Digital twin auxiliary intelligent network connection vehicle carbon emission prediction system and method

Abstract

The invention relates to the technical field of vehicle carbon emission prediction, and particularly provides a digital twinning-assisted intelligent network-connected vehicle carbon emission prediction system and method, which combines digital twinning and deep learning, applies the combined digital twinning and deep learning to a carbon emission prediction link, and simulates various driving behaviors to obtain a carbon emission prediction result. And finally, giving an optimal carbon emission reduction mode based on the current driving state according to comparison, and feeding back the optimal carbon emission reduction mode to a user. Meanwhile, the whole prediction part can continuously compare a prediction result with an actual execution result to optimize the whole structure, so that the prediction accuracy and generalization are improved, and the user requirements are met.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle carbon emission prediction, and specifically proposes a digital twin-assisted intelligent connected vehicle carbon emission prediction system and method. Background Art

[0002] The assessment and optimization of carbon emissions of intelligent connected vehicles refers to obtaining the vehicle's carbon emission data by analyzing the vehicle's current driving status, and simulating driving behavior through digital twin technology based on the assessment results to obtain various carbon emission reduction plans, so that the target can take corresponding measures in a timely manner to reduce the vehicle's carbon emissions.

[0003] Digital twin technology is to build a virtual digital model in the virtual world that can reflect the various characteristics of real physical entities. Digital twin technology has now gradually penetrated into various fields, including vehicle carbon emission prediction.

[0004] At present, the carbon emission predictions used by vehicles are mostly based on mileage, fuel consumption and statistical models. The mileage-based method calculates the total carbon emissions based on the vehicle's mileage and the average carbon emission factor per unit mileage. However, the average carbon emission factor cannot accurately reflect the impact of factors such as different vehicle models, vehicle conditions, driving habits and road conditions, which will lead to inaccurate prediction results. The method based on fuel consumption measures or estimates the vehicle's fuel consumption and calculates carbon emissions in combination with the fuel's carbon emission coefficient. The measurement or estimation of fuel consumption may have errors, and does not take into account the changes in the vehicle's fuel efficiency under different working conditions. The method based on statistical models uses a large amount of historical data to establish a statistical model, such as a regression analysis model, to predict the vehicle's carbon emissions. There are problems such as insufficient model integrity and susceptibility to environmental factors, which will make the reliability and accuracy of vehicle carbon emission data not guaranteed. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method for carbon emission assessment and optimization of intelligent networked vehicles combined with digital twin technology, by combining digital twin with deep learning and applying it to the prediction link of carbon emissions, simulating various driving behaviors to obtain the prediction results of carbon emissions, and finally giving the optimal carbon emission reduction method based on the current driving state according to the comparison and feeding it back to the user. At the same time, the entire prediction part can continuously compare the prediction results with the actual execution results to optimize the overall structure, improve the accuracy and generalization of the prediction, and meet user needs.

[0006] In order to solve the above technical problems, the present invention provides a digital twin-assisted intelligent networked vehicle carbon emission prediction system, the system comprising: Data acquisition module: collects and records relevant data during the vehicle operation process through data sensors located on various devices and components of the vehicle (drive motor, air conditioning system, battery management system, on-board electronic equipment, power steering system and braking system, etc.), that is, records environmental parameters and the energy consumption of vehicle components at the current speed under a given external environment to build a vehicle operation sensor data set.

[0007] Digital twin module: Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set obtained by the data acquisition module, the operating parameters of other components in the vehicle's operation (components whose energy consumption is not related to the vehicle's operating status), and some static information of the vehicle (vehicle design parameters, technical specifications provided by the manufacturer, etc.), and place it on the on-board edge device. At the same time, update the relevant parameters of the digital twin model based on the vehicle component information and environmental information collected in real time by the on-board edge device to ensure that the digital twin model can reflect the real-time status of the vehicle.

[0008] Evaluation module: The total carbon emissions of each vehicle component are calculated in real time based on the energy consumption of each vehicle component collected in the data acquisition module, and compared with the carbon emissions exceeding the limit to determine whether the carbon emissions exceed the limit. If the emissions exceed the limit, the prediction module will obtain the optimal carbon emission reduction driving plan.

[0009] Prediction module: Establish a dynamic prediction model of carbon emissions based on CNN, import the vehicle's operating sensor data set for learning, so that the dynamic prediction model can output the prediction results of the energy consumption of vehicle components (drive motor, air conditioning system, battery management system, etc.) at later moments when the vehicle speed and environmental parameters are input. The dynamic prediction model of carbon emissions is combined with the digital twin model to simulate various driving behaviors. Specifically, the digital twin model completes the simulation and simulation of driving behavior based on the current moment, obtains the vehicle operation status at later moments, and then combines the dynamic prediction model of carbon emissions to realize the prediction based on the simulated vehicle condition, and finally obtains the expected power consumption of each component of the vehicle at later moments, and then obtains the predicted carbon emissions of the vehicle under various simulated driving behaviors, and thus screens out the solution with the lowest carbon emissions.

[0010] Visualization module: The on-board screen dynamically displays the carbon emissions of each component collected by the sensor, giving users intuitive vehicle carbon emission information. When the evaluation and optimization module collects the best carbon emission reduction plan, the plan will be fed back to the user through the screen.

[0011] On the other hand, the present invention provides a digital twin-assisted intelligent connected vehicle carbon emission prediction method, the method comprising: S1. Record the information collected by the sensors installed on the vehicle; install sensors on each major operating component of the intelligent connected vehicle to detect the specific energy consumption of the components and collect relevant environmental parameters; S2, forming a vehicle operation sensor data set based on the data collected in step S1 during the vehicle operation process, that is, the energy consumption of components of the vehicle at a certain speed under a given external environment; S3. Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set, the real-time working status of vehicle components collected by sensors, and the static information of the vehicle; load the built vehicle digital twin model on the on-board edge device to process the information collected by the sensors, and update the relevant parameters of the digital twin model based on the real-time collected information of each component and environmental information, so that the digital twin model can reflect the real-time vehicle status; S4, based on the vehicle operation sensor data set constructed in S1, the energy consumption of each component of the vehicle under various conditions during operation is obtained, and a dynamic prediction model of carbon emissions is established using a convolutional neural network (CNN); S5. The vehicle-mounted edge device evaluates the carbon emission information obtained by the vehicle sensor. If the overall carbon emission of the vehicle exceeds the standard, the digital twin model will be combined with the carbon emission prediction model to simulate the driving behavior and enter step S6. If it does not exceed the standard, the vehicle sensor will continue to collect information to improve the energy consumption of various vehicle components under various driving environments, and import the improved data set into CNN for training, and enter step S8. S6. First, the digital twin model constructed in step S3 completes the simulation of a certain driving behavior based on the vehicle state at the current time t, obtains the vehicle operation state at the next time t+1 and the parameters a and b, and then the convolutional neural network constructed in step S3 realizes the energy consumption prediction of each component of the vehicle at time t+1, and then obtains the predicted carbon emissions when the vehicle adopts this driving behavior scheme; the above process is repeated multiple times to obtain the carbon emission data obtained by simulating multiple driving behavior schemes for screening, and selects a reasonable and optimal scheme under the condition that the emission does not exceed the standard; S7, screening the carbon emission data obtained by simulating various driving behaviors, selecting a reasonable and optimal driving behavior plan under the condition that the emission does not exceed the standard, and proceeding to step S8; S8. Collect information for visualization. Visualize the carbon emission data collected by the sensors on the vehicle, feed back the energy consumption of each component to the user through the on-board screen, complete dynamic display, and let the user understand the overall carbon emission of the vehicle; when the optimal carbon emission reduction driving behavior plan is obtained in step S7, this plan will be fed back to the user through the on-board screen for the user to choose to execute.

[0012] Furthermore, in step S1, based on the obtained energy consumption of vehicle components and the electric energy carbon emission factor, the two are multiplied to obtain the carbon emissions of the vehicle.

[0013] Further, in step S2, the sensor data set is operated using triples<a,b,c> The parameter a represents the real-time speed of the vehicle collected by the sensor; the parameter b represents the environmental parameters of the vehicle collected by the sensor, which consists of a three-digit group. The first digit in the group represents the weather environment outside the vehicle, the second digit represents the current driving state of the vehicle, and the third digit represents the current road conditions, including the degree of uphill slope, the degree of downhill slope, flat road surface without slope, and bumpy road surface; the parameter c represents the energy consumption corresponding to the vehicle components at this time, which consists of a two-digit group. The first digit in the group represents the number of a certain component of the vehicle, and the second digit represents the power consumption of the corresponding component at this time.

[0014] Furthermore, in step S4, the process of establishing the dynamic prediction model of carbon emissions based on CNN is as follows: the operation sensor data set established in step S2 is imported into CNN for training, and the input is parameter a and parameter b in the vehicle operation data set, that is, the current speed of the vehicle and the environmental parameters at this time are input, and the output is parameter c in the operation data set, that is, the power consumption of the corresponding component when working normally; when the vehicle speed and environmental parameters are input, the trained CNN dynamic prediction model of carbon emissions can output the prediction results of the energy consumption that the vehicle components will generate at the next moment; at the same time, the vehicle operation sensor data set is continuously supplemented and imported into the CNN dynamic prediction model of carbon emissions for training to ensure the accuracy of the prediction results.

[0015] Furthermore, in step S6, the process of predicting the energy consumption of driving behavior is as follows: first, the on-board navigation determines the currently feasible routes and corresponding road conditions, and records them as route 1, route 2, ..., route n, where n≥1 and is an integer; then the direction parameter A, the speed parameter B, and the on-board equipment parameter C are set to obtain a series of driving behaviors; wherein A is a driving behavior related to the vehicle's driving direction, such as turning left, turning right, and going straight; wherein B is a driving behavior related to the vehicle's driving speed, including acceleration, deceleration, and maintaining the current speed; wherein C is a driving behavior related to the on-board equipment; thereby, a series of driving behavior adjustment strategies composed of a certain route and the three driving behaviors A, B, and C are obtained, and then the driving strategies are predicted by the digital twin model and the carbon emission dynamic prediction model to obtain the corresponding carbon emission reduction schemes and their carbon emissions, namely scheme 1, scheme 2, ..., scheme m, where m≥1, and m is an integer.

[0016] Furthermore, in step S7, the reasonable and optimal judgment of the driving behavior plan includes the following screening process. The first step is to screen for removing plans with high carbon emissions, that is, the dynamic prediction module of carbon emissions excludes plans whose carbon emissions still exceed the standard, and obtains plan 1, plan 2, ..., plan v, where m≥v≥1, and v is an integer; the second step is to screen based on the rationality of the plans, exclude plans with too low speed and too high speed, and plans that conflict with the current vehicle driving environment, and obtain plan 1, plan 2, ..., plan k, where m≥v≥k≥1, and k is an integer; the third step of screening is to sort according to the standard of the lowest carbon emissions, select the plan with the lowest carbon emissions among the remaining plans after the above two steps of screening, and obtain the optimal driving behavior plan.

[0017] Furthermore, in step S8, when the user chooses to execute a carbon emission reduction driving behavior plan, the digital twin model and the carbon emission prediction model compare the simulated vehicle condition and carbon emission data with the data obtained in actual execution, and continuously optimize and adjust the relevant parameters of the model according to the error.

[0018] Compared with the prior art, the digital twin-assisted intelligent connected vehicle carbon emission prediction system and method provided by the present invention has the following beneficial effects: (1) The digital twin-assisted intelligent connected vehicle carbon emission prediction system and method provided in this application constructs a digital twin model of the target vehicle by processing and analyzing the sensor data set of vehicle operation and some specification information of the vehicle itself, and updates the relevant parameters of the operating sensor data set and the digital twin model based on the real-time collected component information and environmental information. At the same time, combined with the constructed vehicle carbon emission prediction model, it can predict the carbon emissions of the vehicle under specific driving behavior, and screen out the method with the lowest carbon emissions and feedback to the user, thereby achieving carbon emission optimization.

[0019] (2) When the user chooses to execute a driving behavior plan, the digital twin model and the carbon emission prediction model can compare the simulated data with the actual execution data, and continuously optimize and adjust the relevant parameters of the model based on the error, thereby improving the accuracy and reliability of the model simulation.

[0020] (3) The continuous updating of the operating sensor data set and the combination of the digital twin model and the carbon emission prediction model solve the problem of low accuracy in the existing technology due to insufficient equipment completeness and susceptibility to external factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a schematic diagram of the system structure of the digital twin-assisted intelligent connected vehicle carbon emissions in an embodiment of the present application.

[0022] Figure 2 This is a schematic flow chart of the method for digital twin-assisted carbon emissions of intelligent connected vehicles in an embodiment of the present application.

[0023] Figure 3 Schematic diagram of the structure of the driving behavior prediction link in the embodiment of the present application. DETAILED DESCRIPTION

[0024] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.

[0025] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. In addition, it should be understood that the specific implementation methods described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0026] Reference Figure 1 This embodiment provides a digital twin-assisted intelligent connected vehicle carbon emission prediction system, which specifically includes: Data acquisition module: collects and records relevant data during the vehicle operation process through data sensors located on various devices and components of the vehicle (drive motor, air conditioning system, battery management system, on-board electronic equipment, power steering system and braking system, etc.), that is, records environmental parameters and the energy consumption of vehicle components at the current speed under a given external environment to build a vehicle operation sensor data set.

[0027] Digital twin module: Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set obtained by the data acquisition module, the operating parameters of other components in the vehicle's operation (components whose energy consumption is not related to the vehicle's operating status), and some static information of the vehicle (vehicle design parameters, technical specifications provided by the manufacturer, etc.), and place it on the on-board edge device. At the same time, update the relevant parameters of the digital twin model based on the vehicle component information and environmental information collected in real time by the on-board edge device to ensure that the digital twin model can reflect the real-time status of the vehicle.

[0028] Evaluation module: The total carbon emissions of each vehicle component are calculated in real time based on the energy consumption of each vehicle component collected in the data acquisition module, and compared with the carbon emissions exceeding the limit to determine whether the carbon emissions exceed the limit. If the emissions exceed the limit, the prediction module will obtain the optimal carbon emission reduction driving plan.

[0029] Prediction module: Establish a dynamic prediction model of carbon emissions based on CNN, import the vehicle's operating sensor data set for learning, so that the dynamic prediction model can output the prediction results of the energy consumption of vehicle components (drive motor, air conditioning system, battery management system, etc.) at later moments when the vehicle speed and environmental parameters are input. The dynamic prediction model of carbon emissions is combined with the digital twin model to simulate various driving behaviors. Specifically, the digital twin model completes the simulation and simulation of driving behavior based on the current moment, obtains the vehicle operation status at later moments, and then combines the dynamic prediction model of carbon emissions to realize the prediction based on the simulated vehicle condition, and finally obtains the expected power consumption of each component of the vehicle at later moments, and then obtains the predicted carbon emissions of the vehicle under various simulated driving behaviors, and thus screens out the solution with the lowest carbon emissions.

[0030] Visualization module: The on-board screen dynamically displays the carbon emissions of each component collected by the sensor, giving users intuitive vehicle carbon emission information. When the evaluation and optimization module collects the best carbon emission reduction plan, the plan will be fed back to the user through the screen.

[0031] like Figure 2 As shown, this embodiment provides a digital twin-assisted intelligent connected vehicle carbon emission prediction method, the method comprising: S1. Record the information collected by the sensors installed on the vehicle. There are multiple sensors on the intelligent connected vehicle. The sensors are located on the main components of the vehicle (drive motor, air conditioning system, battery management system, on-board electronic equipment, power steering system and braking system, etc.), mainly to detect its specific energy consumption and collect relevant environmental parameters.

[0032] The energy consumption of the vehicle is obtained in order to obtain the carbon emissions of the vehicle. The carbon emissions obtained based on the energy consumption of the components are determined in combination with the local electric energy carbon emission factor, which is obtained by multiplying the energy consumption of the corresponding components of the vehicle by the local electric energy carbon emission factor. For example, if the local carbon emission factor is 0.5568kgCO2 / kWh, and the energy consumption of the vehicle's drive system collected by the sensor is 2.5kwh, then the amount of carbon dioxide emitted by the consumption of one kilowatt-hour of electricity is 2.5×0.5568=1.392kg.

[0033] S2, according to the relevant data collected in step S1 during the vehicle operation process, that is, the energy consumption of the components of the vehicle at a certain speed under a given external environment, through the triple<a,b,c> This form is used to represent the data collected by the sensor and constitute the vehicle's operating sensor data set. The data set is used to participate in the construction of the digital twin model and provide data support for the establishment of the subsequent dynamic prediction model of carbon emissions.

[0034] The data collected by various sensors during the operation of the vehicle do not include components with little relevance to the vehicle's operating status (such as lights, on-board screens, etc.) in the operating sensor data set. Components related to the vehicle's operating status (such as the drive system, power steering system, etc.) are included in the data set, that is, the energy consumption of the components of the vehicle at a certain speed under a given external environment is recorded.<a,b,c> This form is used to record and constitute the vehicle's operating sensor data set.

[0035] The parameter a in the running sensor data set represents the real-time speed of the vehicle collected by the sensor, such as 17, which means that the current speed of the vehicle is 17 km / h, and the data unit is km / h.

[0036] The parameter b in the running sensor data set represents the environmental parameters of the vehicle collected by the sensor, which is composed of a three-digit group, such as 0-1-3, which means that the vehicle turns right on a downhill section with a slope of 3 in a sunny environment. The first digit in the digital group represents the environment outside the vehicle, such as "0" for sunny days, "1" for cloudy days, "2" for rainy days, and "3" for foggy days. The second digit represents the current vehicle state, such as "0" for turning left, "1" for turning right, "2" for accelerating, and "3" for decelerating. The third digit represents the current degree of uphill or downhill, and the slope analysis is performed through the pressure sensor on the vehicle tire, ranging from 0 to 19. The slope of the downhill section is 1 to 9; the slope of the uphill section is 11 to 19; the two are divided starting from "10", "10" means that the road surface is flat and has no slope; "0" means that the road surface is bumpy, that is, when the pressure sensor changes too fast and the fluctuation range is large, the third digit is recorded as "0".

[0037] The parameter c in the running sensor data set represents the energy consumption of the vehicle components (such as the drive motor, power steering system, etc.) at this time, and is composed of a two-digit group, such as 0-11, which means that the energy consumption of the vehicle's drive motor at this time is 11kwh. The first digit in the digital group is a component of the electric vehicle, such as "0" represents the drive motor of the electric vehicle, and "1" represents the battery management system of the vehicle. The second digit is the power consumption of the corresponding component at this time, in kwh. When this component is not running, stop collecting other information related to this component.

[0038] S3. Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set, the real-time working status of vehicle components collected by sensors, and some static information of the vehicle (design parameters of the vehicle, technical specifications provided by the manufacturer, etc.). The built vehicle digital twin model is placed on the vehicle edge device to process a series of information collected by the sensor, and the relevant parameters of the digital twin model are updated according to the real-time collected information of each component and environmental information, so that the digital twin model can reflect the real-time vehicle status.

[0039] The digital twin model refers to a simulation process that makes full use of physical models, sensor updates, operation history and other data, integrates multi-disciplinary, multi-physical quantity, multi-scale, and multi-probability, completes mapping in virtual space, and thus reflects the various characteristics and states of the corresponding entities.

[0040] The vehicle-mounted edge device in this system can quickly process the data collected by the vehicle's own sensors (such as lidar, cameras, etc.). These data include information about the vehicle's surrounding environment and the vehicle's own status. By placing the constructed vehicle digital twin model on the vehicle-mounted edge device, these real-time data can be directly used to update and calculate the model, avoiding the delay problem caused by data transmission to the remote server, and ensuring a high degree of synchronization between the digital twin model and the actual status of the vehicle.

[0041] The parameters of the digital twin model are updated based on the collected information. Since the vehicle may encounter different operating environments during driving, such as bumpy or smooth roads, it is necessary to update the parameters of the digital twin model based on the collected data.<a,b,c> Update the relevant parameters of the digital twin model in real time, so that the subsequent digital twin model can obtain more accurate component working conditions when simulating a certain driving behavior, and thus make the predicted carbon emissions more accurate. Because the road conditions in different places are not exactly the same, and the energy consumption of components under the same road conditions may also be different, it is necessary to continuously supplement the vehicle's operating sensor data set to reduce the impact of environmental factors when predicting carbon emissions in the future.

[0042] S4. According to the vehicle operation sensor data set constructed in S1, the energy consumption of each component of the vehicle under various conditions during operation (such as when the vehicle is at a speed of 30km / h in a sunny environment) can be obtained, and thus a dynamic prediction model of carbon emissions can be established using a convolutional neural network (CNN). The dynamic prediction model of carbon emissions of a CNN is established, and the operation sensor data set established in step S2 is imported into CNN for training, and the input is the parameters a and b in the vehicle operation data set, that is, the current speed of the vehicle and the environmental parameters at this time, and the output is the parameter c in the operation data set, that is, the power consumption of the corresponding component (during normal operation). When the vehicle speed and environmental parameters are input, the trained CNN dynamic prediction model of carbon emissions can output the prediction results of the energy consumption of vehicle components (drive motor, air conditioning system, battery management system, etc.) at the next moment. At the same time, it is necessary to continuously supplement the vehicle's operation sensor data set and import the CNN dynamic prediction model of carbon emissions for training to ensure the accuracy of the prediction results.

[0043] S5. The on-board edge device conducts an assessment based on the carbon emission information obtained by the vehicle sensors. If the vehicle's overall carbon emissions exceed the standard, the digital twin model will be combined with the carbon emission prediction model to simulate the driving behavior. If it does not exceed the standard, the vehicle sensors will continue to collect information to improve the energy consumption of various vehicle components under various driving environments, and import the improved data set into CNN for training to improve the generalization and accuracy of the CNN prediction model.

[0044] The carbon emission information is assessed by adding up the carbon emissions of the entire vehicle population and comparing it with the carbon emission excess limit for such vehicles.

[0045] S6. The on-board edge device calculates the overall carbon emissions of the vehicle in real time. When the carbon emissions exceed the standard, the dynamic prediction model of carbon emissions will be combined with the digital twin model. First, the digital twin model constructed in step S3 completes the simulation of a driving behavior based on the vehicle state at the current time t, obtains the vehicle operation state and parameters a and b at the next time t+1, and then the convolutional neural network constructed in step S3 realizes the energy consumption prediction of each component of the vehicle at time t+1, and then obtains the predicted carbon emissions when the vehicle adopts a driving behavior plan. Finally, the carbon emission data obtained by simulating multiple driving behavior plans are screened to select a reasonable and optimal plan under the condition that the emissions do not exceed the standard.

[0046] When obtaining the predicted carbon emissions, some components related to the vehicle's driving conditions are predicted through CNN, while other components that are weakly related to the vehicle's operating conditions do not need to be predicted through CNN, such as the lighting inside the vehicle. The characteristics of such equipment are that their energy consumption is basically at a fixed value after being turned on, and is not affected by the external environment such as vehicle speed. Therefore, when making a prediction, if such equipment is in the turned-on state, the predicted energy consumption is the energy consumption when it is turned on.

[0047] Furthermore, the predicted driving behavior described in S6 is as follows: Figure 3As shown, the vehicle navigation first determines the currently feasible route and the corresponding road conditions, and records them as "route 1, route 2, ..., route n" (where n ≥ 1 and is an integer); then a series of driving behaviors are obtained by permuting and combining A (direction), B (speed), and C (vehicle-mounted equipment), wherein A is a driving behavior related to the vehicle's driving direction, such as turning left, turning right, and going straight, etc.; wherein B is a driving behavior related to the vehicle's driving speed, such as accelerating, decelerating, maintaining the current speed, etc., and the acceleration and deceleration amplitude is 5km / h. For example, if the current speed is 40km / h, an acceleration strategy is adopted during the simulation of driving behavior, and then acceleration solutions such as accelerating the vehicle speed to 45km / h, 50km / h, and 55km / h can be obtained; wherein C is a driving behavior related to vehicle-mounted equipment (vehicle lighting, air conditioning, etc.), such as turning off vehicle-mounted equipment 1, turning off vehicle-mounted equipment 2, etc. In this way, we can obtain a series of driving behavior adjustment strategies consisting of a certain route and three driving behaviors A, B, and C. Then, we can predict the driving strategy through the digital twin model and the carbon emission dynamic prediction model to obtain the corresponding carbon emission reduction plan and its carbon emissions, namely Plan 1, Plan 2,..., Plan m (m≥1, m is an integer).

[0048] S7. Filter the carbon emission data obtained by simulating various driving behaviors, select the reasonable and optimal way to reduce the emission without exceeding the standard, and feedback it to the user on the in-vehicle screen through the next visualization module.

[0049] Furthermore, the reasonable and optimal judgment in S7 includes three aspects. The purpose of the first step of screening is to remove the scheme with high carbon emissions, that is, the dynamic prediction module of carbon emissions excludes the scheme with carbon emissions that still exceed the standard, and obtains scheme 1, scheme 2, ..., scheme v (m≥v≥1, v is an integer). The second step of screening is carried out according to the rationality of the scheme, that is, the dynamic prediction module of carbon emissions excludes the scheme with too low speed (schemes with speed below 10km / h) and too high speed (schemes with speed above 100km / h), and the scheme related to the current environment (turning off the air conditioner when the temperature difference between the inside and outside of the vehicle is too high, turning off the interior lighting system when driving at night, etc.), and obtains scheme 1, scheme 2, ..., scheme k (m≥v≥k≥1, k is an integer). The third step of screening is carried out according to the standard of the lowest carbon emissions, that is, the dynamic prediction module of carbon emissions selects the scheme with the lowest carbon emissions among the remaining schemes after two steps of screening, and obtains the optimal scheme.

[0050] S8, collects information for visualization, visualizes the carbon emission data collected by the sensors on the vehicle, and feeds back the energy consumption of each component to the user through the on-board screen, completing a dynamic display, so that the user can more intuitively understand the overall carbon emission of the vehicle. And when the optimal carbon emission reduction driving behavior is obtained in S7, this solution will be fed back to the user through the on-board screen for the user to choose to execute.

[0051] When the user chooses to implement a certain plan, the digital twin model and the carbon emission prediction model compare the simulated vehicle condition and carbon emission data with the data obtained in the actual implementation, and continuously optimize and adjust the relevant parameters of the model based on the error, thereby continuously improving the accuracy and reliability of the model simulation.

[0052] The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiment. Any equivalent modifications or changes made by ordinary technicians in this field based on the contents disclosed by the present invention should be included in the protection scope recorded in the claims.

Claims

1. Digital twin-assisted intelligent connected vehicle carbon emission prediction system, characterized by: The system comprises the following parts: Data acquisition module: collects and records relevant data during the vehicle operation process through data sensors installed on various devices and components of the vehicle, records environmental parameters and the energy consumption of vehicle components at the current speed under a given external environment, and builds a vehicle operation sensor data set; Digital twin module: Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set, the vehicle's energy consumption during operation, the working parameters of components unrelated to the vehicle's operating status, and the vehicle's static information, and load it on the vehicle's edge device; Evaluation module: Calculate the total carbon emissions of each vehicle component in real time based on the energy consumption of each vehicle component collected in the data acquisition module, and compare it with the limit of carbon emissions exceeding the standard to determine whether the carbon emissions exceed the standard. If the emissions exceed the standard, the prediction module will obtain the optimal carbon emission reduction driving plan; Prediction module: Establish a dynamic prediction model of carbon emissions based on CNN, import the vehicle's operating sensor data set for learning, so that the dynamic prediction model outputs the energy consumption prediction results of vehicle components at the next moment when the vehicle speed and environmental parameters are input; Combine the energy consumption prediction results with the driving behavior simulated by the digital twin model to obtain the driving behavior plan with the lowest carbon emissions; Visualization module: The on-board screen displays the carbon emissions of each component collected by the sensor; when the evaluation module collects the optimal carbon emission reduction plan, it will be fed back to the user through the screen.

2. The digital twin-assisted intelligent connected vehicle carbon emission prediction system according to claim 1 is characterized in that: In the digital twin module, the relevant parameters of the digital twin model are updated based on the vehicle component information and environmental information collected in real time by the on-board edge device to ensure that the digital twin model can reflect the real-time status of the vehicle.

3. The digital twin-assisted intelligent connected vehicle carbon emission prediction system according to claim 1 is characterized in that: In the prediction module, the dynamic prediction model of carbon emissions is combined with the digital twin model to simulate various driving behaviors. The digital twin model completes the simulation and simulation of the driving behavior based on the current moment, and obtains the vehicle operation status at the next moment. The dynamic prediction model of carbon emissions is then combined to realize the prediction based on the simulated vehicle conditions, and finally the expected power consumption of each component of the vehicle at the next moment is obtained, and then the predicted carbon emissions of the vehicle under various simulated driving behaviors are obtained, and the solution with the lowest carbon emissions is screened out.

4. The carbon emission prediction method of the digital twin-assisted intelligent connected vehicle carbon emission prediction system according to any one of claims 1 to 3, characterized in that: The method comprises the following steps: S1. Record the information collected by the sensors installed on the vehicle; install sensors on each major operating component of the intelligent connected vehicle to detect the specific energy consumption of the components and collect relevant environmental parameters; S2, forming a vehicle operation sensor data set based on the data collected in step S1 during the vehicle operation process, that is, the energy consumption of components of the vehicle at a certain speed under a given external environment; S3. Build a digital twin model of the target vehicle based on the vehicle's operating sensor data set, the real-time working status of vehicle components collected by sensors, and the static information of the vehicle; load the built vehicle digital twin model on the on-board edge device to process the information collected by the sensors, and update the relevant parameters of the digital twin model based on the real-time collected information of each component and environmental information, so that the digital twin model can reflect the real-time vehicle status; S4, based on the vehicle operation sensor data set constructed in S1, the energy consumption of each component of the vehicle under various conditions during operation is obtained, and a dynamic prediction model of carbon emissions is established using a convolutional neural network (CNN); S5. The vehicle-mounted edge device evaluates the carbon emission information obtained by the vehicle sensor. If the overall carbon emission of the vehicle exceeds the standard, the digital twin model will be combined with the carbon emission prediction model to simulate the driving behavior and enter step S6. If it does not exceed the standard, the vehicle sensor will continue to collect information to improve the energy consumption of various vehicle components under various driving environments, and import the improved data set into CNN for training, and enter step S8. S6. First, the digital twin model constructed in step S3 completes the simulation of a certain driving behavior based on the vehicle state at the current time t, obtains the vehicle operation state at the next time t+1 and the parameters a and b, and then the convolutional neural network constructed in step S3 realizes the energy consumption prediction of each component of the vehicle at time t+1, and then obtains the predicted carbon emissions when the vehicle adopts this driving behavior scheme; the above process is repeated multiple times to obtain the carbon emission data obtained by simulating multiple driving behavior schemes for screening, and selects a reasonable and optimal scheme under the condition that the emission does not exceed the standard; S7, screening the carbon emission data obtained by simulating various driving behaviors, selecting a reasonable and optimal driving behavior plan under the condition that the emission does not exceed the standard, and proceeding to step S8; S8. Collect information for visualization. Visualize the carbon emission data collected by the sensors on the vehicle, feed back the energy consumption of each component to the user through the on-board screen, complete dynamic display, and let the user understand the overall carbon emission of the vehicle; when the optimal carbon emission reduction driving behavior plan is obtained in step S7, this plan will be fed back to the user through the on-board screen for the user to choose to execute.

5. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized in that: In step S1, based on the obtained energy consumption of vehicle components and the electric energy carbon emission factor, the two are multiplied to obtain the carbon emissions of the vehicle.

6. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized by: In step S2, the sensor data set is run using triples<a,b,c > The parameter a represents the real-time speed of the vehicle collected by the sensor; the parameter b represents the environmental parameters of the vehicle collected by the sensor, which consists of a three-digit group. The first digit in the group represents the weather environment outside the vehicle, the second digit represents the current driving state of the vehicle, and the third digit represents the current road conditions, including the degree of uphill slope, the degree of downhill slope, flat road surface without slope, and bumpy road surface; the parameter c represents the energy consumption corresponding to the vehicle components at this time, which consists of a two-digit group. The first digit in the group represents the number of a certain component of the vehicle, and the second digit represents the power consumption of the corresponding component at this time.

7. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized in that: In step S4, the process of establishing the dynamic prediction model of carbon emissions based on CNN is as follows: the operation sensor data set established in step S2 is imported into CNN for training, and the input is parameter a and parameter b in the vehicle operation data set, that is, the current speed of the vehicle and the environmental parameters at this time are input, and the output is parameter c in the operation data set, that is, the power consumption of the corresponding component when working normally; when the trained CNN dynamic prediction model of carbon emissions inputs the vehicle speed and environmental parameters, it can output the prediction results of the energy consumption that the vehicle components will generate at the next moment; at the same time, the vehicle operation sensor data set is continuously supplemented and imported into the CNN dynamic prediction model of carbon emissions for training to ensure the accuracy of the prediction results.

8. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized by: In step S6, the process of predicting the energy consumption of driving behavior is as follows: first, the on-board navigation determines the currently feasible route and the corresponding road conditions, and records them as route 1, route 2, ..., route n, where n≥1 and is an integer; then the direction parameter A, the speed parameter B, and the on-board equipment parameter C are set to obtain a series of driving behaviors; wherein A is a driving behavior related to the vehicle's driving direction, such as turning left, turning right, and going straight; wherein B is a driving behavior related to the vehicle's driving speed, including acceleration, deceleration, and maintaining the current speed; wherein C is a driving behavior related to the on-board equipment; thereby, a series of driving behavior adjustment strategies composed of a certain route and the three driving behaviors A, B, and C are obtained, and then the driving strategy is predicted by the digital twin model and the carbon emission dynamic prediction model to obtain the corresponding carbon emission reduction plan and its carbon emissions, that is, plan 1, plan 2, ..., plan m, m≥1, and m is an integer.

9. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized in that: In step S7, the reasonable and optimal judgment of the driving behavior plan includes the following screening process. The first step is to screen for removing plans with high carbon emissions, that is, the dynamic prediction module of carbon emissions eliminates plans whose carbon emissions still exceed the standard, and obtains plan 1, plan 2, ..., plan v, where m≥v≥1, and v is an integer; the second step is to screen based on the rationality of the plans, exclude plans with too low speed and too high speed, and plans that conflict with the current vehicle driving environment, and obtain plan 1, plan 2, ..., plan k, where m≥v≥k≥1, and k is an integer; the third step of screening is to sort according to the standard of the lowest carbon emissions, select the plan with the lowest carbon emissions among the remaining plans after the above two steps of screening, and obtain the optimal driving behavior plan.

10. The digital twin-assisted intelligent connected vehicle carbon emission prediction method according to claim 4 is characterized in that: In step S8, when the user chooses to execute a carbon emission reduction driving behavior plan, the digital twin model and the carbon emission prediction model compare the simulated vehicle condition and carbon emission data with the data obtained in actual execution, and continuously optimize and adjust the relevant parameters of the model based on the error.

Images