A smart traffic system environment benefit evaluation test method based on Markov chain construction of standard working condition library

Abstract

This invention discloses a method for evaluating the environmental benefits of intelligent transportation systems based on a standard operating condition library constructed using Markov chains. Specifically, it involves a method for constructing representative driving conditions for motor vehicles based on Markov chains, followed by a coupled environmental benefit evaluation using a machine learning model. The method includes a data acquisition and processing module, a vehicle speed guidance standard operating condition library module, an exhaust emission model template, and a coupled evaluation module. This method unifies evaluation standards and accurately assesses benefits, providing direction for improvements in vehicle speed guidance technology and promoting the reduction of motor vehicle exhaust emissions and the improvement of the atmospheric environment.

Description

Technical Field

[0001] This invention discloses a test method for evaluating the environmental benefits of intelligent transportation systems based on a standard operating condition library constructed using Markov chains. Specifically, it relates to a test method for evaluating the environmental benefits of intelligent transportation systems based on a standard operating condition library constructed using Markov chains to simulate the emissions of motor vehicles under different operating conditions. Background Technology

[0002] Traffic congestion, traffic pollution, and traffic energy consumption are common problems faced by many large cities worldwide. In congested conditions, vehicles frequently operate at idling, low speeds, and acceleration / deceleration, leading to increased energy consumption and pollutant emissions. Intelligent transportation optimization systems, as a novel traffic congestion relief solution, are being explored and tested by an increasing number of countries and regions, with some areas achieving significant results in congestion relief and energy conservation / emission reduction. Vehicle speed guidance technology, as a pioneering technology in intelligent transportation systems, will alter the driving conditions of motor vehicles, thereby affecting vehicle exhaust emissions. Currently, the main methods used are traffic-emission model simulation and real-world testing. However, methods such as traffic-emission model simulation rely on older, established models that are subject to reliability challenges due to advancements in motor vehicle technology and a lack of emission process models for new vehicle models. Real-world testing is limited by the fact that vehicle speed guidance technology is often applied locally, resulting in limited data collection, poor representativeness of evaluation results, high testing costs, difficulty in reproducing the testing environment affecting vehicle driving conditions and emission processes, difficulty in comparing test results, and high testing costs, making it impossible to collect sufficient data to characterize emission features. This method constructs a speed-guided coupled evaluation model of motor vehicle driving conditions and emissions, aiming to provide theoretical and technical support for the improvement of speed-guided technology and the refined emission control of exhaust gases, and to help the continuous improvement of motor vehicle exhaust emissions. Summary of the Invention

[0003] To address the problems existing in the background technology, this invention proposes a method for testing energy conservation and emission reduction of motor vehicles by constructing a speed-guided coupled evaluation model of motor vehicle operating conditions and emissions, and by using machine learning and Markov chains. This method is a test method for evaluating the environmental benefits of intelligent transportation systems.

[0004] This invention is achieved through the following technical solution: an environmental benefit assessment and testing method for intelligent transportation systems based on a standard operating condition library constructed using Markov chains, comprising the following modules: a data acquisition and processing module, a vehicle speed guidance standard operating condition library module, an exhaust emission model template, and a coupling assessment module.

[0005] The data acquisition and processing module is used to collect the operating condition data of motor vehicles during the drum test and the exhaust emission data of motor vehicles under typical actual road tests, and to pre-clean the collected data to reduce data errors.

[0006] The vehicle speed guidance standard operating condition library module uses the Markov chain method to process the operating condition data obtained from the data acquisition and processing module and generate candidate operating condition curves; curves that meet the set conditions are combined into a representative operating condition library.

[0007] The exhaust emission module is used to process exhaust emission data, mine proxy parameters and reduce noise to accurately predict emissions; construct emission models, use multiple algorithms to build models and optimize hyperparameters to improve model performance.

[0008] The coupling evaluation module is used to establish coupling parameters and coupling mechanisms between standard operating conditions and transient emission models of motor vehicles, thereby building a coupled evaluation model of motor vehicle driving conditions and emissions, and realizing a quantitative evaluation of the traffic pollution reduction benefits before and after the use of vehicle speed guidance.

[0009] Furthermore, the data acquisition and processing module includes a real road test submodule, a motor vehicle exhaust emission data acquisition submodule, and a data preprocessing module.

[0010] The actual road test submodule is used to collect data on the operating conditions of motor vehicles with and without speed guidance. It uses a GPS positioning system and on-board terminal equipment (such as OBD) to obtain data such as the vehicle's speed, acceleration, and trajectory during operation, thereby constructing an operating condition database.

[0011] The vehicle exhaust emission data submodule is used to collect vehicle exhaust emission data, and the vehicle emission data is obtained directly through drum testing, AVL four-wheel drive drum and gas monitoring, and particulate matter monitoring equipment.

[0012] The data preprocessing module is used to pre-clean the collected data to reduce data errors. The data cleaning process includes outlier detection and missing data filling.

[0013] Furthermore, the vehicle speed guidance standard operating condition library module includes an operating condition data processing submodule and a standard operating condition library verification submodule.

[0014] The working condition data processing submodule is used to process the working condition data obtained by the data acquisition and processing module. It uses the Markov chain method to process the data and generate working condition curves that can be used as candidates. After verification that they meet the set conditions, they form a representative working condition library.

[0015] The standard operating condition library verification submodule is used to verify the reliability of the standard operating condition libraries corresponding to light-duty gasoline vehicles, light-duty new energy vehicles, and diesel trucks. By analyzing the similarities and differences between the constructed standard operating conditions and the original collected operating condition features, the reliability and effectiveness of the method are verified, thereby providing a basis for elucidating the influence mechanism of vehicle speed guidance on motor vehicle driving conditions.

[0016] Furthermore, the exhaust emission model module includes an emission data processing submodule and an emission model submodule.

[0017] The emission data processing submodule is used to process the motor vehicle exhaust emission data obtained by the data acquisition and processing module. First, based on the large sample size of exhaust emission data obtained, the proxy parameters of the emission model are deeply mined to construct a parameter combination that conforms to the characteristics of the emission process.

[0018] Then, noise reduction is performed on features that may affect vehicle exhaust emissions to reduce interference from data noise on the algorithm model, resulting in more reasonable and accurate predictions of vehicle exhaust emissions. The emission model submodule is used to construct the vehicle exhaust emission model.

[0019] First, 70% of the vehicle exhaust emission data was divided into a training set and 30% into a test set. Then, five cutting-edge ensemble learning algorithms in machine learning were selected (Random Forest, GBDT, XGBoost, LightGBM, and CatBoost). XGBoost, an improvement on the GBDT algorithm, uses the feature columns of the training set as the model input and the emission data columns of the training set as the model's prediction data. Transient emission models of motor vehicles were constructed using these five machine learning algorithms.

[0020] For each algorithm, the hyperparameters of the corresponding model are automatically optimized using 10-fold cross-validation and Bayesian optimization. Accuracy and R-squared are then used to measure these optimizations. 2 The model performance is evaluated using metrics such as root mean square error, thereby improving the efficiency and accuracy of parameter tuning, and ultimately enhancing the model's robustness and generalization ability.

[0021] Furthermore, the coupling evaluation module includes a coupling parameter submodule and an emission reduction evaluation submodule.

[0022] The coupling parameter submodule, based on the constructed motor vehicle exhaust emission model and the vehicle speed guidance standard operating condition library, determines the coupling parameters of the standard operating condition and the transient emission model of motor vehicles, and establishes a vehicle speed guidance-oriented coupled evaluation model for motor vehicle driving conditions and emissions.

[0023] The emission reduction assessment submodule uses the constructed coupled assessment model to calculate the emission process of motor vehicles under standard operating conditions with and without vehicle speed guidance.

[0024] By using a transient model of motor vehicles, the average emission factor under standard operating conditions with and without vehicle speed guidance is calculated to quantitatively assess the emission reduction benefits of vehicles using speed guidance.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Unified evaluation standard: The present invention provides a unified standard for evaluating the impact of vehicle speed guidance on motor vehicle exhaust emissions by establishing a coupled evaluation model, which solves the problem of large differences and difficulty in comparison of results in existing evaluation methods.

[0026] 2. Accurate assessment of benefits: It can accurately quantify the environmental benefits of speed guidance, clearly present the actual effect of speed guidance technology in reducing motor vehicle exhaust emissions, and provide strong data support for relevant decision-making.

[0027] 3. Promote technological development: Provide direction for the improvement and optimization of vehicle speed guidance technology, which will help the technology be widely applied in actual traffic scenarios, thereby promoting the work of reducing motor vehicle exhaust emissions and playing an important role in improving the quality of the atmospheric environment. Attached Figure Description

[0028] Figure 1 This is a flowchart of the testing method of the present invention.

[0029] Figure 2 This is the technical roadmap for this testing method. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. The described examples are only some embodiments of the present invention, and not all embodiments.

[0031] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0032] Example: This experiment was conducted in Wuqing District of Tianjin and Zibo City of Shandong Province to investigate the impact of vehicle speed guidance on emissions from different types of motor vehicles.

[0033] The test subjects included diesel trucks (such as Jiefang brand cargo trucks and Geerfa A5X cargo trucks) and a plug-in hybrid electric vehicle that meets the China VI emission standard.

[0034] Test process

[0035] In Wuqing District of Tianjin and Zibo City of Shandong Province, typical driving roads for gasoline cars and diesel trucks were selected to conduct actual road tests covering both situations with and without speed guidance.

[0036] Data acquisition and processing module: Using the P-10 Pro professional high-precision GNSS recorder, the latitude, longitude, speed and altitude information of the vehicle are collected in real time, second by second, to construct vehicle driving condition data.

[0037] To minimize the interference of traffic conditions and driver behavior on the experimental results, measures such as alternating data collection, fixing the driver, and simultaneously operating two devices were adopted to ensure the accuracy and reliability of the data.

[0038] In addition, preprocessing work such as outlier detection and missing value imputation was performed on the collected data to ensure that the constructed working condition data is complete and effective, providing a solid data foundation for subsequent research.

[0039] Diesel truck emission data collection: Select a chassis dynamometer that can simulate the driving conditions of diesel trucks on different roads such as urban areas, suburbs, and highways, and connect it to the AVL MOVE on-board exhaust gas analysis system.

[0040] Different initial load conditions were set to preheat, leak-check, zero-check, and calibrate the equipment. Tests were conducted on a chassis dynamometer according to the settings of simulated actual road conditions and standard test conditions. Environmental conditions were strictly controlled, and exhaust emission data and vehicle operating parameters were recorded in real time.

[0041] Plug-in hybrid electric vehicle emissions data acquisition: Emissions testing was conducted in the laboratory using a chassis dynamometer. The AVL MOVE on-board exhaust gas analysis system was used to test exhaust emissions, and the eDrive Analyzer acquired real-time changes in state of charge (SOC).

[0042] The tests take into account actual road conditions and WLTC conditions, setting initial states of charge at 20% and 50%, as well as cold start and hot start tests.

[0043] Vehicle speed guidance standard operating condition library module: The standard operating conditions are constructed using the Markov chain method. The original operating condition data is discretized, the state transition probability matrix is ​​calculated, the operating conditions are generated and filtered, and finally 100 standard operating conditions with and without vehicle speed guidance are obtained to form a standard operating condition group, thus constructing the vehicle speed guidance standard operating condition library.

[0044] The reliability of the standard operating condition was verified by analyzing the deviation of the operating condition characteristics and the kernel density estimation of the speed distribution. The results show that the standard operating condition can effectively represent the actual driving characteristics of the vehicle.

[0045] Exhaust emission model module: Proposes a "different pollution and different model" vehicle emission modeling framework, and adopts a "knowledge-driven + data-driven" model.

[0046] Transient emission model for China VI diesel trucks: Based on test data from 10 diesel truck tests on a chassis dynamometer, a transient emission model for CO2 and NOx was established using a Super-learner model. After vehicle validation and comparison with the MOVES model, it showed good predictive performance and advantages.

[0047] Transient emission model of plug-in hybrid electric vehicles: A PHEV meeting the China VI emission standard was selected and tested on a chassis dynamometer. A two-stage model method was used to predict the transient emissions of CO2 and NOx. The SHAP method was used to explain the importance of the model input parameters, and it was found that VSP and SOC are the main factors affecting CO2 and NOx emissions, respectively.

[0048] Coupled evaluation module: Based on the vehicle speed-guided standard operating condition library and the transient emission model of motor vehicles, a coupled evaluation model of motor vehicle driving conditions and emissions is constructed.

[0049] The emission values ​​of motor vehicles under different emission processes with and without speed guidance are obtained by model calculation, converted into emission factors, and the environmental benefits are quantitatively assessed.

[0050] Data on the changes in emission factors for China VI diesel trucks and plug-in hybrid electric vehicles before and after speed guidance have been presented.

[0051] Starting with these data differences, we can delve into the reasons for the differences in emission reduction effects among different vehicle models. This is crucial for understanding the application effect of speed-guided technology.

[0052] For example, diesel trucks are mainly used for cargo transportation, with complex operating conditions and large load variations. Their engine characteristics and exhaust treatment systems are different from those of plug-in hybrid electric vehicles.

[0053] Diesel truck engines prioritize power output. Under speed guidance, CO2 emission factors decrease significantly. This is because speed guidance optimizes driving conditions, reduces unnecessary acceleration and deceleration, and makes the engine run more efficiently. However, NOx emission factors do not decrease significantly, possibly because the effectiveness of the exhaust gas treatment device in treating NOx is less affected by speed guidance.

[0054] For plug-in hybrid electric vehicles, speed guidance has a significant impact on CO2 and NOx emission factors at different initial state of charge (SOC).

[0055] When the State of Charge (SOC) is low, the vehicle relies more on fuel-powered operation. Speed-guided operation optimizes driving conditions, making the engine work more efficiently and reducing emissions more significantly. When the SOC is high, the proportion of electric drive increases, and speed-guided operation reduces energy loss when switching between the electric motor and the engine, thus reducing emissions.

[0056] This difference provides a targeted basis for applying speed guidance technology to different types of vehicles, helps optimize speed guidance strategies, improves overall emission reduction efficiency, promotes the precise application of speed guidance technology in different vehicle models, and better achieves the goal of reducing vehicle exhaust emissions.

[0057] emissions Speed-free average emission factor Speed-guided average emission factor range of change CO2 433.2g / km 403.7g / km Down 6.8% NOx 0.0394g / km 0.0391g / km Decrease of 0.7%

[0058] Changes in emission factors for China VI diesel trucks

[0059] emissions Initial SOC state Speed-free average emission factor Speed-guided average emission factor range of change CO2 20% 169.8g / km 141.3g / km Down 16.6% CO2 50% 67.6g / km 52.2g / km Decrease of 22.9% NOx 20% 0.22g / km 0.17g / km Down 21.1% NOx 50% 0.17g / km 0.13g / km Decrease of 27.7%

[0060] Changes in emission factors of plug-in hybrid electric vehicles.

Claims

1. A method for environmental benefit assessment and testing of intelligent transportation systems based on a standard operating condition library constructed using Markov chains, characterized in that: The environmental benefit assessment method for the intelligent transportation system includes a data acquisition and processing module, a vehicle speed guidance standard operating condition library module, an exhaust emission model module, and a coupled assessment module. The data acquisition and processing module is used to collect the operating condition data of the motor vehicle during the drum test and the exhaust emission data of the motor vehicle under typical actual road tests, and to perform data pre-cleaning on the collected data to reduce data errors. The vehicle speed guidance standard operating condition library module uses the Markov chain method to process the operating condition data obtained from the data acquisition and processing module, generate candidate operating condition curves, and form a representative operating condition library by selecting curves that meet the set conditions. The exhaust emission model module is used to process exhaust emission data, mine proxy parameters and reduce noise to accurately predict emissions. Build emission models, use multiple algorithms to build models and optimize parameters to improve model performance; The coupling evaluation module is used to establish coupling parameters and coupling mechanisms between standard operating conditions and transient emission models of motor vehicles, thereby building a coupled evaluation model of motor vehicle driving conditions and emissions, and realizing a quantitative evaluation of the traffic pollution reduction benefits before and after the use of vehicle speed guidance.

2. The environmental benefit assessment and testing method for intelligent transportation systems as described in claim 1, characterized in that: The data acquisition and processing module includes a real-road testing submodule, a vehicle exhaust emission data collection submodule, and a data preprocessing module. The actual road test submodule is used to collect data on the operating conditions of motor vehicles with and without speed guidance, providing data support for building a standard operating condition library; The motor vehicle exhaust emission data submodule is used to collect motor vehicle exhaust emission data, obtain motor vehicle emission data directly through drum testing, and establish an emission database based on this data. The data preprocessing module is used to pre-clean the collected data to reduce data errors. The data cleaning process includes outlier detection and missing data filling.

3. The environmental benefit assessment and testing method for intelligent transportation systems as described in claim 1, characterized in that: The vehicle speed guidance standard operating condition library module includes an operating condition data processing submodule and a standard operating condition library verification submodule. The working condition data processing submodule is used to process the working condition data obtained by the data acquisition and processing module. It uses the Markov chain method to process the data and generate working condition curves that can be used as candidates. After verification that the set conditions are met, a representative working condition library is formed. The standard operating condition library verification submodule is used to verify the reliability of the standard operating condition libraries corresponding to light-duty gasoline vehicles, light-duty new energy vehicles, and diesel trucks. By analyzing the similarities and differences between the constructed standard operating conditions and the original collected operating condition features, the reliability and effectiveness of the method are verified, and the influence mechanism of vehicle speed guidance on the driving conditions of motor vehicles is clarified.

4. The environmental benefit assessment and testing method for intelligent transportation systems as described in claim 1, characterized in that: The exhaust emission model module includes an emission data processing submodule and an emission model submodule. The emission data processing submodule is used to acquire and process motor vehicle exhaust emission data provided by the data acquisition and processing module. Based on a large sample size of data, it performs in-depth mining and constructs emission model proxy parameters, and then reduces noise on relevant features to accurately predict exhaust emissions. The emission model submodule is used to construct a vehicle exhaust emission model. First, the data is divided into training and testing sets. A cutting-edge ensemble learning algorithm is selected, using the training set feature columns and emission data columns as input and prediction data, respectively, to construct a transient emission model. Then, cross-validation and optimization methods are used to automatically adjust the model's hyperparameters, improving the quality of parameter tuning and enhancing the model's stability and applicability.

5. The environmental benefit assessment and testing method for intelligent transportation systems as described in claim 1, characterized in that: The coupling evaluation module includes a coupling parameter submodule and an emission reduction evaluation submodule. The coupling parameter submodule determines the coupling parameters based on existing motor vehicle exhaust emission models and speed-guided standard operating condition libraries, and establishes a speed-guided coupled evaluation model for motor vehicle driving conditions and emissions. The emission reduction assessment submodule uses the constructed coupled assessment model to calculate the average emission factor of motor vehicles under standard operating conditions with and without vehicle speed guidance, thereby quantitatively assessing the emission reduction benefits after vehicle speed guidance.

Images