has become one of the main threats to urban air quality. Carbon emissions from motor vehicles cause serious pollution to the atmospheric environment, aggravate the degree of global climate change, and have a direct impact on the health and quality of life of residents. Therefore, it is necessary to accurately predict and monitor carbon emissions from motor vehicles on the road network in order to take effective measures to reduce emissions and protect the environment and human health.

[0003] In the current technological context, although we have made significant progress in many areas, there are still challenges in accurately predicting carbon emissions from motor vehicle travel on road networks. Traditional prediction methods often rely on simple statistical models or linear regression, which often ignore the complex relationships and long-term dependencies in time series data. Therefore, the prediction accuracy is often not high and it is difficult to meet actual needs. Summary of the invention

[0004] The problem to be solved by the present invention is to accurately predict road carbon emissions, and a method for predicting carbon emissions from motor vehicles on a road network is proposed.

[0005] To achieve the above object, the present invention is implemented through the following technical solutions:

[0006] A method for predicting carbon emissions of motor vehicles on a road network comprises the following steps:

[0007] S1. Take the urban road network as the research object, divide it into sections with intersections as boundaries, collect road network data for each section, and build a road network data model;

[0008] S2. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types, build a carbon emission calculation model for motor vehicles on the urban road network, and calculate carbon emission data for motor vehicles on the urban road network;

[0009] S3. Based on the urban road network motor vehicle driving carbon emission data calculated in step S2, a road network motor vehicle driving carbon emission grouping model is constructed to obtain grouped urban road network motor vehicle driving carbon emission data;

[0010] S4. Based on the grouped urban road network motor vehicle driving carbon emission data obtained in step S3, a road network motor vehicle driving carbon emission prediction model is constructed using a long short-term memory neural network to predict the road network motor vehicle driving carbon emission.

[0011] Furthermore, the specific implementation method of constructing the road network data model in step S1 is to take the urban road network as the research object and divide it into road segment links with intersections as boundaries. The research object is the links including all urban highways, main roads, secondary roads and some branches. The fields of each link include the segment number, the previous segment number, the next segment number, the road type, the length, the speed limit value, the direction and the name.

[0012] Furthermore, the length of the road section in step S1 is divided into sections not exceeding 500 meters.

[0013] Furthermore, the specific implementation method of step S2 includes the following steps:

[0014] S2.1. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types;

[0015] S2.2. Based on the real-time dynamic traffic information data collected in step S2.1, calculate the Type of vehicle The period is Carbon emissions from driving on road sections , the calculation formula is:

[0016] ;

[0017] in, Indicates working conditions In the state Type of vehicle The period is The number of vehicles traveling on the road segment, Indicates working conditions In the state The emission coefficients of the vehicles of different types, Indicates The length of the road section;

[0018] S2.3. Calculate the The period is Carbon emissions from driving on road sections , the calculation formula is:

[0019] ;

[0020] Real-time dynamic traffic information such as road section traffic flow, vehicle driving conditions, and vehicle types are extracted, and the emissions of the road section are dynamically estimated through a bottom-up approach.

[0021] Furthermore, in step S2, the working condition It consists of three parameters: average vehicle speed, relative positive acceleration and parking time ratio, and the three parameters have the same weight.

[0022] Furthermore, the specific implementation method of step S3 includes the following steps:

[0023] S3.1. Considering that the carbon emissions of motor vehicles on the road network are time series data, a time series curve is constructed for each research unit section at an interval of 1 hour. The carbon emission series of motor vehicles on the road network is obtained by 24 hours a day. , different from Other sections of the road The carbon emission sequence of motor vehicles on the road network is obtained every 24 hours ;

[0024] S3.2. Create a 24×24 matrix T and replace the Euclidean distance of the dynamic time warping algorithm with the square of the difference between the estimated derivatives of the two sequences obtained in step S3.1 to obtain The first section Carbon emissions during the period The first section Distance of time period The calculation formula is:

[0025] ;

[0026] S3.3. The cumulative distance for a regular path is the sum of the distance of the current point and the distance of the smallest neighboring element that can reach this point. The calculation formula is:

[0027] ;

[0028] in, For the The first section Carbon emissions during the period The first section Cumulative similarity of carbon emissions over time periods;

[0029] S3.4. Perform iterative calculations based on steps S3.1-S3.3, starting from the first road segment, and calculate the cumulative similarity between the first road segment and all other road segments. The cumulative similarity value is assumed to be 0.01, that is, all road segments with cumulative similarity values ​​less than or equal to 0.01 with the first road segment are considered as a group. For those that are not grouped together with the first road segment, repeat the calculation of the cumulative similarity between each other and ensure that all road segments are divided into corresponding groups.

[0030] Furthermore, the specific implementation method of step S4 using the improved long short-term memory neural network considering the association between adjacent road sections to construct a road network motor vehicle driving carbon emission prediction model includes the following steps:

[0031] S4.1. Perform input calculation: Calculate the linear transformation result of the input data and the hidden state of the previous time step, multiply the grouped urban road network motor vehicle driving carbon emission data obtained in step S3 by the corresponding learnable weight matrix, and then add the bias parameter to obtain the linear transformation result of the input data. The calculation formula is as follows:

[0032] ;

[0033] in, is the output of the input gate, is the hidden state at the previous moment, are the learnable weights of the input, is the input bias parameter, After grouping The period is Carbon emissions from driving on each road segment;

[0034] S4.2. Calculate candidate memory cells using the following formula:

[0035] ;

[0036] in, is the output of the candidate memory unit, tanh is the hyperbolic tangent activation function, is the hidden state at the previous moment, are candidate learnable weights, is a candidate bias parameter;

[0037] S4.3. Perform forgetting calculation and calculate the output value of the forgetting gate. The calculation formula is as follows:

[0038] ;

[0039] in, is the output of the forget gate, is the sigmoid activation function, are the forgettable learnable weights; is the forgotten bias parameter;

[0040] S4.4. Update the memory unit according to the forgetting calculation, calculate the forgetting degree of the memory unit by the forget gate, and use the output of the forget gate Perform element-level multiplication on the memory unit of the previous time step. The calculation formula is: forget gate output × previous time step memory unit; calculate the input gate For candidate memory cells To determine the update degree, use the input gate to perform element-level multiplication operation on the candidate memory unit of the current time step. The calculation formula is: input gate output × current time step candidate memory unit; update memory unit , add the product of the forget gate output and the input gate output to update the memory unit. The comprehensive calculation formula is:

[0041] ;

[0042] S4.5. Update the hidden state and calculate the output degree of the output gate to the memory unit. Use the output gate Perform element-level multiplication on the memory unit of the current time step, and the calculation formula is:

[0043] ;

[0044] in, is the hidden state at the current moment, The output of the output gate;

[0045] S4.6. Perform output calculation, the calculation formula is:

[0046] ;

[0047] in, are the learnable weights of the output;

[0048] S4.7. Repeat steps S4.1 to S4.6 until all time steps have been processed. Calculate the value of the loss function based on the model output and the true label;

[0049] S4.8. Update parameters and repeat steps S4.1 to S4.7 until the predetermined number of training rounds or convergence condition is reached. Use the trained model to predict new input data to obtain future prediction values.

[0050] Beneficial effects of the present invention:

[0051] The method for predicting carbon emissions from motor vehicles on a road network described in the present invention takes into account the long-term trend and short-term characteristic changes of carbon emissions from vehicles on the road network, that is, there may be long-term periodic laws, but short-term changes can more affect future trends and show more important weights, which is specifically reflected in the fact that carbon emissions from a road network are similar to the traffic conditions of the road network and have a transmission mechanism. The traffic conditions or carbon emissions of a certain section of a road are caused by the upstream. Accurate predictions can provide a reliable reference for the design of policies under different scenarios.

[0052] The method for predicting carbon emissions of motor vehicles on a road network described in the present invention can reliably predict carbon emissions for the next 7 times 24 hours based on historical data of each hour for at least 8 weeks.

[0053] The method for predicting carbon emissions from motor vehicles on a road network described in the present invention utilizes the long-term and short-term dependencies in the sequence data to capture subtle changes in past carbon emission trends and make accurate future predictions based on them. This feature ensures that the model will not have gradient vanishing or gradient exploding problems when processing long-term sequence data, thereby improving the stability and accuracy of the prediction.

[0054] The method for predicting carbon emissions of motor vehicles on a road network described in the present invention can achieve a prediction accuracy of more than 95% in the absence of emergencies.

[0055] The method for predicting carbon emissions from motor vehicles on a road network described in the present invention can more accurately capture the long-term dependencies in time series data by introducing advanced sequence data processing technology and utilizing the characteristics of long short-term memory networks, thereby improving the accuracy and stability of carbon emission prediction. This method can not only provide more reliable decision support for fields such as traffic management and environmental protection, but also help reduce the impact of air pollution on human health. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of a method for predicting carbon emissions from motor vehicles traveling on a road network according to the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the specific embodiments described are only part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0058] Therefore, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected specific embodiments of the present invention. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0059] In order to further understand the content, features and effects of the present invention, the following specific implementation methods are given as examples, and the attached Figure 1The detailed instructions are as follows:

[0060] Embodiment 1:

[0061] A method for predicting carbon emissions of motor vehicles on a road network comprises the following steps:

[0062] S1. Take the urban road network as the research object, divide it into sections with intersections as boundaries, collect road network data for each section, and build a road network data model;

[0063] Furthermore, the specific implementation method of step S1 to construct the road network data model is to take the urban road network as the research object, and divide it into road segment links with intersections as boundaries. The research object is the links including all urban highways, main roads, secondary roads and some branches. The fields of each link include the road segment number, the previous road segment number, the next road segment number, the road type, the length, the speed limit value, the direction and the name;

[0064] Furthermore, the length of the road section in step S1 is divided into sections not exceeding 500 meters;

[0065] S2. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types, build a carbon emission calculation model for motor vehicles on the urban road network, and calculate carbon emission data for motor vehicles on the urban road network;

[0066] Furthermore, the specific implementation method of step S2 includes the following steps:

[0067] S2.1. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types;

[0068] S2.2. Based on the real-time dynamic traffic information data collected in step S2.1, calculate the Type of vehicle The period is Carbon emissions from driving on road sections , the calculation formula is:

[0069] ;

[0070] in, Indicates working conditions In the state Type of vehicle The period is The number of vehicles traveling on the road segment, Indicates working conditions In the state The emission coefficients of the vehicles of different types, Indicates The length of the road section;

[0071] S2.3. Calculate the The period is Carbon emissions from driving on road sections , the calculation formula is:

[0072] ;

[0073] Extract real-time dynamic traffic information such as road section traffic flow, vehicle driving conditions, and vehicle types, and dynamically estimate the emissions of the road section through a bottom-up approach;

[0074] Furthermore, in step S2, the working condition It consists of three parameters: average vehicle speed, relative positive acceleration and parking time ratio, and the three parameters have the same weight;

[0075] Furthermore, the road network traffic information is derived from multi-source traffic data, among which the real-time speed information of the road network is extracted from the urban road traffic operation index system, and the trajectory data can be used for operating vehicles. The road section vehicle type data is derived from the checkpoint license plate data, and the data fields include checkpoint number, recognition time, license plate color, license plate type, etc. Some sections do not have checkpoint videos, and radar monitoring data and microwave detection are used to detect the traffic flow of each section in both directions.

[0076] The carbon emission factor refers to the carbon emissions generated when a vehicle travels one kilometer, which involves two parts: static properties such as vehicle type and dynamic properties of driving conditions.

[0077] Vehicle type: According to the National Technical Guidelines for Air Pollutant Emission Inventories of Road Motor Vehicles in China, it can be divided into vehicle type, fuel type, and emission standard classification.

[0078] The driving conditions obtain carbon emission factor data of highways, expressways, main roads, secondary roads and branch roads at different levels, such as smooth, basically smooth and slow. The service level is composed of three indicators: average speed, stop time ratio and relative positive acceleration.

[0079] S3. Based on the urban road network motor vehicle driving carbon emission data calculated in step S2, a road network motor vehicle driving carbon emission grouping model is constructed to obtain grouped urban road network motor vehicle driving carbon emission data;

[0080] Furthermore, the specific implementation method of step S3 includes the following steps:

[0081] S3.1. Considering that the carbon emissions of motor vehicles on the road network are time series data, a time series curve is constructed for each research unit section at an interval of 1 hour. The carbon emission series of motor vehicles on the road network is obtained by 24 hours a day. , different from Other sections of the road The carbon emission sequence of motor vehicles on the road network is obtained every 24 hours ;

[0082] S3.2. Create a 24×24 matrix T and replace the Euclidean distance of the dynamic time warping algorithm with the square of the difference between the estimated derivatives of the two sequences obtained in step S3.1 to obtain The first section Carbon emissions during the period The first section Distance of time period The calculation formula is:

[0083] ;

[0084] S3.3. The cumulative distance for a regular path is the sum of the distance of the current point and the distance of the smallest neighboring element that can reach this point. The calculation formula is:

[0085] ;

[0086] in, For the The first section Carbon emissions during the period The first section Cumulative similarity of carbon emissions over time periods;

[0087] S3.4. Perform iterative calculations based on steps S3.1-S3.3, starting from the first road segment, respectively calculate the cumulative similarity between the first road segment and all other road segments, and assume that the cumulative similarity value is 0.01, that is, all road segments with a cumulative similarity value less than or equal to 0.01 with the first road segment are considered as a group. For those that are not grouped together with the first road segment, repeat the calculation of the cumulative similarity between them, and ensure that all road segments are grouped into corresponding groups;

[0088] S4. Based on the grouped urban road network motor vehicle driving carbon emission data obtained in step S3, a road network motor vehicle driving carbon emission prediction model is constructed using a long short-term memory neural network to predict the road network motor vehicle driving carbon emission.

[0089] Furthermore, the specific implementation method of step S4 using the improved long short-term memory neural network considering the association between adjacent road sections to construct a road network motor vehicle driving carbon emission prediction model includes the following steps:

[0090] S4.1. Perform input calculation: Calculate the linear transformation result of the input data and the hidden state of the previous time step, multiply the grouped urban road network motor vehicle driving carbon emission data obtained in step S3 by the corresponding learnable weight matrix, and then add the bias parameter to obtain the linear transformation result of the input data. The calculation formula is as follows:

[0091] ;

[0092] in, is the output of the input gate, is the hidden state at the previous moment, are the learnable weights of the input, is the input bias parameter, After grouping The period is Carbon emissions from driving on each road segment;

[0093] S4.2. Calculate candidate memory cells using the following formula:

[0094] ;

[0095] in, is the output of the candidate memory unit, tanh is the hyperbolic tangent activation function, is the hidden state at the previous moment, are candidate learnable weights, is a candidate bias parameter;

[0096] Furthermore, the carbon emission data of each road section per hour in the hidden state of the previous time step is compared with the corresponding weight matrix Multiply and add the bias term to get the linear transformation result of the input data. Perform a nonlinear transformation (hyperbolic tangent activation function). Here, the linear transformation results of the input data and the hidden state are added, and then input into the hyperbolic tangent activation function for nonlinear transformation. The output range of the hyperbolic tangent function is between -1 and 1, which represents the value range of the candidate memory unit. The output of the hyperbolic tangent activation function is used as the output value of the candidate memory unit. The output value of the candidate memory unit represents the candidate memory content of the carbon emissions of each road section per hour in the input data of the current time step, which will be used to update the information in the memory unit.

[0097] S4.3. Perform forgetting calculation and calculate the output value of the forgetting gate. The calculation formula is as follows:

[0098] ;

[0099] in, is the output of the forget gate, is the sigmoid activation function, are the forgettable learnable weights; is the forgotten bias parameter;

[0100] Furthermore, the carbon emission data of each road section per hour in the current time step is multiplied by the corresponding weight matrix, and the bias parameter is added to obtain the linear transformation result of the input data. The hidden state of the previous time step is multiplied by the corresponding weight matrix, and the bias term is added to obtain the linear transformation result of the hidden state. The linear transformation results of the input data and the hidden state are added, and then input into the sigmoid activation function for nonlinear transformation. The output range of the Sigmoid function is between 0 and 1, indicating the importance of each input. The output of the sigmoid activation function is used as the output value of the forget gate. The output value of the forget gate indicates which parts of the carbon emissions of each road section per hour in the input data of the current time step need to be forgotten and will not be updated to the memory unit.

[0101] S4.4. Update the memory unit according to the forgetting calculation, calculate the forgetting degree of the memory unit by the forget gate, and use the output of the forget gate Perform element-level multiplication on the memory unit of the previous time step. The calculation formula is: forget gate output × previous time step memory unit; calculate the input gate For candidate memory cells To determine the update degree, use the input gate to perform element-level multiplication operation on the candidate memory unit of the current time step. The calculation formula is: input gate output × current time step candidate memory unit; update memory unit , add the product of the forget gate output and the input gate output to update the memory unit. The comprehensive calculation formula is:

[0102] ;

[0103] S4.5. Update the hidden state and calculate the output degree of the output gate to the memory unit. Use the output gate Perform element-level multiplication on the memory unit of the current time step, and the calculation formula is:

[0104] ;

[0105] in, is the hidden state at the current moment, The output of the output gate;

[0106] Furthermore, the output degree of the output gate to the memory unit is calculated. Perform an element-wise multiplication operation on the memory cell at the current time step to determine which information output is read from the memory cell. Multiply the output of the output gate by the value of the memory cell at the current time step after the hyperbolic tangent function: This step ensures that the value of the hidden state is between -1 and 1 while preserving the information in the memory cell.

[0107] S4.6. Perform output calculation, the calculation formula is:

[0108] ;

[0109] in, are the learnable weights of the output;

[0110] Furthermore, the carbon emission data of each section per hour in the current time step is multiplied by the corresponding weight matrix, and then the bias term is added to obtain the linear transformation result of the input data. The hidden state of the previous time step is multiplied by the corresponding weight matrix, and then the bias term is added to obtain the linear transformation result of the hidden state. A nonlinear transformation (sigmoid activation function) is performed, the linear transformation results of the input data and the hidden state are added, and then input into the sigmoid activation function for nonlinear transformation. The output range of the Sigmoid function is between 0 and 1, indicating the importance of each input. The output value of the output gate is calculated, and the output of the sigmoid activation function is used as the output value of the output gate. The output value of the output gate indicates which information is read from the memory unit and the output amount.

[0111] S4.7. Repeat steps S4.1 to S4.6 until all time steps have been processed. Calculate the value of the loss function based on the model output and the true label;

[0112] S4.8. Update parameters and repeat steps S4.1 to S4.7 until the predetermined number of training rounds or convergence condition is reached. Use the trained model to predict new input data to obtain future prediction values.

[0113] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0114] Although the present application has been described above with reference to specific embodiments, various modifications may be made thereto and parts thereof may be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for predicting carbon emissions of motor vehicles on a road network, characterized in that: The steps include: S1. Taking the urban road network as the research object, it is divided into sections with intersections as boundaries, and the road network data of each section is collected to build a road network data model; S2. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types, build a carbon emission calculation model for motor vehicles on the urban road network, and calculate carbon emission data for motor vehicles on the urban road network; S3. Based on the carbon emission data of motor vehicles on the urban road network calculated in step S2, a road network motor vehicle carbon emission grouping model is constructed to obtain grouped carbon emission data of motor vehicles on the urban road network; The specific implementation method of step S3 includes the following steps: S3.

1. Considering that the carbon emission of motor vehicles on the road network is time series data, a time series curve is constructed for each research unit section with an interval of 1 hour. For the kth section, the carbon emission sequence of motor vehicles on the road network is obtained for 24 hours every day {TE 1,k ,TE 2,k …TE 24,k }, different from the kth road segment, the other road segments l are used every 24 hours to obtain the road network vehicle driving carbon emission sequence {TE 1,l ,TE 2,l …TE 24,l }; S3.

2. Establish a 24×24 matrix T. Replace the Euclidean distance of the dynamic time warping algorithm with the square of the estimated derivative difference between the two sequences obtained in step S3.

1. The calculation formula for the distance ED(h,e) between the carbon emissions of the kth section in the hth period and the lth section in the eth period is: S3.

3. The cumulative distance for a regular path is the sum of the distance of the current point and the distance of the smallest neighboring element that can reach this point. The calculation formula is: r(h,e)=ED(h,e)+min(r(h-1,e-1),r(h-1,e),r(h,e-1)) Among them, r(h,e) is the cumulative similarity between the carbon emissions of the kth section in the hth period and the carbon emissions of the lth section in the eth period; S3.

4. Perform iterative calculation based on steps S3.1-S3.3, starting from the first road segment, respectively calculate the cumulative similarity between the first road segment and all other road segments, and assume that the cumulative similarity value is 0.01, that is, all road segments with a cumulative similarity value less than or equal to 0.01 with the first road segment are considered as a group. For those that are not grouped with the first road segment, repeat the calculation of the cumulative similarity between each other, and ensure that all road segments are grouped into corresponding groups; S4. Based on the grouped urban road network motor vehicle driving carbon emission data obtained in step S3, a road network motor vehicle driving carbon emission prediction model is constructed using a long short-term memory neural network to predict the road network motor vehicle driving carbon emission.

2. A method for predicting carbon emissions of motor vehicles on a road network according to claim 1, characterized in that: The specific implementation method of step S1 to construct a road network data model is to take the urban road network as the research object and divide it into road segment links with intersections as boundaries. The research object is the links including all urban highways, main roads, secondary roads and some branches. The fields of each link include the road segment number, the previous road segment number, the next road segment number, the road type, the length, the speed limit value, the direction and the name.

3. A method for predicting carbon emissions of motor vehicles on a road network according to claim 2, characterized in that: In step S1, the length of the road section is divided into sections not exceeding 500 meters.

4. A method for predicting carbon emissions of motor vehicles on a road network according to claim 3, characterized in that: The specific implementation method of step S2 includes the following steps: S2.

1. Collect real-time dynamic traffic information data on road traffic flow, vehicle driving conditions, and vehicle types; S2.

2. Based on the real-time dynamic traffic information data collected in step S2.1, calculate the carbon emission TE of the i-th type of vehicle traveling on the k-th road section in the h-th period h,k,i , the calculation formula is: TEA h,k,i =q h,k,i (g)×EF i (g)×L k Among them, q h,k,i (g) represents the number of vehicles of the i-th type traveling on the k-th road section in the h-th period under the working condition g, EF i (g) represents the emission coefficient of the i-th type of vehicle under the operating condition g, L k represents the length of the kth road segment; S2.

3. Calculate the carbon emission TE for driving on the kth road section during the hth period h,k , the calculation formula is: Real-time dynamic traffic information such as road section traffic flow, vehicle driving conditions, and vehicle types are extracted, and the emissions of the road section are dynamically estimated through a bottom-up approach.

5. A method for predicting carbon emissions of motor vehicles on a road network according to claim 4, characterized in that: In step S2, the operating condition g is composed of three parameters: average vehicle speed, relative positive acceleration and parking time ratio, and the three parameters have the same weight.

6. A method for predicting carbon emissions of motor vehicles on a road network according to claim 5, characterized in that: The specific implementation method of step S4 of using the improved long short-term memory neural network considering the association between adjacent road sections to construct a road network motor vehicle driving carbon emission prediction model includes the following steps: S4.

1. Perform input calculation: Calculate the linear transformation result of the input data and the hidden state of the previous time step, multiply the grouped urban road network motor vehicle driving carbon emission data obtained in step S3 by the corresponding learnable weight matrix, and then add the bias parameter to obtain the linear transformation result of the input data. The calculation formula is as follows: in t =σ(W i ·[d t-1 ,THE′ h,k ]+b i ) Among them, i t is the output of the input gate, d t-1 is the hidden state at the previous moment, W i is the learnable weight of the input, b i is the input bias parameter, TE h ' ,k is the carbon emission of driving on the kth road section during the hth period after grouping; S4.

2. Calculate the candidate memory cells. The calculation formula is as follows: in, is the output of the candidate memory unit, tanh is the hyperbolic tangent activation function, d t-1 is the hidden state at the previous moment, W C is the candidate learnable weight, b C is a candidate bias parameter; S4.

3. Perform forgetting calculation and calculate the output value of the forgetting gate. The calculation formula is as follows: f t =σ(W f ·[d t-1 ,THE′ h,k ]+b f ) Among them, f t is the output of the forget gate, σ is the sigmoid activation function, W f is the forgotten learnable weight; b f is the forgotten bias parameter; S4.

4. Update the memory unit according to the forgetting calculation, calculate the forgetting degree of the memory unit by the forget gate, and use the output f of the forget gate t Perform element-level multiplication on the memory unit of the previous time step. The calculation formula is: forget gate output × previous time step memory unit; calculate input gate i t For candidate memory cells To determine the update degree, use the input gate to perform element-level multiplication operation on the candidate memory unit of the current time step. The calculation formula is: input gate output × current time step candidate memory unit; update memory unit C t , add the product of the forget gate output and the input gate output to update the memory unit. The comprehensive calculation formula is: S4.

5. Update the hidden state and calculate the output degree of the output gate to the memory unit. Use the output gate o t Perform element-level multiplication on the memory unit of the current time step, and the calculation formula is: h t =o t *tanh(C t ) Among them, h t is the hidden state at the current moment, o t The output of the output gate; S4.

6. Perform output calculation, the calculation formula is: a t =σ(W o ·[d t-1 ,TE′ h,k ] Among them, W o are the learnable weights of the output; S4.

7. Repeat steps S4.1 to S4.6 until all time steps have been processed. Calculate the value of the loss function based on the model output and the true label; S4.

8. Parameter update, repeat steps S4.1 to S4.7 until the predetermined number of training rounds or convergence condition is reached. Use the trained model to predict new input data to obtain future prediction values.

Images