Method for testing energy saving effect of route memory-based predictive energy management system and electronic device

Abstract

The application relates to the technical field of vehicle testing, in particular to a method for testing energy-saving effect of a predictive energy management system based on route memory and electronic equipment, wherein the method comprises the following steps: calculating information entropy corresponding to a plurality of candidate test routes based on route segment features of a plurality of route segments in the plurality of candidate test routes to select a target test route; under the condition that route memory and energy-saving effect of a predictive energy management system of a test vehicle respectively meet preset requirements and preset energy-saving conditions, obtaining first carbon emission of the test vehicle under a first preset working condition and second carbon emission of a reference vehicle under a second preset working condition based on the target test route, and calculating energy-saving rate of the test vehicle under the first preset working condition. The application can realize energy-saving effect evaluation of the predictive energy management system which does not depend on navigation activation and relies on historical route memory, and can unify fuel consumption and electricity consumption into carbon emission to realize unified quantitative comparison of energy-saving effect of different power type vehicles.

Description

Technical Field

[0001] This application relates to the field of vehicle testing technology, and in particular to a method and electronic device for testing the energy-saving effect of a predictive energy management system based on route memory. Background Technology

[0002] Predictive Energy Management (PEM) systems can optimize energy distribution in hybrid electric vehicles using information about the road ahead, thereby achieving overall vehicle energy savings. Unlike systems that rely on user-defined navigation destinations, PEM systems come in two types: one uses navigation map information for prediction, and the other can still activate and perform energy-saving functions even when the user has not set a navigation destination.

[0003] In related technologies, the energy-saving test methods for PEM strategies all use indoor rotary drum test benches for simulation testing, and the test objects are all PEM strategies that require navigation activation.

[0004] However, in related technologies, the energy-saving test methods for PEM strategies are all designed for PEM strategies that require navigation activation. They do not take into account that PEM may be in navigation scenarios or non-navigation scenarios in actual applications. There is a lack of test methods for the energy-saving effect of PEM strategies that do not rely on navigation activation. Furthermore, there is a lack of solutions to problems that may be encountered when testing the energy-saving effect of PEM strategies that do not rely on navigation activation, such as interference from external factors, baseline energy consumption drift, and overfitting of the PEM system. As a result, the test results cannot truly reflect the real energy-saving effect of the PEM strategy, which urgently needs to be addressed. Summary of the Invention

[0005] This application provides a method and electronic device for testing the energy-saving effect of a predictive energy management system based on route memory. This addresses the problem that in related technologies, energy-saving testing methods for PEM strategies are all designed for PEM strategies that require navigation activation, lacking a method for testing the energy-saving effect of PEM strategies that do not rely on navigation activation. Furthermore, there is a lack of solutions to problems that may be encountered when testing the energy-saving effect of PEM strategies that do not rely on navigation activation, such as interference from external factors, drift of baseline energy consumption, and overfitting of the PEM system. As a result, the test results cannot truly reflect the real energy-saving effect of the PEM strategy.

[0006] The first aspect of this application provides a method for testing the energy-saving effect of a predictive energy management system based on route memory, comprising the following steps: acquiring the road segment characteristics of multiple road segments among multiple candidate test routes; calculating the information entropy corresponding to the multiple candidate test routes based on the road segment characteristics; selecting a target test route from the multiple candidate test routes based on the information entropy; under the condition that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet preset requirements and preset energy-saving conditions respectively, acquiring the first carbon emission of the test vehicle under a first preset operating condition and the second carbon emission of a reference vehicle under a second preset operating condition based on the target test route; and calculating the energy-saving rate of the test vehicle under the first preset operating condition based on the first carbon emission and the second carbon emission.

[0007] Optionally, in one embodiment of this application, the step of obtaining the road segment features of multiple road segments in multiple candidate test routes and calculating the information entropy corresponding to the multiple candidate test routes based on the road segment features includes the following steps: obtaining the slope features, speed limit features, and historical traffic flow features of multiple road segments in multiple candidate test routes, and dividing the slope features, speed limit features, and historical traffic flow features into multiple levels; statistically analyzing the proportion of road segments corresponding to each level of the slope features, speed limit features, and historical traffic flow features to obtain the probability distribution corresponding to the slope features, speed limit features, and historical traffic flow features; and calculating the information entropy corresponding to the multiple candidate test routes based on the probability distribution.

[0008] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under a first preset operating condition and the second carbon emission of a reference vehicle under a second preset operating condition based on the target test route, when the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet preset requirements and preset energy-saving conditions respectively, the method further includes: determining multiple test sampling points of multiple road segments corresponding to each round of driving data in the multi-round driving data of the test vehicle based on the target test route; calculating the matching deviation between the multiple test sampling points and the target centerline, so as to calculate the trajectory matching quality between the multiple road segments and the target test route based on the matching deviation; determining the proportion of road segments that have reached the convergence state in the target test route based on the trajectory matching quality, so as to determine the route memory of the predictive energy management system of the test vehicle based on the proportion of road segments.

[0009] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, when the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the method further includes: obtaining, based on the target test route, multiple test results of the test vehicle under the navigation mode being off and the predictive energy management system being on, and multiple test results of the reference vehicle under the navigation mode being off and the predictive energy management system being off; and obtaining, based on the multiple test results corresponding to the test vehicle and the reference vehicle, the number of outlier road segments of the test vehicle in each test process, so as to determine the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments.

[0010] Optionally, in one embodiment of this application, determining the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments includes: dividing the multiple road segments into multiple road segment groups based on the road segment characteristics of multiple road segments of the target test route; calculating an initial baseline mean and an initial baseline standard deviation based on the energy-saving rate of a preset road segment in each road segment group of the predictive energy management system of the test vehicle; identifying outlier road segments in the preset road segments and the other road segments based on the initial baseline mean, the initial baseline standard deviation, and the energy-saving rate of other road segments in each road segment group besides the preset road segments; determining the total number of real-time outlier road segments of the test vehicle during the test process based on the number of outlier road segments in the multiple road segment groups; calculating the real-time cumulative outlier rate of the test vehicle during multiple rounds of testing based on the total number of real-time outlier road segments, and determining the energy-saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate.

[0011] Optionally, in one embodiment of this application, determining the energy-saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate includes: calculating the energy-saving rate of the predictive energy management system of the test vehicle based on the test results when the real-time cumulative outlier rate meets a preset outlier condition; calculating the range of energy-saving rates of the test vehicle in the first target round of testing based on the energy-saving rate, and calculating the absolute value of the difference between the average energy-saving rate of the first target round of testing and the average energy-saving rate of the second target round of testing; and determining the energy-saving effect of the predictive energy management system of the test vehicle based on the range of energy-saving rates and the absolute value.

[0012] Optionally, in one embodiment of this application, the step of obtaining the first carbon emission of the test vehicle under a first preset operating condition and the second carbon emission of the reference vehicle under a second preset operating condition based on the target test route further includes: obtaining multiple types of disturbance test data of the test vehicle to calculate the energy saving rate of the predictive energy management system of the test vehicle under multiple types of disturbance tests; calculating multiple types of generalization indicators of the test vehicle based on the energy saving rate under multiple types of disturbance tests; and obtaining the first carbon emission and the second carbon emission based on the target test route when the multiple types of generalization indicators meet the corresponding preset indicator ranges.

[0013] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, the method further includes: determining that the test vehicle is in the first preset operating condition when the navigation condition of the test vehicle is off and the predicted energy management system condition is on; and determining that the reference vehicle is in the second preset operating condition when the navigation condition of the reference vehicle is off and the predicted energy management system condition is off.

[0014] A second aspect of this application provides a testing device for the energy-saving effect of a predictive energy management system based on route memory, comprising: a selection module, configured to acquire road segment features of multiple road segments among multiple candidate test routes, calculate information entropy corresponding to the multiple candidate test routes based on the road segment features, and select a target test route from the multiple candidate test routes based on the information entropy; a first acquisition module, configured to, based on the target test route, acquire a first carbon emission of the test vehicle under a first preset operating condition and a second carbon emission of a reference vehicle under a second preset operating condition, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle respectively meet preset requirements and preset energy-saving conditions; and a first calculation module, configured to calculate the energy-saving rate of the test vehicle under the first preset operating condition based on the first carbon emission and the second carbon emission.

[0015] Optionally, in one embodiment of this application, the selection module includes the following steps: a first acquisition unit, used to acquire slope characteristics, speed limit characteristics, and historical traffic flow characteristics of multiple road segments in multiple candidate test routes, and to divide the slope characteristics, speed limit characteristics, and historical traffic flow characteristics into multiple levels; a statistics unit, used to calculate the proportion of road segments corresponding to each level of the slope characteristics, speed limit characteristics, and historical traffic flow characteristics, and to obtain the probability distribution corresponding to the slope characteristics, speed limit characteristics, and historical traffic flow characteristics; and a first calculation unit, used to calculate the information entropy corresponding to the multiple candidate test routes based on the probability distribution.

[0016] Optionally, in one embodiment of this application, it further includes: a determining module, configured to, before obtaining the first carbon emission of the test vehicle under a first preset operating condition and the second carbon emission of a reference vehicle under a second preset operating condition based on the target test route, determine multiple test sampling points of multiple road segments corresponding to each round of driving data in the multi-round driving data of the test vehicle, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet preset requirements and preset energy-saving conditions respectively; a second calculation module, configured to calculate the matching deviation between the multiple test sampling points and the target centerline, so as to calculate the trajectory matching quality between the multiple road segments and the target test route based on the matching deviation; and a first determining module, configured to determine the proportion of road segments that have reached a convergence state in the target test route based on the trajectory matching quality, so as to determine the route memory of the predictive energy management system of the test vehicle based on the proportion of road segments.

[0017] Optionally, in one embodiment of this application, it further includes: a second acquisition module, configured to, before acquiring the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, and before acquiring the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, acquire the multi-round test results of the test vehicle under the navigation condition being off and the predictive energy management system under the operating condition being on, and the multi-round test results of the reference vehicle under the navigation condition being off and the predictive energy management system under the operating condition being off; and a second determination module, configured to, based on the multi-round test results corresponding to the test vehicle and the reference vehicle, acquire the number of outlier road segments of the test vehicle in each round of testing, so as to determine the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments.

[0018] Optionally, in one embodiment of this application, the second determining module includes: a segmentation unit, configured to divide the multiple road segments into multiple road segment groups based on the road segment characteristics of multiple road segments of the target test route; a second calculation unit, configured to calculate an initial benchmark mean and an initial benchmark standard deviation based on the energy saving rate of a preset road segment in each road segment group by the predictive energy management system of the test vehicle; an identification unit, configured to identify outliers in the preset road segment and the other road segments based on the initial benchmark mean, the initial benchmark standard deviation, and the energy saving rate of other road segments in each road segment group besides the preset road segment; a first determining unit, configured to determine the total number of real-time outliers of the test vehicle during the test process based on the number of outliers in the multiple road segment groups; and a second determining unit, configured to calculate the real-time cumulative outlier rate of the test vehicle during multiple rounds of testing based on the total number of real-time outliers, and determine the energy saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate.

[0019] Optionally, in one embodiment of this application, the second determining unit includes: a first calculation subunit, configured to calculate the energy saving rate of the predictive energy management system of the test vehicle based on the test results, provided that the real-time cumulative outlier rate meets a preset outlier condition; a second calculation subunit, configured to calculate the range of energy saving rates of the test vehicle in the first target round of testing based on the energy saving rate, and calculate the absolute value of the difference between the average energy saving rate of the first target round of testing and the average energy saving rate of the second target round of testing; and a determining subunit, configured to determine the energy saving effect of the predictive energy management system of the test vehicle based on the range of energy saving rates and the absolute value.

[0020] Optionally, in one embodiment of this application, the first acquisition module includes: a third calculation unit, configured to acquire multi-type disturbance test data of the test vehicle to calculate the energy saving rate of the predictive energy management system of the test vehicle under multi-type disturbance tests; and a second acquisition unit, configured to calculate multi-type generalization indicators of the test vehicle based on the energy saving rate under the multi-type disturbance tests, and, when the multi-type generalization indicators meet the corresponding preset indicator range, acquire the first carbon emission and the second carbon emission based on the target test route.

[0021] Optionally, in one embodiment of this application, it further includes: a third determining module, configured to determine that the test vehicle is in the first preset operating condition before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, when the navigation operating condition of the test vehicle is off and the predicted energy management system operating condition is on; and a fourth determining module, configured to determine that the reference vehicle is in the second preset operating condition when the navigation operating condition of the reference vehicle is off and the predicted energy management system operating condition is off.

[0022] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the program to implement the energy-saving effect testing method for a predictive energy management system based on route memory as described in the above embodiments.

[0023] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for testing the energy-saving effect of a predictive energy management system based on route memory.

[0024] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, is used to implement the above-described method for testing the energy-saving effect of a predictive energy management system based on route memory.

[0025] This application embodiment selects a target test route from multiple candidate test routes based on information entropy. It calculates the energy-saving rate of the test vehicle when navigation is off and the predictive energy management system is on, and the energy-saving rate of the reference vehicle is off. This ensures stable energy consumption of the reference vehicle under suitable test route conditions, through the rules of rotating and discontinuous use of multiple reference vehicles, preventing artificially inflated or understated energy-saving rates and improving the credibility of test results. Through strategies such as real-time data quality monitoring and graded correction based on road segment clustering and outlier rate classification, real-time quality control of the test process is achieved, effectively avoiding the risk of entire-round data scrapping or erroneous data participating in termination decisions due to local anomalies. Finally, a test condition determination method combining quantification and experience ensures that the formal test is conducted under the same conditions as the route memorization and verification test phases, completing a unified carbon emission evaluation of the PEM energy-saving effect. Therefore, this solves the problem that in related technologies, the energy-saving test methods for PEM strategies are all set up for PEM strategies that require navigation activation, and there is a lack of test methods for the energy-saving effect of PEM strategies that do not rely on navigation activation. Furthermore, there is a lack of solutions to problems that may be encountered when testing the energy-saving effect of PEM strategies that do not rely on navigation activation, such as interference from external factors, baseline energy consumption drift, and overfitting of PEM systems. As a result, the test results cannot truly reflect the real energy-saving effect of PEM strategies.

[0026] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0027] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for testing the energy-saving effect of a predictive energy management system based on route memory, according to an embodiment of this application. Figure 2 This is a schematic diagram of the energy-saving effect testing device for a route memory-based predictive energy management system provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application.

[0028] Figure label: 10-Energy-saving effect test device for predictive energy management system based on route memory; 100-Selection module, 200-First acquisition module and 300-First calculation module; 301-Memory, 302-Processor and 303-Communication interface. Detailed Implementation

[0029] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0030] The following describes, with reference to the accompanying drawings, a method and electronic device for testing the energy-saving effect of a predictive energy management system based on route memory, according to embodiments of this application. Addressing the problems mentioned in the background section, this application provides a method for testing the energy-saving effect of a predictive energy management system based on route memory. In this method, a target test route can be selected from multiple candidate test routes based on information entropy. The energy-saving rate of the test vehicle is calculated when navigation is off and the predictive energy management system is on, and when a reference vehicle is off and the predictive energy management system is off. This ensures stable energy consumption of the reference vehicle under suitable test route conditions, through the rules of rotating and discontinuous use of multiple reference vehicles, avoiding artificially inflated or understated energy-saving rates, and improving the credibility of the test results. Real-time data quality monitoring and graded correction strategies based on road segment clustering and outlier rate classification achieve real-time quality control of the test process, effectively avoiding the risk of entire-round data scrapping or erroneous data participating in termination decisions due to local anomalies. Finally, a test condition determination method combining quantification and experience ensures that the formal test is conducted under the same conditions as the route memory and verification test stages, completing a unified carbon emission evaluation of the PEM energy-saving effect. Therefore, this solves the problem that in related technologies, the energy-saving test methods for PEM strategies are all set up for PEM strategies that require navigation activation, and there is a lack of test methods for the energy-saving effect of PEM strategies that do not rely on navigation activation. Furthermore, there is a lack of solutions to problems that may be encountered when testing the energy-saving effect of PEM strategies that do not rely on navigation activation, such as interference from external factors, baseline energy consumption drift, and overfitting of PEM systems. As a result, the test results cannot truly reflect the real energy-saving effect of PEM strategies.

[0031] Vehicle PEM (Power Estimation) can optimize energy distribution in hybrid electric vehicles using information about the road ahead, thereby achieving overall vehicle energy savings. Unlike technologies that rely on user-defined navigation destinations, PEM technology generally falls into two categories: one type requires prediction based on navigation map information; the other type can still be activated and perform energy-saving functions even when the user has not set navigation. This latter type of PEM technology mainly utilizes the vehicle's historical route memory (such as frequently used routes) or cloud-based push notifications to perform predictive optimization and energy allocation management for routes already memorized by the vehicle. For example, without navigation enabled, the cloud platform can determine whether the vehicle is traveling on the target route and perform targeted predictive energy management.

[0032] The engineering implementation of this type of route memory-based non-navigation PEM system typically follows a two-stage working mode of "memory pre-construction → non-navigation invocation".

[0033] Pre-construction of navigation data: With navigation enabled, the vehicle travels the target route several times. The system uses navigation map information to pre-extract and store static road features such as slope, curvature, speed limit, and traffic light density of the route, forming a digital map memory for the route in the cloud or on the vehicle, and marking the route as a "frequently traveled route".

[0034] During the navigation-free phase: When the vehicle travels on the same route again and the user closes the navigation interface, the PEM system obtains the vehicle's current location via GPS and matches this location with the habitual route stored in the cloud memory. Upon successful matching, the system generates the optimal energy planning strategy based on pre-stored route data (rather than real-time navigation information) and real-time traffic conditions pushed from the cloud, thereby optimizing vehicle energy efficiency under navigation-free conditions.

[0035] However, existing testing methods for the energy-saving effects of predictive energy management systems (PEM) based on PEM strategies are all designed for PEM strategies that require navigation activation, and they employ indoor rotary drum test benches, which cannot reflect real-world road environments, traffic conditions, and weather conditions. There are currently no publicly available testing methods for the energy-saving effects of PEM strategies that do not rely on navigation activation.

[0036] On the other hand, the testing methods for the energy-saving effect of predictive energy management systems based on PEM strategies in related technologies also have several technical problems: (1) The complex usage scenarios of PEM strategy in actual application are ignored: some vehicles rely on navigation, some rely on existing route memory, and each vehicle has a different route memory. If traditional testing methods are used directly, the route memory library of the vehicle PEM system may not have been established or the pre-built quality may be insufficient during testing. This may cause the PEM system to fail to match the habitual route correctly in the absence of navigation, resulting in the energy-saving effect being seriously underestimated or not being reflected.

[0037] (2) Lack of objective means to determine the stable output of energy-saving effect. With the route memory database established, whether the PEM system can stably and repeatedly exert its energy-saving effect during multiple navigation-free calls is a key indicator for evaluating its practical value. Existing technologies lack testing methods for navigation-free PEM systems and do not provide a technical solution for determining whether "the energy-saving effect has been stably output." If judgment is based solely on the subjective experience of testers, the results are not reproducible and lack comparability.

[0038] (3) There is a lack of systematic methods for identifying and eliminating interference factors in real-world road testing. Moving the test from indoor drums to real roads improves the external validity of the test, but it introduces a series of uncontrollable factors such as traffic flow fluctuations, weather changes, and road surface condition changes. These factors can lead to local abnormal road sections in single-round test data (such as sudden congestion causing abnormally low or high energy efficiency in a small section). The methods used in related technologies, such as discarding the entire round of data or manually removing outliers, either result in a large amount of test waste or introduce subjective bias.

[0039] (4) Reference vehicle adaptive learning-induced baseline energy consumption drift. Although the reference vehicle has its PEM system turned off, it may still have adaptive learning functions based on driving habits (such as transmission shift adaptation, engine A / F learning value, etc.). When the same reference vehicle repeatedly drives the same route in multiple rounds of testing, its energy consumption may decrease successively, resulting in a systematic deviation in the energy saving rate calculation results.

[0040] (5) There is a lack of means to identify the overfitting risk of PEM systems. PEM systems based on route memory may only be able to "memorize" specific trajectories and cannot generalize to different driving conditions (such as different traffic flows, weather, load) on the same route. If overfitting is not identified, a system that has little effect in actual use may be wrongly evaluated as an efficient system.

[0041] (6) Lack of a unified evaluation dimension for carbon emissions. The PEM energy-saving effect test methods in related technologies do not uniformly convert the electric energy consumption and fuel consumption of hybrid vehicles into carbon emissions for evaluation. Instead, they use fuel consumption per 100 kilometers, electric consumption per 100 kilometers, or comprehensive energy consumption as evaluation indicators respectively.

[0042] Based on this, this application provides a method for testing the energy-saving effect of a predictive energy management system based on route memory. This method mainly tests the energy-saving effect of a vehicle's predictive energy management system based on route memory. By setting a prerequisite for the full functioning of PEM technology—the test vehicle should drive multiple times on a specific route in advance to establish "route memory"—a unified vehicle testing standard is established, thereby unifying the consumption of different energy types to the environmental benefit indicator of carbon emissions, and providing a unified standard for evaluating energy-saving effects.

[0043] The technical principles and ideas of this method can be expressed, but are not limited to, as follows: To address the prerequisite that such PEM systems rely on vehicles repeatedly driving a specific route to establish "route memory," a closed-loop testing framework is proposed, consisting of "information entropy calibration → establishment of habitual route memory bank → random rotation verification → real-time anomaly monitoring and hierarchical correction → inflection point adaptive termination → generalization verification against disturbances → conservative evaluation." First, by calibrating the information entropy of the test route, the complexity of the route is quantified, and a standard route with moderate difficulty is selected to avoid the impact of the ceiling or floor effect on the test results. Secondly, a method for establishing a habitual route memory bank based on the convergence of GPS trajectory and map road network matching is proposed. By calculating the coefficient of variation (CV) of GPS lateral matching deviation in multiple route memory trips for each road segment, the map matching stability of each road segment is quantitatively determined. The method terminates when the proportion of convergent road segments in the whole route reaches a threshold, replacing the traditional fixed number of memory trips. Third, a random rotation test of the reference vehicle was designed to solve the problem of adaptive energy consumption drift caused by the reference vehicle driving multiple times, and to ensure the stability of the baseline energy consumption. Fourth, in the navigation-free verification phase, a real-time data quality monitoring and hierarchical correction method based on road segment clustering and outlier rate classification is proposed: Test road segments are grouped according to sensitivity, allowing for consistency verification within each group, thus solving the problem of outlier comparisons at the same scale due to inherent differences in energy-saving characteristics among different road types; an outlier rate classification processing system is established, achieving multi-level fine-grained processing at the road segment level; and a correction effectiveness verification method is proposed to ensure the reliability of the corrected data. Then, based on the energy-saving rate in each round, the sliding window range and mean difference are calculated. When the energy-saving rate range for three consecutive rounds is less than or equal to a certain value (e.g., ≤0.2%) and the difference from the mean of the previous three rounds is less than or equal to a certain value (e.g., ≤0.2%), the vehicle's energy-saving effect is considered to have entered a stable state, and the verification test is terminated. If the significance threshold is not reached, the PEM system is deemed to have excessively fluctuating energy-saving effects in actual use, lacking stable and reliable practical value, and the test is terminated, thus avoiding invalid testing. Furthermore, a generalization verification is introduced, applying three types of disturbances: traffic flow, weather, and load, and calculating the ratio of the energy saving rate after the disturbance to the standard energy saving rate. If any ratio is less than 0.7, the system is considered to be overfitting, and the test is terminated. Otherwise, driving tests are conducted on M-wheel test vehicles and reference vehicles in a road scenario without navigation, and the results are compared. The conservatively estimated energy saving rate is calculated using the confidence interval limit values ​​at 90% confidence level (upper limit for test vehicles, lower limit for reference vehicles) to prevent exaggeration of energy saving effects due to accidental factors. Finally, the energy and fuel consumption are converted into carbon emissions to accurately evaluate their energy-saving effect in actual use.

[0044] Specifically, Figure 1 This is a flowchart illustrating a method for testing the energy-saving effect of a predictive energy management system based on route memory, provided in an embodiment of this application.

[0045] like Figure 1 As shown, the energy-saving effect testing method of the predictive energy management system based on route memory includes the following steps: In step S101, the road segment features of multiple road segments in multiple candidate test routes are obtained, and the information entropy corresponding to the multiple candidate test routes is calculated based on the road segment features, so as to select the target test route from the multiple candidate test routes based on the information entropy.

[0046] In some embodiments, in order to ensure that the testing process can cover various situations in the actual driving process of the vehicle, this application can first calculate the information entropy of multiple candidate routes based on the road segment characteristics of multiple road segments in multiple candidate test routes, so as to select a route that can be used for actual testing based on the information entropy of multiple candidate routes.

[0047] In this embodiment, the information entropy used is Shannon entropy. Therefore, a "test route" can be considered an "information source," and the distribution of various features along the route (such as gradient, speed limit, etc.) can be considered "information." The more uniform and diverse the feature distribution of a route, the higher its "uncertainty," the larger its Shannon entropy value, and the greater the "challenge" it poses to the PEM system.

[0048] By calculating information entropy, the embodiments of this application can concretize the vague concept of route complexity for each candidate test route into a clear and comparable value, thereby objectively selecting target test routes with complexity within a suitable range (e.g., normalized information entropy between 0.6 and 0.8).

[0049] It should be noted that the specific appropriate range can be set and adjusted by those skilled in the art according to the actual situation, such as (0.6~0.65), (0.7~0.8), etc. The embodiments in this application are only illustrative and are not specifically limited.

[0050] Here, the target test route can be understood as the test route that meets the testing requirements after screening multiple candidate test routes. For example, it may have multiple road segments, and these segments may be evenly distributed throughout the entire test route.

[0051] The selected target test routes generally have the following road characteristics: 20%–30% of the mileage is urban roads (speed limit ≤ 40 km / h), 35%–50% is suburban roads (speed limit 41–80 km / h), and 25%–40% is highways (speed limit > 80 km / h); in terms of gradient, it includes uphill sections of no less than 3% and downhill sections of no less than 3%; in terms of traffic flow, sections that are not completely unobstructed (with speed fluctuations or congestion) account for no less than 60%. It should be noted that in actual application scenarios, the specific road characteristics of the selected roads may slightly differ from the example numerical ranges in this application embodiment. This application embodiment is only for illustrative purposes and does not impose specific limitations.

[0052] Therefore, the target test route is neither a "simple route" that is easy to achieve optimal control, nor a "complex route" that frequently encounters extreme conditions, thus effectively avoiding the ceiling and floor effects in the evaluation.

[0053] Optionally, in one embodiment of this application, obtaining the segment features of multiple road segments in multiple candidate test routes and calculating the information entropy corresponding to the multiple candidate test routes based on the segment features includes the following steps: obtaining the slope features, speed limit features, and historical traffic flow features of multiple road segments in multiple candidate test routes, and classifying the slope features, speed limit features, and historical traffic flow features into multiple levels; statistically analyzing the proportion of road segments corresponding to each level of slope features, speed limit features, and historical traffic flow features to obtain the probability distribution corresponding to slope features, speed limit features, and historical traffic flow features; and calculating the information entropy corresponding to the multiple candidate test routes based on the probability distribution.

[0054] In actual implementation, the road segment characteristics of multiple road segments of each candidate test route obtained in this application include, but are not limited to, slope characteristics, speed limit characteristics, historical traffic flow characteristics, etc.

[0055] For example, this application may first select candidate test routes from a unified test route library. The total length of the test routes is 90 to 110 km, and they are set in the order of urban roads, suburban roads, expressways, suburban roads and urban roads, and include sloping road sections.

[0056] Then, in this embodiment of the application, the test route can be divided into several basic road segments (e.g., divided into segments of 1-5km in length), and multiple road segment features can be extracted for each segment. For example, three types of road segment features can be extracted for each segment: slope features, speed limit features, and historical traffic flow features.

[0057] The specific road segment features extracted can be set and adjusted by those skilled in the art according to actual conditions and needs. For example, road segment curvature features, traffic light density features, etc. can be added on the basis of slope features, speed limit features, and historical traffic flow features. The embodiments in this application are only illustrative and do not impose specific limitations.

[0058] Then, in this embodiment of the application, the slope feature, speed limit feature and historical traffic flow feature can be divided into multiple levels, and the proportion of road segments in each level of these features can be counted to calculate the probability distribution corresponding to the slope feature, speed limit feature and historical traffic flow feature, and then the information entropy corresponding to multiple candidate test routes can be calculated using the probability distribution.

[0059] For example, this application can classify slope characteristics, speed limit characteristics, and historical traffic flow characteristics into the following multiple levels: (1) Slope characteristics: Based on the average slope (%), it is divided into 4 levels: <-3% (downhill), -3%-0% (gentle downhill), 0%-3% (flat road / gentle uphill), >3% (uphill).

[0060] (2) Speed ​​limit characteristics: Based on the maximum speed limit (km / h) of the road section, it is divided into 3 levels: ≤40 (low speed), 41-80 (medium speed), and >80 (high speed).

[0061] (3) Historical traffic flow characteristics: Based on the historical average vehicle speed fluctuation coefficient (standard deviation / mean, etc.) provided by the map service provider for this road segment, it is divided into 3 levels: <0.1 (stable), 0.1-0.3 (fluctuation), >0.3 (congestion).

[0062] The specific grading criteria used for features such as slope characteristics, speed limit characteristics, and historical traffic flow characteristics can be set and adjusted by professionals in this field according to actual conditions. The embodiments in this application are only illustrative and do not impose specific limitations.

[0063] Next, in this embodiment of the application, the proportion of road segments at each level of each road segment characteristic can be statistically analyzed to obtain a probability distribution. Calculate the information entropy of each road segment feature on the candidate test route. :

[0064] Information entropy of all road segment features The arithmetic mean is taken to obtain the comprehensive information entropy of each candidate test route. And normalize it as follows:

[0065] in, The maximum number of levels to be assigned to the characteristics of each road segment (taken as an example, 4); This is the comprehensive information entropy of the test route.

[0066] The higher the normalized information entropy value of a route, the greater the challenge to the predictive energy management system's memory and energy-saving capabilities. Therefore, in this embodiment, a route with a normalized information entropy at a moderate level (e.g., 0.6-0.8) can be selected as the target test route to avoid the ceiling effect or floor effect caused by extreme routes.

[0067] This application embodiment can quantify the road conditions of each candidate test route into information entropy. By selecting a test route with an appropriate level of information entropy as the target test route, it avoids the inability to effectively distinguish the true capabilities of the PEM system due to routes that are too complex or too simple, so that the final test results can truly reflect the performance differences of the PEM system.

[0068] Step S102: If the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, based on the target test route, obtain the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition.

[0069] In other embodiments, after selecting a target test route from multiple candidate test routes based on the information entropy of multiple candidate test routes, this application may use the target test route for testing.

[0070] Furthermore, in order to ensure the accuracy of the energy-saving test results of the test vehicle, the embodiments of this application can detect the route memory status and energy-saving status of the predictive energy management system of the test vehicle. Then, if the route memory status of the predictive energy management system of the test vehicle meets the preset requirements and the energy-saving effect of the predictive energy management system of the test vehicle meets the preset energy-saving conditions, a formal multi-round following test can be conducted on the test vehicle and the reference vehicle.

[0071] In this context, "route memory" refers to the test vehicle's memory of the target test route. The preset requirements can be understood as the standards / requirements that the route memory in the vehicle must meet before a test can be conducted. For example, the test vehicle must have over 90% valid route memory of the target test route.

[0072] Energy-saving effect here can be understood as the energy-saving effect of the predictive energy management system of the test vehicle. Preset energy-saving conditions here can be understood as the pre-set standards / requirements that the predictive energy management system of the test vehicle must meet to determine whether the test vehicle can be tested. For example, the energy-saving effect of the predictive energy management system of the test vehicle has reached its optimal state, meaning that the energy-saving rate of the predictive energy management system of the test vehicle cannot be significantly improved further.

[0073] Specific preset requirements and preset energy-saving effects can be set by those skilled in the art based on actual conditions and needs. The embodiments in this application are for illustrative purposes only and are not intended to impose specific limitations.

[0074] Then, in this embodiment of the application, if the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the test vehicle under the first preset operating condition and the reference vehicle under the second preset operating condition shall conduct multiple rounds of following tests on the target test route to obtain the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition to calculate the energy-saving rate of the PEM system in the test vehicle.

[0075] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, the method further includes: determining that the test vehicle is under the first preset operating condition when the navigation condition of the test vehicle is off and the predicted energy management system condition is on; and determining that the reference vehicle is under the second preset operating condition when the navigation condition of the reference vehicle is off and the predicted energy management system condition is off.

[0076] Based on the relevant circumstances of other embodiments, it is understood that this application can perform the test when the test vehicle is in a first preset working condition and the reference vehicle is in a second preset working condition.

[0077] In actual implementation, this application allows the test vehicle and the reference vehicle to be tested under navigation conditions and without navigation conditions, respectively.

[0078] Specifically, in the embodiments of this application, when the test vehicle is in a first preset operating condition with navigation off but the predictive energy management system on, the test vehicle can be determined to be in a first preset operating condition. That is, the first preset operating condition can be understood as a combined operating condition where the navigation of the test vehicle is off and the predictive energy management system is on.

[0079] Similarly, in the embodiments of this application, when the reference vehicle is in a state where navigation is not turned on and the predictive energy management system is also turned off, it can be determined that the reference vehicle is in a second preset operating condition. That is, the second preset operating condition can be understood as a combined operating condition where the reference vehicle's navigation is off and the predictive energy management system is also off.

[0080] This application embodiment can conduct tests on a target test route when both the test vehicle and the reference vehicle have navigation turned off. It simulates a real scenario where the user does not use navigation but the vehicle can still activate PEM through route memory. This enables road testing of the energy-saving effect of the predictive energy management strategy that does not rely on navigation activation. It provides a scientific testing and evaluation method for PEM strategies based on route memory, provides a theoretical basis for real road testing under navigation-free conditions, and makes up for the shortcomings of related technologies. The testing methods are all for PEM strategies that require navigation activation and are indoor drum simulation tests, lacking practical testing methods for the energy-saving effect of PEM strategies without navigation activation.

[0081] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, when the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the method further includes: determining multiple test sampling points for multiple road segments corresponding to each set of driving data based on multiple sets of driving data of the test vehicle on the target test route; calculating the matching deviation between the multiple test sampling points and the target centerline, so as to calculate the trajectory matching quality between the multiple road segments and the target test route based on the matching deviation; and determining the proportion of road segments that have reached the convergence state in the target test route based on the trajectory matching quality, so as to determine the route memory of the predictive energy management system of the test vehicle based on the proportion of road segments.

[0082] As one possible approach, in order to ensure that the test vehicle has the route memory that a general vehicle can activate the PEM strategy when driving on the road during the formal test, the embodiments of this application can first have the test vehicle drive multiple times on the target test route, and determine the route memory status of the test vehicle based on the multiple driving data (for ease of distinction, this stage can be simply referred to as the route memory stage).

[0083] Specifically, this application can first have the test vehicle drive multiple times on the target test route to obtain multiple sets of driving data, and then determine multiple test sampling points for multiple road segments corresponding to each set of driving data.

[0084] For example, in this application embodiment, the test vehicle can drive along the target test route multiple times with navigation enabled, and a minimum number of memory attempts m_min=3 times is set to ensure that the test vehicle has sufficient trajectory samples for each segment of the target test route.

[0085] Then, in this embodiment of the application, the target test route can be divided into several basic road segments (such as 1km) that are the same as those in the information entropy calculation process, and in each memory drive, for each road segment, all GPS sampling points (i.e. test sampling points) of the test vehicle in that road segment are recorded.

[0086] Next, embodiments of this application can calculate the matching deviation between these test sampling points and the target centerline, so as to calculate the trajectory matching quality between multiple road segments and the target test route based on the matching deviation.

[0087] For each sampling point, this embodiment of the application can calculate its lateral matching deviation from the target centerline, that is, the vertical distance from the GPS measured position of each sampling point to the centerline of the map road network. Here, the target centerline refers to the center of the map road network, that is, the "central axis" of the current road.

[0088] Then, the root mean square of the lateral matching deviation can be used in the embodiments of this application. Characterizes the trajectory matching quality between road segment k and the target test route during the j-th trip:

[0089] Where N is the total number of sampling points in this road segment, The RMS value represents the vertical distance from the measured GPS location of each sampling point to the center line of the map road network. This indicator comprehensively reflects the matching accuracy between GPS positioning and the map road network during this trip. The larger the RMS value, the more severe the interference with the GPS signal (such as obstruction by overpasses), and the less stable the map matching.

[0090] Next, the embodiments of this application can determine the sufficiency of the route memory of the test vehicle, that is, determine whether the route memory of the predictive energy management system of the test vehicle meets the preset requirements.

[0091] Specifically, in this application embodiment, the proportion of road segments that have reached the convergence state in the target test route can be determined based on the calculated trajectory matching quality, and then the route memory status of the predictive energy management system of the test vehicle can be determined based on the proportion of road segments, so as to determine whether the route memory of the predictive energy management system of the test vehicle meets the preset requirements.

[0092] For example, after the test vehicle completes m route memory drives (m ≥ m_min), this application can calculate the coefficient of variation of the RMS value of each road segment k over all m drives. :

[0093] in, ( The ) represents the trajectory matching quality between road segment k and the target test route during the j-th trip. for The arithmetic mean; for The standard deviation of . When According to the embodiments of this application, it can be determined that the GPS trajectory and map matching of the road segment have reached a stable convergence state, that is, the road segment has been fully memorized, which means that the predictive energy management system of the test vehicle already has the route memory corresponding to the road segment.

[0094] The specific method used to determine whether the GPS trajectory and map matching of a road segment has reached a stable convergence state. The value can be set by those skilled in the art based on actual conditions and needs. The embodiments in this application are for illustrative purposes only and are not intended to impose specific limitations.

[0095] Next, the embodiments of this application can statistically analyze the percentage of road segments that have reached convergence along the entire test route. :

[0096] in, The number of road segments required to reach a convergent state; This represents the total number of road segments on the test route.

[0097] when When the established route memory database is complete, the test vehicle is considered to have completed its memory journey; otherwise, a new memory journey is added, and the data is collected and recalculated. It should be noted that the test vehicle still needs to have navigation enabled during the added memory journey, and the time should be within the same time period of the day as the previous memory journey. That is, all memory journeys should be conducted within the same time period each day to ensure consistency. If those in this technical field have generalization needs, they can adjust the conditions for each memory journey appropriately according to their own requirements.

[0098] The route memory termination condition can be set to, but is not limited to, reaching the maximum number of times the route can be memorized, i.e., the maximum number of trips. For example, the maximum number of times the route can be memorized can be set to m_max = 8 times. If the route memory is not completed when m_max is reached, it is determined that the test route is unsuitable as a standard test route for the PEM system of the test vehicle due to severe GPS signal obstruction (such as a high proportion of elevated underpasses, tunnels, etc.), and the test route should be returned to select a new target test route.

[0099] After this step is completed, at least 90% of the route memory in the test vehicle's established route memory has passed the GPS trajectory matching convergence verification, ensuring that the PEM system can be effectively activated when there is no subsequent navigation call, and that the system's positioning and route recognition are highly reliable.

[0100] This application embodiment can determine the route memory establishment status of the PEM system of the test vehicle based on the convergence of GPS trajectory matching, replacing the empirical approach of establishing habitual routes with a fixed number of trips. By using the coefficient of variation of GPS lateral matching deviation to quantify the stability of map matching for each road segment, closed-loop control of the route memory establishment process can be achieved, thereby significantly reducing the risk of subsequent navigation failure or poor quality due to insufficient quality of the habitual route memory database.

[0101] Optionally, in one embodiment of this application, before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, when the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the method further includes: obtaining the multi-round test results of the test vehicle under the navigation condition of being off and the predictive energy management system under the operating condition of being on, and the multi-round test results of the reference vehicle under the navigation condition of being off and the predictive energy management system under the operating condition of being off, based on the target test route; obtaining the number of outlier road segments of the test vehicle in each round of testing based on the multi-round test results of the test vehicle and the reference vehicle, so as to determine the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments.

[0102] In some embodiments, considering that vehicles can activate the PEM strategy based on route memory, in order to effectively verify the energy-saving effect of the vehicle's PEM strategy activated without relying on navigation, this application can conduct multiple rounds of testing on both the test vehicle and the reference vehicle under the condition of navigation being turned off before the formal test (this stage can be called the verification test stage to distinguish it from the subsequent formal test). Based on the test results of the multiple rounds of verification tests, it can be determined whether the energy-saving effect of the test vehicle has reached the optimal state. When the energy-saving effect of the test vehicle reaches the optimal state, the formal test will then be conducted.

[0103] Specifically, the embodiments of this application can obtain in real time the number of outlier road segments in multiple rounds of verification tests before the formal test for the test vehicle and the reference vehicle. Based on the number of outlier road segments, the reliability of the data of the test vehicle in each round of verification test is determined. Then, when the number of outlier road segments meets certain conditions, the energy-saving effect of the predictive energy management system of the test vehicle is determined, that is, whether the energy-saving effect of the test vehicle has reached the optimal level is determined.

[0104] In this context, outlier road sections can be understood as road sections where, during multiple rounds of testing, the corresponding data of the vehicles (test vehicles and reference vehicles) deviates significantly from the baseline range that the normal data should be.

[0105] It should be noted that in the multiple rounds of verification tests before the formal test, the navigation function of the test vehicle must be turned off throughout and the predictive energy management system must be turned on. Similarly, the navigation function of the reference vehicle must be turned off, but the predictive energy management system must also be turned off.

[0106] Furthermore, in the formal testing and verification testing of both the test vehicle and the reference vehicle, multiple rounds of testing (i.e., multiple rounds of formal testing and multiple rounds of verification testing) are required, such as 3 rounds, 5 rounds, etc., and all of them adopt the following test of rotating reference vehicles, that is, multiple reference vehicles are set, such as 2 vehicles, 3 vehicles, etc.

[0107] The specific number of test rounds and the number of reference vehicles can be set and adjusted by professionals in this field according to actual conditions and needs. The embodiments in this application are only illustrative and do not impose specific limitations.

[0108] By using multiple reference vehicles for wheel rotation and the constraint that the same vehicle cannot be selected for two consecutive wheels, the embodiments of this application can effectively eliminate the energy consumption benchmark drift caused by the adaptive learning effect of the reference vehicles, and ensure the accuracy of energy saving rate calculation.

[0109] For example, this application may prepare one test vehicle V_test (with PEM function enabled) and at least three reference vehicles V_ref1, V_ref2, and V_ref3 of the same model and configuration (all reference vehicles have PEM function permanently disabled or disabled via a diagnostic tool). The test vehicle V_test remains unchanged throughout the entire testing process.

[0110] Before the start of each round of verification testing (round i), one reference vehicle can be randomly selected from the reference vehicles to be paired with the test vehicle V_test in this round, denoted as V_ref,i. The same reference vehicle cannot be selected in two consecutive rounds (for example, if V_ref1 is selected in round i, then in round i+1, it can only be selected from V_ref2, V_ref3, etc.), to prevent a single reference vehicle from accumulating adaptive learning effects due to continuous use, and to avoid baseline drift caused by its energy consumption decreasing with each iteration.

[0111] The test vehicle V_test participated in each round of verification testing, completing the entire test route for rounds 1, 2, 3, and so on. In each round of verification testing, V_test had navigation turned off throughout. The PEM system used GPS for real-time positioning and matched it with the habitual route memory stored in S2. After a successful match, the PEM system autonomously activated and executed the predictive energy management strategy based on the pre-stored static feature data of the route in the memory bank and the real-time traffic information pushed from the cloud.

[0112] Furthermore, the habitual route memory of V_test has been established, and no new substantial memory updates occur in subsequent rounds of testing (i.e., the route memory content remains unchanged), to ensure that each round of testing is conducted under the same baseline memory state, thereby guaranteeing the comparability of test results.

[0113] This application embodiment allows for random rotation testing of reference vehicles, resolving the baseline drift problem caused by adaptive learning of the reference vehicles. By rotating multiple reference vehicles and using discontinuous usage rules, the energy consumption of the reference vehicles can be ensured to remain stable, preventing artificially inflated or understated energy-saving rates, thereby significantly improving the reliability of the test results. Through multiple rounds of repeated driving, this application embodiment can effectively verify whether the PEM system can continuously and stably identify the preferred route and reproduce its energy-saving effect in multiple independent runs.

[0114] Optionally, in one embodiment of this application, determining the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments includes: dividing multiple road segments into multiple road segment groups based on the road segment characteristics of multiple road segments of the target test route; calculating an initial baseline mean and an initial baseline standard deviation based on the energy-saving rate of the road segments preset in each road segment group by the predictive energy management system of the test vehicle; identifying outlier road segments in other road segments based on the initial baseline mean, the initial baseline standard deviation, and the energy-saving rate of other road segments in each road segment group (excluding the preset road segments), and identifying outlier road segments in the preset road segments when the number of non-outlier road segments in other road segments meets a preset number; determining the total number of real-time outlier road segments of the test vehicle during the test process based on the outlier road segments in other road segments and the outlier road segments in the preset road segments; calculating the real-time cumulative outlier rate of the test vehicle during multiple rounds of testing based on the total number of real-time outlier road segments, so as to determine the energy-saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate.

[0115] As one possible approach, this application proposes a real-time data quality monitoring and classification correction method based on road segment clustering and outlier rate classification when determining the energy-saving effect of the predictive energy management system of the test vehicle. This method can be implemented, but is not limited to, the following steps: First, a memory clearing operation is performed on all vehicles (including test vehicles and all reference vehicles) before the test begins. The specific method is as follows: Execute the manufacturer's standardized memory clearing command via the OBD diagnostic interface (clearing historical route GPS tracks, energy management strategy learning values, etc.), and then drive a "demagnetizing section" unrelated to the test route (e.g., randomly driving in a closed plaza for 10 minutes, including non-repetitive actions such as acceleration, deceleration, and steering) to clear all temporarily stored adaptive data. For the test vehicle V_test, only its non-route memory adaptive data (such as short-term driving habit learning values) is cleared, while the established habitual route memory remains unchanged.

[0116] For hybrid vehicles, before testing, ensure that the state of charge (SOC) of both batteries is at least 10% above their respective battery balance points. After fully warming up the vehicles, fill both vehicles with fuel using the same fuel nozzle until the nozzle clicks off. Begin testing immediately after refueling. During the test, the reference vehicle V_ref,i drives in front of the test vehicle V_test. Both vehicles maintain the same load, tire specifications, and tire pressure. Onboard accessories are either disabled or kept in the same state. Fuel consumption data should be recorded using an onboard fuel consumption meter (with fuel nozzle data used as auxiliary verification), and SOC changes should also be recorded.

[0117] Second, based on the road segment characteristics of multiple road segments of the target test route, multiple road segments are divided into multiple road segment groups: that is, considering that the actual scenarios of different road segments are different, in order to effectively determine whether a certain road segment belongs to an outlier, the embodiments of this application can first divide road segments with similar characteristics into a road segment group according to the road segment characteristics of multiple road segments, so that each road segment can be checked for consistency within the group, and only the data within the road segment group is used to determine whether the road segment within the road segment group belongs to an outlier, ensuring the authenticity and reliability of the outlier determination, and solving the problem that different road types of road segments cannot be compared on the same scale due to the inherent differences in energy-saving characteristics.

[0118] For example, this application can divide all road segments into several road segment groups based on the extracted feature parameters such as slope characteristics, speed limit characteristics and historical traffic flow characteristics of each road segment, using the K-means clustering algorithm, so that the road segments in the same road segment group have statistical similarity in terms of energy saving sensitivity of the PEM system, and then perform outlier determination on the road segments in each road segment group.

[0119] Third, based on the energy-saving rates of the test vehicles and reference vehicles in each road segment group, the initial baseline mean and initial baseline standard deviation are initially set to determine whether other road segments besides the preset road segments belong to outliers.

[0120] The energy-saving rate of each road segment, including the preset road segments and other road segments, is calculated in real time.

[0121] Furthermore, the preset road segments can be understood as one or more road segments pre-set in each road segment group and used when calculating the initial baseline mean and initial baseline standard deviation. For example, the first three road segments or the first five road segments in each road segment group may be used. The specific preset road segments can be set or adjusted by those skilled in the art according to the actual situation. The embodiments in this application are only illustrative and do not impose specific limitations.

[0122] IV. Based on the initial baseline mean, initial baseline standard deviation, and energy saving rate of other road segments in each road segment group (excluding the preset road segments), identify outlier road segments in other road segments, and identify outlier road segments in the preset road segments when the number of non-outlier road segments in other road segments meets the preset number.

[0123] For example, this application may use the first three road segments in each road segment group as preset road segments, and then perform the following process: (1) First, do not make outlier determinations for the first 3 road segments, and temporarily accept and record the energy-saving rate data of the first 3 road segments. Among them, the energy-saving rate of each road segment The calculation formula can be expressed, but is not limited to, as follows:

[0124] in, and These represent the combined carbon emission equivalent per 100 kilometers for the reference vehicle and the test vehicle on the k-th road segment, respectively.

[0125] Based on the energy-saving rate data of the first three road sections, the initial baseline mean was calculated. and initial baseline standard deviation :

[0126]

[0127] (2) Starting from the 4th road segment, use the current baseline average. and the current benchmark standard deviation Outlier detection is performed. The energy-saving rate of each newly added road segment is calculated in real time, and the road segment is marked as an outlier when the energy-saving rate of the newly added road segment k satisfies the following formula:

[0128] If the road segment (every road segment starting from the 4th road segment) is a normal road segment, it is included in the baseline (baseline mean and baseline standard deviation) update; outlier road segments are not included in the baseline update.

[0129] It should be noted that in this application's embodiments, "starting from the 4th road segment, the current benchmark average is used" means... and the current benchmark standard deviation "Outlier detection" here refers to starting with four road segments. The baseline mean and standard deviation used for each road segment are actually derived from the baseline mean and standard deviation of the road segments preceding that segment, and so on. When a road segment is determined to be an outlier, the data for that outlier segment is not included in the update calculation of the baseline mean and standard deviation.

[0130] Therefore, for the fourth road segment, the current benchmark average is used. and the current benchmark standard deviation The baseline mean and baseline standard deviation are calculated from the data of the first three road segments. For the fifth road segment, the current baseline mean is used. and the current benchmark standard deviation The baseline mean and baseline standard deviation are calculated from the data of the first four road segments... and so on, until all road segments in the group have been identified as outliers.

[0131] (3) When the number of normal road segments accumulated in a certain road segment group reaches a preset number (e.g., 10, which can be set and adjusted by professionals in this field according to the actual situation. This application embodiment is only an example and does not impose any specific restrictions), the first 3 road segments are checked back to identify the outlier road segments in the first 3 road segments.

[0132] First, the median of normal road segments within the group of this road segment. Replace current mean ; With median absolute deviation Alternate standard deviation :

[0133] Then, outlier detection was performed on the energy-saving rate data of the first three road sections:

[0134] If there are outliers in the first 3 road segments, the data will be corrected to "outlier" and the overall energy saving rate of that round will be recalculated; if the first 3 road segments are all normal, the original judgment result will be maintained.

[0135] (4) Road segments marked as outliers will not participate in subsequent baseline updates in this round, but the original data will be retained for final correction processing.

[0136] Fifth, based on outliers in other road segments and outliers in preset road segments, determine the real-time number of outliers on the target test route during the test process. That is, in this embodiment of the application, after each outlier is identified, the total number of outliers in the verification test can be counted in real time, which is referred to here as the real-time outlier number.

[0137] Corresponding to the real-time calculation of the energy-saving rate of each road segment, after the real-time calculation of the energy-saving rate of each road segment in each road segment group and the determination of whether each road segment is an outlier, the total number of outlier road segments up to the current time can be calculated in real time, so as to calculate the real-time cumulative outlier rate of the target test route in each round of testing based on the total number of outlier road segments in real time.

[0138] For example, during each round of testing, whenever a basic road segment is tested and outlier determination is completed, this application can immediately update the cumulative number of outlier road segments for that round. This refers to the total number of outbound road segments in real time, and the real-time cumulative outbound rate is calculated after obtaining the total number of outbound road segments in real time. (Current road segment as of the current moment):

[0139] in, This represents the total number of road segments (100) on the test route. This is the cumulative number of road segments marked as outliers from the start of this round to the current road segment.

[0140] The following section provides a further explanation of how the energy-saving effect of the predictive energy management system is determined based on the real-time cumulative outlier rate of the test vehicle and the reference vehicle in the embodiments of this application.

[0141] Optionally, in one embodiment of this application, determining the energy-saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate includes: calculating the energy-saving rate of the predictive energy management system of the test vehicle based on the test results when the real-time cumulative outlier rate meets the preset outlier condition; calculating the range of energy-saving rates of the test vehicle in the first target round of testing based on the energy-saving rate, and calculating the absolute value of the difference between the average energy-saving rate of the first target round of testing and the average energy-saving rate of the second target round of testing; and determining the energy-saving effect of the predictive energy management system of the test vehicle based on the range and absolute value of the energy-saving rate.

[0142] In other embodiments, after calculating the real-time cumulative outlier rate, this application can calculate the energy-saving rate of the test vehicle when the real-time cumulative outlier rate meets the preset outlier condition.

[0143] Here, the preset outlier condition can be understood as a pre-set condition / requirement that the real-time cumulative outlier rate must meet when determining the test data used to calculate the energy-saving rate of the test vehicle. For example, the real-time cumulative outlier rate needs to be within a certain range to determine the validity of the data in this round of verification test. The specific preset outlier condition can be set or adjusted by those skilled in the art according to the actual situation. The embodiments in this application are only illustrative and do not impose specific limitations.

[0144] For example, embodiments of this application can be based on the real-time cumulative outlier rate. The following exception handling is performed on the values: (1) When hour: The test data for this round is determined to be severely abnormal, and the data for this round is unreliable as a whole. This round of testing is to be terminated immediately, all data is to be discarded, and this round of testing must be re-executed (using the same vehicle as specified in this round). To ensure weather consistency, the test can be conducted the following day.

[0145] (2) When hour: The test data for this round was found to contain local anomalies, and data correction was performed: when In this process, a weighted average method is used, where normal road segments are assigned a weight of 1.0 and outlier road segments a weight of 0.3. The overall energy-saving rate for that round is then calculated using a weighted average. In actual implementation, the weights of normal and outlier road segments can be set and adjusted by those skilled in the art based on the actual situation. The embodiments in this application are merely illustrative and do not impose specific limitations.

[0146] when When the outlier road segment is replaced by the arithmetic mean of the energy-saving rates of all normal road segments in the same group and in the same round, the in-group mean interpolation method is used.

[0147] After the correction is completed, calculate the change in the overall energy saving rate before and after the correction:

[0148] in, , These represent the energy saving rate of the entire test after correction and the energy saving rate of the entire test before correction, respectively.

[0149] like (The 0.2% here refers to the difference between the values ​​before and after correction, not the measurement accuracy.) If the difference before and after correction is less than 0.2%, it means that the quality of the test data in this round is good enough and the corrected data can be used as the valid data for this round.

[0150] like If the value is too large, it indicates that the correction will significantly change the energy saving rate, which means that the impact of abnormal road sections in the original data cannot be reliably eliminated by simple correction. Therefore, the entire round of data should be discarded and retested.

[0151] (3) When hour: The test data for this round is deemed to be of acceptable quality. The number of outlier road sections is within the allowable range for normal noise measurement, requiring no correction; the data from this round is directly considered valid.

[0152] Once the data passes the validity determination, this application embodiment can use the energy consumption data in the multi-round test results of the reference vehicle V_ref,i as the benchmark for each round of testing to calculate the energy consumption and energy-saving effect of the test vehicle in each round of driving.

[0153] It should be noted that the specific numerical value or range used when determining whether the real-time cumulative outlier rate meets the preset outlier criteria is, for example, , , , The percentages (3%, 10%), (5%, 10%), (3%, 5%), 3%, 5%, 10%, 0.2%, etc., can be adapted by those skilled in the art according to actual needs. The embodiments in this application are merely illustrative and do not impose specific limitations. It is recommended to prefer the example values / ranges in the embodiments of this application.

[0154] After each round of testing, the combined carbon emissions per 100 kilometers for the test vehicle V_test and the reference vehicle V_ref,i are first calculated: Obtain the fuel consumption and electricity consumption per 100 kilometers for the test vehicle V_test and the reference vehicle V_ref,i, and then convert them into carbon emissions per 100 kilometers: Fuel carbon emissions = fuel consumption per 100 kilometers × fuel carbon emission factor; Carbon emissions from electricity = electricity consumption per 100 kilometers × carbon emission factor for electricity; Total carbon emissions = carbon emissions from fuel oil + carbon emissions from electricity.

[0155] The carbon emission factor for fuel oil can be, but is not limited to, the value provided in national standards, such as the K_CO2 conversion coefficient in GB 19578-2024, and the carbon emission factor for electricity can be, but is not limited to, the national average CO2 emission factor for electricity. The specific values ​​for the carbon emission factors for fuel oil and electricity can be set and adjusted by those skilled in the art according to actual circumstances. The embodiments in this application are merely illustrative and do not impose specific limitations.

[0156] Then, the fuel efficiency of the test vehicle in this round of testing is calculated. :

[0157] in, and These represent the combined carbon emissions per 100 kilometers for the reference vehicle and the test vehicle, respectively, during this round of testing.

[0158] Based on the energy saving rate of each test round, the embodiments of this application can calculate the range of energy saving rates of the test vehicle in the first target test round, and calculate the absolute value of the difference between the average energy saving rate of the first target test round and the average energy saving rate of the second target test round, so as to determine the energy saving effect of the predictive energy management system of the test vehicle.

[0159] Here, the first target round and the second target round can be understood as one or several rounds of testing in a pre-set multi-round test of the vehicle.

[0160] For example, this application may perform the following calculations based on the energy saving rate from multiple rounds of testing: (1) The energy saving rate of the most recent three consecutive rounds (rounds i-2, i-1, and i) is extremely poor: , in, These represent the energy-saving rates for the (i-2), (i-1), and (i-1)th rounds of testing, respectively.

[0161] (2) The arithmetic mean of the energy saving rates of the most recent three rounds The arithmetic mean of the previous 3 rounds (rounds i-5, i-4, and i-3) The absolute value of the difference:

[0162] when and When the predictive energy management system of the test vehicle has reached a stable output state, the verification test can be terminated, and the current round N is recorded. The criteria used for this determination are as follows: and The embodiments described herein can be adapted by those skilled in the art to meet actual needs. These embodiments are merely illustrative and do not constitute specific limitations. It is recommended that the examples in these embodiments be preferred.

[0163] If the above conditions are not met for 10 consecutive rounds, the PEM system of the test vehicle is deemed to lack stable application capability and has poor reliability, and the test is terminated.

[0164] This application's embodiments employ a statistical criterion of sliding window range and mean difference to replace the traditional fixed-round setting or subjective judgment by testers, establishing a unified testing starting point for evaluating the energy-saving effect of navigation-free PEM systems. Simultaneously, this statistical criterion of sliding window range and mean difference can provide an effective criterion for terminating testing when some systems may exhibit continuous oscillations in energy-saving rates that fail to converge, thereby avoiding the waste of testing resources on ineffective systems.

[0165] Once the energy-saving effect of the PEM system of the test vehicle reaches a stable output state, it can be determined that the energy-saving effect of the PEM system of the test vehicle meets the preset energy-saving conditions. Combined with the fact that the test vehicle has route memory of the target test route, it can be determined that the test vehicle can carry out formal testing.

[0166] This application's embodiments address the issue of outlier comparisons at the same scale due to inherent differences in energy-saving characteristics among road segments of different types by clustering test road segments according to their sensitivity. Furthermore, this application's embodiments establish a hierarchical outlier rate processing system, achieving a multi-level refined processing upgrade from the "round level" to the "road segment level." A correction validity verification method is also proposed to ensure the reliability of the corrected data. This enables real-time quality control of the testing process, effectively avoiding the risk of entire rounds of data being discarded or erroneous data being used in termination decisions due to local anomalies.

[0167] Optionally, in one embodiment of this application, obtaining the first carbon emission of a test vehicle under a first preset operating condition and the second carbon emission of a reference vehicle under a second preset operating condition based on a target test route further includes: obtaining multi-type disturbance test data of the test vehicle to calculate the energy saving rate of the predictive energy management system of the test vehicle under multi-type disturbance tests; calculating multi-type generalization indicators of the test vehicle based on the energy saving rate under multi-type disturbance tests; and obtaining the first carbon emission and the second carbon emission based on the target test route when the multi-type generalization indicators meet the corresponding preset indicator range.

[0168] In some embodiments, after the verification test is completed, the present application may also conduct disturbance tests (such as traffic flow disturbances, weather disturbances, load disturbances, etc.) on the test vehicle to ensure the robustness of the test vehicle's PEM system.

[0169] Each type of disturbance test can be repeated a certain number of times, such as 2-3 times. The specific number of repetitions can be set and adjusted by those skilled in the art according to the actual situation. The embodiments in this application are only illustrative and do not impose specific limitations.

[0170] Specifically, embodiments of this application can be based on the energy efficiency of the PEM system of the test vehicle under standard conditions (the same conditions as during formal testing, the route memory phase before formal testing, and the verification testing phase). Energy saving rate of multi-type disturbance test under various disturbance conditions (i.e., energy saving rate of multi-type disturbance test corresponding to multi-type disturbance test data; the calculation method of energy saving rate is the same as that of the aforementioned embodiment, and will not be repeated here). (j=1, 2, 3 correspond to traffic, weather, and load disturbance, respectively), calculate the generalization index. :

[0171] Only when all types of generalization indicators meet the corresponding preset indicator ranges can the test vehicle be subjected to formal testing in this application embodiment.

[0172] Here, the preset index range can be understood as a pre-set range of indices used to determine whether the vehicle's generalization index meets the requirements of the formal test. Only when all the generalization indices of the test vehicle's PEM system are within the corresponding preset index range can the subsequent formal test be carried out.

[0173] The specific preset index range can be set and adjusted by those skilled in the art according to the actual situation. The embodiments in this application are only illustrative and do not impose specific limitations.

[0174] For example, only when the generalization index corresponding to each type of disturbance test is ≥0.7 can the vehicle's PEM system be considered robust; if any... If the energy-saving effect decreases by more than 30% after the disturbance, the PEM system is determined to be overfitted and has no practical value. The test is terminated and an "overfitting" report is issued.

[0175] The embodiments of this application can effectively identify PEM systems that can only memorize specific trajectories and cannot generalize to different driving conditions (such as different traffic flows, weather, and loads) on the same route through generalization verification, thus avoiding the misjudgment of low robustness PEM systems as high efficiency PEM systems.

[0176] Step S103: Calculate the energy saving rate of the test vehicle under the first preset operating condition based on the first carbon emission and the second carbon emission.

[0177] As a possible approach, once the PEM system of the test vehicle has effective route memory and the energy-saving effect reaches a certain level, the traffic conditions, weather conditions, and driver style during the formal testing phase should be kept consistent with those during the verification testing phase.

[0178] The method for determining traffic status may include, but is not limited to: obtaining real-time traffic flow characteristic data for each road segment along the test route provided by the map service provider, and using a 4D evaluation criterion model to calculate the traffic flow characteristic data for the verification test phase and the formal test phase separately. When the 4D result of the formal test phase is within the threshold range generated by the traffic flow data of the verification test phase, the traffic status is determined to be consistent.

[0179] For example, the traffic condition assessment criterion model in this application embodiment may, but is not limited to, start from the traffic data in the verification test phase and summarize and integrate the following four-dimensional feature data: (1) average driving speed of road segment; (2) road traffic density; (3) hourly flow rate of single lane; (4) average headway. Each road segment is extracted separately and compared with its respective base value to form a four-dimensional feature dataset of the entire test route.

[0180] The base values ​​for the four parameters include, but are not limited to: average road segment speed base value of 100 km / h (corresponding to average road segment speed); road traffic density base value of 150 vehicles / km (corresponding to road traffic density); single lane hourly traffic base value of 2000 PCU (corresponding to single lane hourly flow rate); and average headway base value of 5s (corresponding to average headway).

[0181] Subsequently, in this embodiment of the application, the average value M and standard deviation σ of the data set for each dimension of the 4-dimensional parameters can be calculated, and (M-kσ, M+kσ) is set as the preset threshold range for that dimension. Here, k is a dimensionless coefficient used to adjust the traffic state determination threshold, which can be flexibly set to values ​​such as 1, 2, and 3. When the value of k is large, such as 2 or 3, the threshold range is wide, which means that the judgment conditions for traffic state are lenient. When the value of k is small, such as 1, it means that the judgment conditions are strict, that is, the traffic flow during the formal test is required to be highly consistent with the traffic flow during the route memory and verification test phases. Here, it can be set to 1, but is not limited to.

[0182] After obtaining the 4D feature dataset of each road segment in the formal testing phase using the same method, the embodiments of this application can calculate the average value of the four dimensions and examine whether the four parameters are respectively within the preset threshold range of each dimension. If they are, the traffic state is considered to be consistent.

[0183] By using methods such as 4D traffic condition determination, weather condition determination, and driver style recognition, the comparability and reproducibility of test results at each stage can be ensured, eliminating systematic biases introduced into energy efficiency measurement due to changes in external factors such as time, traffic, and weather.

[0184] Here, the preset threshold range can be understood as a pre-defined numerical range in which each dimension of the feature should be located. In this embodiment, it can be set to (M-kσ, M+kσ). The specific preset threshold range can be set and adjusted by those skilled in the art according to the actual situation. This embodiment is only for illustrative purposes and does not impose any specific limitations.

[0185] Furthermore, in the embodiments of this application, the method for determining weather conditions may be, but is not limited to, subjective judgment by testers based on general meteorological classification standards.

[0186] Furthermore, the driving style recognition in this application embodiment may, but is not limited to, the following methods: classifying driving styles through cluster analysis algorithms to obtain no less than three typical driving styles (aggressive, normal, and mild), constructing a driving style recognition model based on the clustering results, and automatically determining the driving style category of new drivers.

[0187] Once the generalization verification is passed, the embodiments of this application can repeat the formal test for M rounds (M≥3) under the test conditions corresponding to the route memory and verification test phases (navigation is turned off throughout, and PEM is activated by relying on the completed habitual route memory (i.e., the route memory of the target test route by the PEM system of the test vehicle)). The real-time anomaly monitoring and hierarchical correction method (steps one to five and anomaly handling of the real-time cumulative outlier rate P) in the aforementioned embodiments are also adopted to ensure the consistency of the quality of the formal test data.

[0188] For example, this application can adjust the battery SOC of both vehicles V_test and V_ref,i to 10% above the battery balance point; after driving at 70km / h for 25 minutes to fully warm up the engine, fuel is added to the same fuel nozzle until the nozzle clicks off, and the vehicle is driven out of the gas station immediately after refueling; the load, tires, air conditioning settings of the two vehicles are the same as those in the route memory and verification test phase, and OTA updates are prohibited for both vehicles.

[0189] The reference vehicle (V_ref,i) starts first, followed by the test vehicle (V_test). Both vehicles depart simultaneously, maintaining a following distance of approximately 50-100 meters. The test vehicle has its in-vehicle network enabled but navigation is disabled (navigation interface closed), and its predictive energy management system is activated. The reference vehicle also has its in-vehicle network enabled but navigation is disabled (navigation interface closed), and its predictive energy management system is directly disabled by the manufacturer using a diagnostic tool. The test route follows a pre-selected target test route, and multiple path branches are not allowed.

[0190] After each test round, both the test vehicle and the reference vehicle returned to the same gas station, filled up to the brim using the same fuel nozzle, and recorded the refueling data. Subsequent calculations primarily used onboard fuel consumption data from the vehicle's fuel consumption meter, with fuel nozzle data serving as supplementary verification.

[0191] Then, the embodiments of this application can calculate the fuel consumption per 100 kilometers and the electricity consumption per 100 kilometers of the test vehicle and the reference vehicle, and then convert them into carbon emissions per 100 kilometers, as described above: Fuel carbon emissions = fuel consumption per 100 kilometers × fuel carbon emission factor; Carbon emissions from electricity = electricity consumption per 100 kilometers × carbon emission factor for electricity; Total carbon emissions = carbon emissions from fuel oil + carbon emissions from electricity.

[0192] Carbon emission data for the M wheels of the test vehicle , ,..., Calculate the upper limit of one-sided confidence at a 90% confidence level as the final carbon emission value of the test vehicle: (1) Average carbon emissions :

[0193] in, This represents the carbon emission data for the kth round.

[0194] (2) Standard deviation of carbon emissions :

[0195] (3) Error in carbon emission standards :

[0196] From the t-distribution table, we can obtain... The upper confidence limit for:

[0197] in, The critical value of the t-distribution with M-1 degrees of freedom at a one-sided 90% confidence level can be obtained from the t-distribution critical value table.

[0198] For the reference vehicle, the 90% confidence level one-sided lower limit of the carbon emissions from wheel M is also calculated as the final carbon emission value of the reference vehicle:

[0199] in, This represents the lower confidence limit of a single-sided test at a 90% confidence level for the reference vehicle in the M-wheel test. This indicates the average carbon emissions of the reference vehicle during the M round of testing. This indicates an error in referencing vehicle carbon emission standards.

[0200] If an unforeseen event occurs during the test, the test should be interrupted and restarted at the same time the following day.

[0201] Final energy-saving effect evaluation: Energy saving rate = (referring to total carbon emissions of the vehicle) Total carbon emissions of the test vehicle / Total carbon emissions of the reference vehicle × 100%.

[0202] This application's embodiments can ensure consistency of test conditions between the formal testing phase and the route memory and verification testing phases based on a test condition determination method that combines quantification and experience. Then, it uses confidence interval boundary values ​​(upper limit for the test vehicle, lower limit for the reference vehicle) to calculate the energy-saving rate, significantly reducing the risk of exaggerating energy-saving effects due to accidental factors and effectively improving the robustness of the evaluation results. This unifies the energy consumption of vehicles with different power types into carbon emissions per 100 kilometers, achieving a unified carbon emission evaluation of PEM energy-saving effects.

[0203] The following is a detailed explanation of the energy-saving effect testing method of the predictive energy management system based on route memory in this application, using a specific embodiment.

[0204] Example 1: Hybrid Electric Vehicle (PHEV) Testing This embodiment uses a plug-in hybrid electric vehicle as the test object. This vehicle is equipped with a predictive energy management system based on historical route memory (not relying on navigation activation). Its strategy compares the vehicle's current GPS location with the database of memorized routes. When it is determined that the vehicle is traveling on a frequently used route, the historically optimal energy management strategy is invoked.

[0205] S1: Information Entropy Calibration and Selection of Test Routes A candidate route of approximately 100km was selected from a unified test route database, encompassing urban areas, suburbs, highways, and sloping sections. This route was divided into 100 1km segments, and the probability distributions of three characteristics—gradient, speed limit, and historical traffic flow—were statistically analyzed. (1) Slope: 8 downhill sections (8%), 18 gentle downhill sections (18%), 58 flat / gentle uphill sections (58%), 16 uphill sections (16%), entropy H_slope=1.62.

[0206] (2) Speed ​​limit: 22 low speeds (22%), 45 medium speeds (45%), 33 high speeds (33%), H_speed=1.55; where H_speed represents the information entropy of the "speed limit" feature on the selected target test route.

[0207] (3) Traffic flow: 30 stable (30%), 55 fluctuating (55%), 15 congested (15%), H_traffic=1.42; where H_traffic represents the information entropy of the "historical traffic flow" feature on the selected target test route.

[0208] The total information entropy H_total = (1.62 + 1.55 + 1.42) / 3 = 1.53.

[0209] Normalizing the total information entropy H_total yields the normalized information entropy H_norm: H_norm=1.53 / log24=1.53 / 2=0.765, If the value is within the range of 0.6-0.8, it meets the requirements, and this route is selected as the standard test route.

[0210] S2: Establishing a Familiar Route Memory The test vehicle activated its navigation interface and drove along the selected route via S1. The minimum number of times the GPS signal was recorded was set to m_min = 3, and the maximum number of times it was recorded to m_max = 8. The onboard GPS receiver continuously recorded the trajectory at a sampling frequency of 1Hz.

[0211] The road assessment process is illustrated using section 42 as an example. This section is located under an urban elevated highway and is 1 km long.

[0212] First driving experience (memorized): The vehicle traveled through this section of road at a speed of approximately 50 km / h, taking about 72 seconds and collecting 72 GPS sampling points. The lateral matching deviation d_i of each sampling point is distributed as follows (unit: meters): 0.62, 0.58, 0.71, 0.65, 0.59, 0.73, 0.68, 0.61, 0.66, 0.70, ... (72 values ​​in total) Calculate the root mean square of the lateral matching deviation for each memory drive: First driving experience (memorized): rice; Second memory drive: rice 3rd memory drive: rice After completing 3 memory driving sessions, calculate the coefficient of variation for road segment 42:

[0213] At this point, convergence has not been achieved. This section of the road is located under an overpass, and the GPS signal is blocked by the bridge structure, resulting in insufficient matching stability.

[0214] Compared to section 78 (a typical urban arterial road with no obstructions): The three RMS values ​​are 0.52m, 0.51m, and 0.54m, respectively. Therefore, the coefficient of variation for road segment 78 is: It reaches a convergent state.

[0215] in, , , , These are the root mean square values ​​of the lateral matching deviations from multiple drives on the corresponding road segment. RMS The corresponding standard deviation and arithmetic mean.

[0216] After three memory driving sessions, a full route statistical analysis was conducted: 70 road segments had converged, resulting in a memory saturation rate of [missing information]. The fourth memory driving test is required.

[0217] After four memory driving sessions, a full route statistical analysis was conducted: 85 road segments had converged, resulting in a memory sufficiency rate. The fifth memory driving test is required.

[0218] After 5 trips with memory access, a full route analysis was performed: 92 sections had converged, resulting in a memory saturation rate of [missing information]. Once the termination condition is met, the habitual route memory bank is established.

[0219] For 92% of the test route, the matching accuracy between the GPS trajectory and the map road network stabilized and converged after multiple drives (coefficient of variation ≤ 5%). The remaining 8% of the road sections that did not meet the standard accounted for no more than 10%, and their impact on the overall energy-saving effect of the route was controlled within an acceptable range. In the subsequent S4 verification test without navigation, if the vehicle deviates from the PEM control strategy due to insufficient GPS matching accuracy on such road sections, the real-time outlier detection mechanism in S4-1 to S4-2 will capture the abnormal energy-saving rate of the road section and handle it through graded correction or retesting.

[0220] S3: Refer to the vehicle random rotation test Prepare one test vehicle V_test (navigation off, PEM on) and three reference vehicles of the same model V_ref1, V_ref2, and V_ref3 (navigation off, PEM permanently off). Perform memory initialization on all vehicles.

[0221] The reference vehicles are randomly paired: Round 1: V_ref1 is drawn; Round 2: V_ref3 is drawn; Round 3: V_ref2 is drawn; Round 4: V_ref1 is drawn; Round 5: V_ref3 is drawn; ...

[0222] The test vehicle V_test participates continuously in each round.

[0223] S4: Activation verification and real-time anomaly monitoring of habitual route memory in the absence of navigation. At this point, the habitual route memory of the test vehicle V_test has been established in S2 (full route memory coverage rate 92%). After entering the S4 stage, the vehicle navigation interface remains closed, and V_test relies solely on GPS positioning to match the established habitual route memory library to activate the PEM function.

[0224] Before the verification test, the memory was cleared for all reference vehicles; for the test vehicle V_test, only its short-term driving habit learning value was cleared, while the habit route memory library built in S2 remained unchanged.

[0225] Then, the SOC of both vehicles' batteries was adjusted to 10% above the balance point (25%), i.e., 35%. The vehicles were driven at 70 km / h for 25 minutes to fully warm up the engines. Once the coolant temperature reached 90°C, 92-octane gasoline was added using the same nozzle until the pump clicked off. The test was started immediately after refueling, ensuring the reference vehicle was in front of the test vehicle throughout the test. Both vehicles carried one driver, with weight differences within permissible limits, and were balanced using counterweights. Tires were the same brand and model, with a cold tire pressure of 2.5 bar. Air conditioning was turned off.

[0226] Multiple rounds of testing were conducted on the selected test route. The test vehicle was driven by a driver identified as a standard driver. During each round, the vehicle was ensured to complete the predetermined route, with a reference vehicle ahead of the test vehicle. Fuel consumption data for each trip was recorded using an onboard fuel consumption meter (fuel nozzle data was used as auxiliary verification), SOC changes were recorded, and the energy consumption of V_test and the corresponding reference vehicle per round and per road segment was calculated, along with the energy saving rate. The traffic condition determination module verified that the traffic condition level at each stage met the requirements. Weather conditions were confirmed by the test personnel to be free of extreme weather.

[0227] S4-1: Real-time data collection and consistency determination of road segment energy saving rate In this embodiment, the test route was divided into 100 1km segments. Based on the slope, speed limit, and historical traffic flow characteristic parameters extracted in S1, the K-means clustering algorithm was used to divide the 100 segments into 6 sensitivity groups. Taking one of the sensitivity groups as an example, the consistency determination process within the group is explained: The first three road sections: The energy-saving rates for the first three road sections in this group are 7.8%, 8.2%, and 7.5%, respectively. This is tentatively accepted and recorded.

[0228] The initial baseline mean was calculated to be 7.83%, and the initial baseline standard deviation was 0.351%.

[0229] The fourth road segment (gradually expanding the determination): The energy-saving rate of this road section is 8.5%. Judgment:

[0230] No outlier was triggered; the road segment is normal and will be included in the baseline update.

[0231] 5th road segment (gradually expanding the determination): The energy saving rate for this road section is 2.1%. After being included in the fourth road section for benchmark updates, the benchmark mean is 8.0%, and the initial benchmark standard deviation is 0.38%. Therefore:

[0232] If an outlier is triggered, the road segment is marked as an outlier and will not be included in the baseline update.

[0233] When the cumulative number of normal road segments in the group is ≥10, the data of the first 3 road segments are re-verified: When the cumulative number of normal road segments in the group reached 10, the median of 7.95 and median absolute deviation of 0.40% were used to re-examine the first 3 road segments. No outliers were detected, and the original judgment was maintained.

[0234] S4-2: Outlier Calculation and Hierarchical Anomaly Handling Assume that by the 50th road segment in this round of testing, 6 outlier segments have been identified. Current cumulative outlier rate:

[0235] The test continues.

[0236] Assume that by the 80th road segment, 12 outlier segments have been identified. Current cumulative outlier rate:

[0237] The test should be terminated immediately, all data should be discarded, and the test should be restarted the following day.

[0238] Assuming this round of testing successfully completes 100 road segments, a total of 8 outlier road segments are ultimately identified. Final outlier rate:

[0239] If the energy saving rate is in the 5%-10% range, a weighted average method is used for correction: Overall energy saving rate before correction. =4.85%, corrected overall energy saving rate =4.78%, change = 0.07% ≤ 0.2%, correction is effective.

[0240] S4-3: Activation Verification Termination Decision Multiple rounds of testing were conducted, and the energy saving rate was recorded. Table 1 is an energy saving rate recording table of one embodiment of this application, which can be represented as follows: Table 1

[0241] calculate: The range for rounds 7-9 is 4.9 - 4.8 = 0.1% ≤ 0.2%. The average value for rounds 7-9 is (4.8 + 4.8 + 4.9) / 3 = 4.83%. The average value for rounds 4-6 is (4.6 + 4.7 + 4.7) / 3 = 4.67%. The mean difference is 0.16% ≤ 0.2%. If both conditions are met simultaneously, it is determined that the energy-saving effect has entered a stable output state, the verification test is terminated, and N=9 is recorded.

[0242] S5: Generalization Validation Energy saving rate under standard conditions: =Average energy saving rate for rounds 7-9=4.83% Traffic flow disturbance: Guide vehicles randomly accelerate and decelerate, and two tests are conducted to determine the average energy saving rate. =3.8%, =0.787.

[0243] Weather disturbances: Two tests were conducted on wet and slippery surfaces, with an average energy saving rate. =3.6%, =0.745.

[0244] Load disturbance: Increase the counterweight by 10%, conduct two tests, and calculate the average energy saving rate. =4.1%, =0.849.

[0245] all ≥0.7, no overfitting, continue with formal testing.

[0246] S6: Formal Performance Evaluation and Conservative Estimation The S4 phase has completed the stable output determination (N=9 rounds), but the data from these 9 rounds has already been used for termination determination and cannot be reused. Therefore, under the premise that the system state is stable, a formal test should be conducted again.

[0247] The battery SOC of both the test vehicle and the reference vehicle paired with this round was adjusted to 35%; after driving at 70km / h for 25 minutes to fully warm up the engine, 92-octane gasoline was added using the same refueling nozzle until the nozzle clicked off, and the vehicle was driven out of the gas station immediately after refueling; the load, tires, air conditioning settings and route memory of the two vehicles were the same as those in the verification test phase; OTA updates were prohibited.

[0248] The reference vehicle leads, followed by the test vehicle, both departing simultaneously while maintaining a following distance of approximately 50-100 meters. After the test, return to the gas station, fill the tank to the brim with the same pump, and record the amount of fuel dispensed. Repeat this test five times. The primary fuel consumption data comes from the vehicle's onboard fuel consumption meter, with the pump data used for supplementary verification.

[0249] Calculate carbon emissions per round (unit: g / km): Test vehicles: 118.2, 119.5, 117.8, 120.1, 118.9; Mean: 118.9; Standard deviation: 0.935; Standard error: 0.418; =1.533; upper confidence limit =118.9 + 1.533 × 0.418 = 119.5; Reference vehicles: 125.1, 124.9, 125.0, 124.8, 125.0; Mean: 125.0; Standard deviation: 0.114; Standard error: 0.051; =1.533; lower confidence limit =125.0 - 1.533 × 0.051 = 124.9; Energy saving rate = (124.9-119.5) / 124.9×100%=4.32%.

[0250] Comparative Example 1: Overfitting System Detection Assuming another PEM system, the standard energy saving rate during the S4 verification phase =5.2%. Average energy saving rate during disturbance testing under traffic flow disturbances. =2.5%, =0.48 < 0.7, indicating overfitting, the test is terminated and an "overfitting" report is output, avoiding an erroneous overestimation of its energy-saving effect. This comparison shows that this application can effectively identify and terminate invalid tests on systems without practical value in advance, saving a significant amount of testing resources.

[0251] Comparative Example 2: Outlier rate exceeding the limit leads to round termination Suppose that during a test run on the 35th road segment, 11 outlier segments have been identified, resulting in a cumulative outlier rate P = 11 / 100 × 100% = 11% > 10%. The system immediately terminates this test run, discarding all data and avoiding invalid testing on the remaining 65 road segments. The test run is then re-executed the following day. This comparative example demonstrates that the real-time monitoring mechanism of this invention can proactively detect serious anomalies and interrupt testing in advance, effectively conserving testing resources.

[0252] Comparative Example 3: Insufficient route memorization led to route changes Assuming a candidate route includes numerous elevated underpasses and tunnel entrances / exits (resulting in severe GPS signal obstruction), after m_max = 8 memory drives, what is the full memory utilization rate of the entire route? Only 72% was achieved, failing to reach the 90% threshold. The system determined that this test route was unsuitable as a standard test route for the PEM system due to severe GPS signal obstruction, and automatically returned to S1 to select an alternative route. This comparative example demonstrates that the habitual route memory bank establishment method of the present invention can effectively identify difficulties in memory bank construction caused by geographical environmental factors, avoiding invalid subsequent tests on routes with inherently poor conditions.

[0253] Comparative Example 4: Identification of Non-convergent Systems Suppose that the energy-saving rate of a certain PEM system exhibits continuous oscillations over 10 rounds of testing: 3.2%, 4.1%, 3.5%, 4.3%, 3.8%, 4.5%, 3.7%, 4.2%, 3.9%, 4.4%. Calculations show that the range for three consecutive rounds is consistently greater than 0.2%, and the convergence criterion is still not met after 10 rounds. The system determines that the energy-saving effect of this PEM system fluctuates too much in actual use and lacks stable and reliable practical value. Therefore, the testing is terminated to avoid continuously consuming testing resources on an ineffective system.

[0254] Example 2: Testing of Traditional Fuel-Powered Vehicles Similar to Example 1, but without involving SOC control and energy consumption calculation, only fuel consumption per 100 kilometers is calculated and converted into carbon emissions. The remaining steps are the same.

[0255] Example 3: Pure Electric Vehicle Testing Similar to Example 1, but the SOC starts from a full charge and does not involve fuel consumption; only the electricity consumption per 100 kilometers is calculated and converted into carbon emissions. The remaining steps are the same.

[0256] The energy-saving effect testing method for the predictive energy management system based on route memory proposed in this application can select a target test route from multiple candidate test routes based on information entropy. The method calculates the energy-saving rate of the test vehicle when navigation is off and the predictive energy management system is on, and when the reference vehicle is off and the predictive energy management system is off. This ensures stable energy consumption of the reference vehicle under suitable test route conditions, through the rules of rotating and discontinuous use of multiple reference vehicles, avoiding artificially inflated or understated energy-saving rates and improving the credibility of the test results. Real-time data quality monitoring and graded correction strategies based on road segment clustering and outlier rate classification achieve real-time quality control of the testing process, effectively avoiding the risk of entire-round data scrapping or erroneous data participating in termination decisions due to local anomalies. Finally, a test condition determination method combining quantification and experience ensures that the formal test is conducted under the same conditions as the route memory and verification test phases, completing a unified carbon emission evaluation of the PEM energy-saving effect. This solves the problems in related technologies, such as the fact that energy-saving test methods for PEM strategies are all set up for PEM strategies that require navigation activation, without taking into account the actual navigation scenarios of PEM in real applications, and the lack of energy-saving effect test methods for PEM strategies that do not rely on navigation activation, resulting in test results that cannot truly reflect the real energy-saving effect of PEM strategies.

[0257] Next, referring to the accompanying drawings, a test device for the energy-saving effect of a predictive energy management system based on route memory, according to an embodiment of this application, is described.

[0258] Figure 2This is a schematic diagram of the energy-saving effect testing device of the predictive energy management system based on route memory, according to an embodiment of this application.

[0259] like Figure 2 As shown, the energy-saving effect testing device 10 of the predictive energy management system based on route memory includes: a selection module 100, a first acquisition module 200, and a first calculation module 300.

[0260] The selection module 100 is used to acquire the road segment features of multiple road segments among multiple candidate test routes, calculate the information entropy corresponding to the multiple candidate test routes based on the road segment features, and select the target test route from the multiple candidate test routes based on the information entropy; the first acquisition module 200 is used to acquire the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively; the first calculation module 300 is used to calculate the energy-saving rate of the test vehicle under the first preset operating condition based on the first carbon emission and the second carbon emission.

[0261] Optionally, in one embodiment of this application, the selection module 100 includes the following steps: a first acquisition unit, used to acquire the slope characteristics, speed limit characteristics, and historical traffic flow characteristics of multiple road segments in multiple candidate test routes, and to divide the slope characteristics, speed limit characteristics, and historical traffic flow characteristics into multiple levels; a statistics unit, used to calculate the proportion of road segments corresponding to each level of slope characteristics, speed limit characteristics, and historical traffic flow characteristics, and to obtain the probability distribution corresponding to slope characteristics, speed limit characteristics, and historical traffic flow characteristics; and a first calculation unit, used to calculate the information entropy corresponding to multiple candidate test routes based on the probability distribution.

[0262] Optionally, in one embodiment of this application, it further includes: a determining module, configured to, before obtaining the first carbon emission of the test vehicle under a first preset operating condition and the second carbon emission of the reference vehicle under a second preset operating condition based on the target test route, determine multiple test sampling points of multiple road segments corresponding to each round of driving data in the multi-round driving data of the test vehicle, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet preset requirements and preset energy-saving conditions respectively; a second calculation module, configured to calculate the matching deviation between the multiple test sampling points and the target centerline, so as to calculate the trajectory matching quality between the multiple road segments and the target test route based on the matching deviation; and a first determining module, configured to determine the proportion of road segments that have reached the convergence state in the target test route based on the trajectory matching quality, so as to determine the route memory of the predictive energy management system of the test vehicle based on the proportion of road segments.

[0263] Optionally, in one embodiment of this application, it further includes: a second acquisition module, configured to, before acquiring the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, acquire, based on the target test route, multiple test results of the test vehicle under the navigation condition being off and the predictive energy management system under the predictive energy management system being on, and multiple test results of the reference vehicle under the navigation condition being off and the predictive energy management system under the predictive energy management system being off; and a second determination module, configured to, based on the multiple test results corresponding to the test vehicle and the reference vehicle, acquire the number of outlier road segments in each test process of the test vehicle, so as to determine the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments.

[0264] Optionally, in one embodiment of this application, the second determining module includes: a segmentation unit, used to divide multiple road segments into multiple road segment groups based on the road segment characteristics of multiple road segments of the target test route; a second calculation unit, used to calculate an initial benchmark mean and an initial benchmark standard deviation based on the energy saving rate of a preset road segment in each road segment group by the predictive energy management system of the test vehicle; an identification unit, used to identify outliers in the preset road segment and other road segments based on the initial benchmark mean, the initial benchmark standard deviation, and the energy saving rate of other road segments in each road segment group besides the preset road segment; a first determining unit, used to determine the total number of real-time outliers of the test vehicle during the test process based on the number of outliers in the multiple road segment groups; and a second determining unit, used to calculate the real-time cumulative outlier rate of the test vehicle during multiple rounds of testing based on the total number of real-time outliers, and determine the energy saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate.

[0265] Optionally, in one embodiment of this application, the second determining unit includes: a first calculation subunit, configured to calculate the energy saving rate of the predictive energy management system of the test vehicle based on the test results, provided that the real-time cumulative outlier rate meets the preset outlier condition; a second calculation subunit, configured to calculate the range of energy saving rates of the test vehicle in the first target round of testing based on the energy saving rate, and calculate the absolute value of the difference between the average energy saving rate of the first target round of testing and the average energy saving rate of the second target round of testing; and a determining subunit, configured to determine the energy saving effect of the predictive energy management system of the test vehicle based on the range and absolute value of the energy saving rate.

[0266] Optionally, in one embodiment of this application, the first acquisition module 200 includes: a third calculation unit, used to acquire multi-type disturbance test data of the test vehicle to calculate the energy saving rate of the predictive energy management system of the test vehicle under multi-type disturbance tests; and a second acquisition unit, used to calculate multi-type generalization indicators of the test vehicle based on the energy saving rate under multi-type disturbance tests, and when the multi-type generalization indicators meet the corresponding preset indicator range, to acquire the first carbon emission and the second carbon emission based on the target test route.

[0267] Optionally, in one embodiment of this application, it further includes: a third determining module, configured to determine that the test vehicle is in the first preset operating condition before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, when the navigation operating condition of the test vehicle is off and the predicted energy management system operating condition is on; and a fourth determining module, configured to determine that the reference vehicle is in the second preset operating condition when the navigation operating condition of the reference vehicle is off and the predicted energy management system operating condition is off.

[0268] It should be noted that the foregoing explanation of the embodiment of the test method for the energy-saving effect of the predictive energy management system based on route memory also applies to the test device for the energy-saving effect of the predictive energy management system based on route memory in this embodiment, and will not be repeated here.

[0269] The energy-saving effect testing device for the predictive energy management system based on route memory proposed in this application can select a target test route from multiple candidate test routes based on information entropy. It calculates the energy-saving rate of the test vehicle when navigation is off and the predictive energy management system is on, and calculates the energy-saving rate of the reference vehicle when navigation is off and the predictive energy management system is off. This ensures stable energy consumption of the reference vehicle under suitable test route conditions by using a rotation and discontinuous use rule for multiple reference vehicles, preventing artificially high or low energy-saving rates and improving the credibility of the test results. Through strategies such as real-time data quality monitoring and graded correction based on road segment clustering and outlier rate classification, real-time quality control of the testing process is achieved, effectively avoiding the risk of entire-round data scrapping or erroneous data participating in termination judgment due to local anomalies. Finally, a test condition judgment method combining quantification and experience ensures that the formal test is conducted under the same conditions as the route memory and verification test stages, completing a unified evaluation of carbon emissions for the PEM energy-saving effect. This solves the problems in related technologies, such as the fact that energy-saving test methods for PEM strategies are all set up for PEM strategies that require navigation activation, without taking into account the actual navigation scenarios of PEM in real applications, and the lack of energy-saving effect test methods for PEM strategies that do not rely on navigation activation, resulting in test results that cannot truly reflect the real energy-saving effect of PEM strategies.

[0270] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 301, the processor 302, and the computer program stored on the memory 301 and capable of running on the processor 302.

[0271] When the processor 302 executes the program, it implements the energy-saving effect testing method of the predictive energy management system based on route memory provided in the above embodiments.

[0272] Furthermore, electronic devices also include: Communication interface 303 is used for communication between memory 301 and processor 302.

[0273] The memory 301 is used to store computer programs that can run on the processor 302.

[0274] The memory 301 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage.

[0275] If the memory 301, processor 302, and communication interface 303 are implemented independently, then the communication interface 303, memory 301, and processor 302 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0276] Optionally, in a specific implementation, if the memory 301, processor 302, and communication interface 303 are integrated on a single chip, then the memory 301, processor 302, and communication interface 303 can communicate with each other through an internal interface.

[0277] Processor 302 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0278] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for testing the energy-saving effect of a predictive energy management system based on route memory.

[0279] This application also provides a computer program product, including a computer program that can run computer instructions. When the computer instructions are executed by a processor, they implement the energy-saving effect testing method of the predictive energy management system based on route memory provided in this application.

[0280] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0281] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0282] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0283] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0284] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0285] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0286] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0287] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for testing the energy-saving effect of a predictive energy management system based on route memory, characterized in that, Includes the following steps: Obtain the segment features of multiple road segments from multiple candidate test routes, calculate the information entropy corresponding to the multiple candidate test routes based on the segment features, and select the target test route from the multiple candidate test routes based on the information entropy; Under the condition that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, based on the target test route, the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition are obtained. Based on the first carbon emission and the second carbon emission, the energy saving rate of the test vehicle under the first preset operating condition is calculated.

2. The method according to claim 1, characterized in that, The step of obtaining the road segment features of multiple road segments from multiple candidate test routes, and calculating the information entropy corresponding to the multiple candidate test routes based on the road segment features, includes the following steps: The slope characteristics, speed limit characteristics, and historical traffic flow characteristics of multiple road segments in the candidate test routes are obtained, and the slope characteristics, speed limit characteristics, and historical traffic flow characteristics are all divided into multiple levels; The probability distributions corresponding to the slope characteristics, speed limit characteristics, and historical traffic flow characteristics are obtained by statistically analyzing the proportion of road segments at each level. Based on the probability distribution, the information entropy corresponding to multiple candidate test routes is calculated.

3. The method according to claim 1, characterized in that, Before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition, based on the target test route, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the method further includes: Based on the target test route, determine multiple test sampling points for multiple road segments corresponding to each round of driving data in the multi-round driving data of the test vehicle; Calculate the matching deviation between multiple test sampling points and the target centerline, and calculate the trajectory matching quality between multiple road segments and the target test route based on the matching deviation; Based on the trajectory matching quality, the proportion of road segments that have reached convergence in the target test route is determined, and the route memory of the predictive energy management system of the test vehicle is determined according to the proportion of road segments.

4. The method according to claim 1, characterized in that, Before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition, based on the target test route, provided that the route memory and energy-saving effect of the predictive energy management system of the test vehicle meet the preset requirements and preset energy-saving conditions respectively, the method further includes: Based on the target test route, obtain the test results of the test vehicle under the condition that the navigation is off and the predicted energy management system is on, as well as the test results of the reference vehicle under the condition that the navigation is off and the predicted energy management system is off. Based on the results of multiple rounds of testing of the test vehicle and the reference vehicle, the number of outlier road segments of the test vehicle in each round of testing is obtained, and the energy-saving effect of the predictive energy management system of the test vehicle is determined based on the number of outlier road segments.

5. The method according to claim 4, characterized in that, The determination of the energy-saving effect of the predictive energy management system of the test vehicle based on the number of outlier road segments includes: Based on the segment characteristics of multiple segments of the target test route, the multiple segments are divided into multiple segment groups; Based on the predictive energy management system of the test vehicle, the energy saving rate of the preset road segments in each road segment group is calculated, and the initial benchmark mean and initial benchmark standard deviation are calculated. Based on the initial baseline mean, the initial baseline standard deviation, and the energy-saving rate of other road segments in each road segment group (excluding the preset road segment), outlier road segments are identified in the preset road segment and the other road segments. Based on the number of outlier road segments in the multiple road segment groups, the total number of outlier road segments in real time for the test vehicle during the test process is determined. The real-time cumulative outlier rate of the test vehicle during multiple rounds of testing is calculated based on the total number of outlier road segments in real time, and the energy-saving effect of the predictive energy management system of the test vehicle is determined based on the real-time cumulative outlier rate.

6. The method according to claim 5, characterized in that, The determination of the energy-saving effect of the predictive energy management system of the test vehicle based on the real-time cumulative outlier rate includes: If the real-time cumulative outlier rate meets the preset outlier condition, the energy saving rate of the predictive energy management system of the test vehicle is calculated based on the test results. Based on the energy saving rate, calculate the range of energy saving rates of the test vehicle in the first target round test, and calculate the absolute value of the difference between the average energy saving rate of the first target round test and the average energy saving rate of the second target round test. Based on the energy saving rate range and the absolute value, the energy saving effect of the predictive energy management system of the test vehicle is determined.

7. The method according to claim 1, characterized in that, The step of obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route further includes: Acquire multi-type disturbance test data of the test vehicle to calculate the energy saving rate of the predictive energy management system of the test vehicle under multi-type disturbance tests; Based on the energy saving rate under various disturbance tests, various generalization indices of the test vehicle are calculated. When the various generalization indices meet the corresponding preset index ranges, the first carbon emission and the second carbon emission are obtained based on the target test route.

8. The method according to claim 1, characterized in that, Before obtaining the first carbon emission of the test vehicle under the first preset operating condition and the second carbon emission of the reference vehicle under the second preset operating condition based on the target test route, the method further includes: When the navigation mode of the test vehicle is off and the predictive energy management system mode is on, the test vehicle is determined to be in the first preset mode. When the navigation mode of the reference vehicle is off and the predictive energy management system mode is off, the reference vehicle is determined to be in the second preset mode.

9. An electronic device, characterized in that, include: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the energy-saving effect testing method for a predictive energy management system based on route memory as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the method for testing the energy-saving effect of the predictive energy management system based on route memory as described in any one of claims 1-8.

Images