A method and apparatus for predicting a range

Abstract

The application discloses a driving range prediction method and device, and relates to the technical field of data processing, and aims to improve the accuracy of driving range prediction. The method comprises the following steps: obtaining a first prediction model; obtaining a first driving range prediction value by using the first prediction model; when the first driving range prediction value does not meet a first preset requirement, adjusting the first prediction model to obtain a second prediction model; and predicting the driving range by using the second prediction model. The embodiment of the application can improve the accuracy of driving range prediction.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and apparatus for predicting driving range. Background Technology

[0002] In recent years, the electric vehicle market has grown rapidly. Among them, pure electric vehicles are the most popular models. The driving range of pure electric vehicles is one of the most important features for electric vehicle sales and directly affects the user experience in actual use.

[0003] Currently, the NEDC (New European Driving Cycle) can be used to predict driving range. However, the test conditions specified in this standard deviate significantly from actual urban driving conditions. In reality, driving range is further affected by the vehicle's own operating conditions (such as vehicle weight, speed, whether windows are open, and whether air conditioning is on) and the surrounding environment (such as temperature, humidity, and altitude).

[0004] Existing technology CN111806239A provides a method for measuring the driving range of electric vehicles. According to this method, the driving range can be obtained by looking up a table based on the vehicle's test power consumption and a pre-established correlation between remaining battery level and driving range. However, vehicles that do not support OTA (Over-the-Air Technology) cannot update this correlation. Existing technology also proposes a method to improve the accuracy of driving range prediction based on machine learning.

[0005] However, when using existing methods to predict driving range, adjustments cannot be made if the prediction is inaccurate, thus affecting the accuracy of the driving range prediction. Summary of the Invention

[0006] This application provides a method and apparatus for predicting driving range, so as to improve the accuracy of predicting driving range.

[0007] In a first aspect, embodiments of this application provide a method for predicting driving range, applied to an in-vehicle terminal, characterized in that it includes:

[0008] Obtain the first prediction model;

[0009] The first predicted driving range is obtained using the first prediction model;

[0010] When the first predicted driving range does not meet the first preset requirement, the first prediction model is adjusted to obtain the second prediction model.

[0011] The second prediction model is used to predict the driving range.

[0012] Before adjusting the first prediction model to obtain the second prediction model, the method further includes:

[0013] Based on SOA (Service Oriented Architecture), establish service subscription and publication relationships with the target domain;

[0014] The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain.

[0015] The establishment of service subscription and publication relationships with the target domain includes:

[0016] After registration is completed in the service management middleware, an assisted driving service subscription request is sent to the service management middleware. The assisted driving service subscription request is used to subscribe to information on the available computing resources of the target domain from the target domain.

[0017] Receive a service publication message sent by the target domain, the service publication message including information on the available computing resources of the target domain.

[0018] The step of adjusting the first prediction model to obtain the second prediction model includes:

[0019] Obtain available computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain;

[0020] If the available computing resources meet the second preset requirements, then the available computing resources are used to train the model to obtain the second prediction model;

[0021] If the available computing resources do not meet the second preset requirements, the second prediction model is obtained from the cloud.

[0022] Wherein, if the available computing resources meet the second preset requirement, then using the available computing resources to train the model to obtain the second prediction model includes:

[0023] If the available computing resources in this domain meet the second preset requirement, then the available computing resources in this domain are used to train the model to obtain the second prediction model.

[0024] If the available computing resources of the target domain meet the second preset requirement, then training data is sent to the target domain and the second prediction model is obtained from the target domain, wherein the target domain trains the model based on the training data to obtain the second prediction model.

[0025] Wherein, if the second prediction model is obtained by using the available computing resources of this domain for model training, or by sending training data to the target domain and obtaining the second prediction model from the target domain, the step of using the second prediction model to predict the driving range includes:

[0026] If the second prediction model is successfully obtained within the first preset time, the remaining driving range is predicted using the second prediction model.

[0027] The method further includes:

[0028] If the second prediction model is not successfully obtained within the first preset time, the second prediction model is obtained from the cloud.

[0029] The step of obtaining the second prediction model from the cloud includes:

[0030] Send a model retrieval request to the cloud;

[0031] The system receives model parameters sent by the cloud based on the model acquisition request, and updates the model based on the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

[0032] The method further includes, after predicting the driving range using the second prediction model:

[0033] Determine whether the second predicted driving range obtained using the second prediction model meets the first preset requirement;

[0034] When the second predicted driving range meets the first preset requirement, the second prediction model is used to predict the driving range.

[0035] When the second predicted driving range does not meet the first preset requirement, if the first prediction model is obtained from the cloud, then the first prediction model is used to predict the driving range; when the second predicted driving range does not meet the first preset requirement, if the first prediction model is not obtained from the cloud, then the prediction model is obtained from the cloud.

[0036] The method further includes:

[0037] When the second predicted driving range meets the first preset requirement, the model parameters of the second prediction model are sent to the cloud.

[0038] Secondly, this application also provides a method for predicting driving range, applied in the cloud, characterized by including:

[0039] Receive model acquisition requests sent by the vehicle-mounted terminal;

[0040] According to the model acquisition request, the vehicle terminal sends model parameters, which are used by the vehicle terminal to adjust the first prediction model to obtain the second prediction model.

[0041] The model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

[0042] The method further includes:

[0043] Generate a single-vehicle prediction model and define a scenario prediction model;

[0044] The single-vehicle prediction model is sent to the vehicle terminal.

[0045] The method further includes:

[0046] The single-vehicle prediction model is updated when at least one of the following conditions is met:

[0047] The adjusted model parameters sent by the vehicle terminal are received.

[0048] Based on the driving status information sent by the vehicle terminal, it is determined that the newly added mileage of the vehicle at the location of the vehicle terminal is greater than the first value.

[0049] Based on the driving status information sent by the vehicle terminal, it is determined that the number of new charge / discharge cycles for the vehicle at which the vehicle terminal is located is greater than the second value.

[0050] The method further includes:

[0051] The prediction model for the defined scenario is updated when at least one of the following conditions is met:

[0052] A new driving scenario has been identified;

[0053] The existing driving scenario data is greater than the third value.

[0054] Thirdly, embodiments of this application also provide a driving range prediction device, applied to an in-vehicle terminal, comprising:

[0055] The first acquisition module is used to acquire the first prediction model;

[0056] The first prediction module is used to obtain a first predicted driving range value using the first prediction model.

[0057] The second acquisition module is used to adjust the first prediction model to obtain a second prediction model when the first predicted driving range does not meet the first preset requirement.

[0058] The second prediction module is used to predict the driving range using the second prediction model.

[0059] The device further includes:

[0060] The first module is used to establish service subscription and publication relationships with the target domain based on SOA.

[0061] The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain.

[0062] Fourthly, embodiments of this application also provide a driving range prediction device, applied in the cloud, including:

[0063] The first receiving module is used to receive model acquisition requests sent by the vehicle terminal;

[0064] The first sending module is used to send model parameters to the vehicle terminal according to the model acquisition request. The model parameters are used by the vehicle terminal to adjust the first prediction model to obtain the second prediction model.

[0065] The model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

[0066] The device further includes:

[0067] The generation module is used to generate prediction models for individual vehicles and prediction models that define scenarios;

[0068] The second sending module is used to send the single-vehicle prediction model to the vehicle terminal.

[0069] In this embodiment, when the predicted driving range obtained using the first prediction model does not meet the first preset requirement, the first prediction model is adjusted to obtain a second prediction model, and the second prediction model is used to predict the driving range. Since the first prediction model can be adjusted when the prediction does not meet the first preset requirement, the solution using this embodiment can obtain a prediction model that better meets the preset requirement by adjusting the prediction model, thereby improving the accuracy of the driving range prediction. Attached Figure Description

[0070] Figure 1 This is one of the flowcharts of the driving range prediction method provided in the embodiments of this application;

[0071] Figure 2 This is an SOA architecture diagram that includes the driving range prediction service;

[0072] Figure 3 This is the second flowchart of the driving range prediction method provided in the embodiments of this application;

[0073] Figure 4 This is the third flowchart of the driving range prediction method provided in the embodiments of this application;

[0074] Figure 5 This is a schematic diagram illustrating the process of training a model in the cloud in an embodiment of this application;

[0075] Figure 6 This is a schematic diagram illustrating the process of adjusting the prediction model at the vehicle end in an embodiment of this application;

[0076] Figure 7 This is a schematic diagram illustrating the process by which the power domain subscribes to services from other domains in an embodiment of this application;

[0077] Figure 8 This is one of the structural diagrams of the driving range prediction device provided in the embodiments of this application;

[0078] Figure 9 This is the second structural diagram of the driving range prediction device provided in the embodiments of this application;

[0079] Figure 10 This is a structural diagram of the first establishing module in the driving range prediction device provided in the embodiments of this application;

[0080] Figure 11 This is a structural diagram of the second acquisition module in the driving range prediction device provided in the embodiments of this application;

[0081] Figure 12 This is a structural diagram of the second processing submodule in the driving range prediction device provided in the embodiments of this application;

[0082] Figure 13 This is the third structural diagram of the driving range prediction device provided in the embodiments of this application;

[0083] Figure 14 This is the fourth structural diagram of the driving range prediction device provided in the embodiments of this application;

[0084] Figure 15 This is the fifth structural diagram of the driving range prediction device provided in the embodiments of this application;

[0085] Figure 16 This is the sixth structural diagram of the driving range prediction device provided in the embodiments of this application;

[0086] Figure 17 This is the seventh structural diagram of the driving range prediction device provided in the embodiments of this application;

[0087] Figure 18 This is the eighth structural diagram of the driving range prediction device provided in the embodiments of this application;

[0088] Figure 19 This is a structural diagram of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0089] In the embodiments of this application, the term "and / or" describes the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following associated objects have an "or" relationship.

[0090] In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar.

[0091] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0092] See Figure 1 , Figure 1 This is a flowchart of the driving range prediction method provided in the embodiments of this application, applied to an in-vehicle terminal. For example... Figure 1 As shown, the method includes the following steps:

[0093] Step 101: Obtain the first prediction model.

[0094] In this embodiment, the first prediction model refers to the model currently used to predict driving range. This first prediction model may be obtained from the cloud, or it may be a prediction model obtained after the in-vehicle terminal itself adjusts the model.

[0095] Step 102: Obtain the first predicted driving range using the first prediction model.

[0096] By running the first prediction model, the predicted first driving range can be obtained.

[0097] Step 103: When the first predicted driving range does not meet the first preset requirement, the first prediction model is adjusted to obtain the second prediction model.

[0098] The first preset requirement can be a prediction accuracy requirement, which can be set according to actual needs. For example, within a cumulative driving range of 30km (or within a battery charge drop of 5kWh), the prediction accuracy of the remaining driving range needs to be greater than 80%. If this first preset requirement is not met, the first prediction model can be adjusted to obtain a second prediction model.

[0099] Adjusting the first prediction model includes adjusting the model by the vehicle terminal itself and adjusting the model by obtaining model parameters from the cloud.

[0100] In practical applications, a large number of automotive control functions are implemented through ECUs (Electronic Control Units). Due to the increasing number of automotive components implementing control, dozens to hundreds of ECUs typically operate within a vehicle. Currently, with the growing acceptance of emerging features such as driver assistance systems and smart cockpits in the automotive market, there is a trend towards domain controller architectures and even central computing architectures. However, due to the extremely high computing power, compliance, and security requirements of central computing architectures, it is currently only considered a possible future direction for automotive evolution. Domain controller-based automotive architectures are the most likely electronic and electrical architecture to be implemented in the present and near future. Therefore, the industry has proposed an architecture based on SOA (Service-Oriented Architecture). SOA is currently widely recognized by the automotive industry as an application architecture for use in conjunction with future domain controllers. SOA can support the integration and planning of business functions by OEMs (Original Equipment Manufacturers) across different operating systems and hardware platforms (controllers).

[0101] Based on the above design, the embodiments of this application can adjust the model based on the SOA architecture. Specifically, before step 103, the method of the embodiments of this application may further include: establishing a service subscription and publication relationship with a target domain based on SOA; wherein, the target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain.

[0102] like Figure 2 The diagram shown is an SOA architecture diagram that includes a driving range prediction service. Figure 2 In the vehicle, the domain can include powertrain domain, chassis domain, body domain, driver assistance domain, infotainment domain, etc. The range prediction service is applied to the powertrain domain. Each domain has corresponding services; for example, the range estimation service is applied to the powertrain domain. The powertrain domain, as described below, can subscribe to services from other domains and establish service subscription and publication relationships with them. The in-vehicle terminal performs corresponding operations through the powertrain domain.

[0103] After a service in this domain completes registration with the service management middleware, it sends an assisted driving service subscription request to the middleware. This request is used to subscribe to information about the available computing resources of the target domain. Subsequently, the service in this domain receives a service publication message from the target domain, which includes information about the available computing resources of the target domain.

[0104] Specifically, in this scenario, the services within this domain can access available computing resources, including either the available computing resources of this domain or the available computing resources of the target domain. If the available computing resources meet a second preset requirement, the model is trained using these resources to obtain the second prediction model; otherwise, the second prediction model is retrieved from the cloud. The second preset requirement can be set as needed, for example, whether there is sufficient computing space.

[0105] If the available computing resources of this domain meet the second preset requirement, then the available computing resources of this domain are used for model training to obtain the second prediction model. If the available computing resources of the target domain meet the second preset requirement, then training data is sent to the target domain and the second prediction model is obtained from the target domain, wherein the target domain trains the model based on the training data to obtain the second prediction model. In this way, prioritizing the computing resources of this domain allows other domains to better play their respective roles and reduces the impact on other domains.

[0106] If the model is obtained from the cloud, a model acquisition request can be sent to the cloud during this process. The cloud can then send model parameters based on the acquisition request, and the model can be updated according to these parameters to obtain the second prediction model. These model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located. The model acquisition request can also carry information related to vehicle driving, such as the road scene (including vehicle geographic coordinates), vehicle speed, outside ambient temperature, air conditioning temperature setting, seat heating level, and window opening status, to facilitate the cloud in selecting matching model parameters. This further improves the accuracy of model adjustments and the accuracy of range prediction.

[0107] Step 104: Predict the driving range using the second prediction model.

[0108] Optionally, to improve model prediction efficiency and reduce the impact on driving range prediction, if the second prediction model is successfully obtained within a first preset time, the driving range is predicted using the second prediction model. If the second prediction model is not successfully obtained within the first preset time, the second prediction model is retrieved from the cloud. The first preset time can be set as needed, for example, to 5 minutes.

[0109] In this embodiment, when the predicted driving range obtained using the first prediction model does not meet the first preset requirement, the first prediction model is adjusted to obtain a second prediction model, and the second prediction model is used to predict the driving range. Since the first prediction model can be adjusted when the prediction does not meet the first preset requirement, the solution using this embodiment can obtain a prediction model that better meets the preset requirement by adjusting the prediction model, thereby improving the accuracy of the driving range prediction.

[0110] Based on the above embodiments, to further ensure the applicability of the obtained second prediction model, after step 104, it can be determined whether the second predicted driving range value obtained using the second prediction model meets the first preset requirement. When the second predicted driving range value meets the first preset requirement, the driving range is predicted using the second prediction model; when the second predicted driving range value does not meet the first preset requirement, if the first prediction model is obtained from the cloud, the driving range is predicted using the first prediction model; when the second predicted driving range value does not meet the first preset requirement, if the first prediction model is not obtained from the cloud, the prediction model is obtained from the cloud.

[0111] When the second predicted driving range meets the first preset requirement, the model parameters of the second prediction model are sent to the cloud, so that the cloud can update its own model parameters and ensure the consistency of cloud and vehicle parameters.

[0112] See Figure 3 , Figure 3 This is a flowchart of the driving range prediction method provided in the embodiments of this application, applied in the cloud. Figure 3 As shown, the method includes the following steps:

[0113] Step 301: Receive the model acquisition request sent by the vehicle terminal.

[0114] Step 302: Send model parameters to the vehicle terminal according to the model acquisition request. The model parameters are used by the vehicle terminal to adjust the first prediction model to obtain the second prediction model.

[0115] The model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

[0116] Optionally, the model acquisition request may also include vehicle driving-related information, such as the road scene (including vehicle geographic coordinates), vehicle speed, outside ambient temperature, air conditioning temperature setting, seat heating status, and window opening indicators. This information facilitates the cloud's selection of matching model parameters, thereby further improving the accuracy of model adjustments and the accuracy of range prediction. Therefore, the cloud can retrieve a matching model from its stored predictive models for defined scenarios based on the model acquisition request.

[0117] In this embodiment, the cloud generates a single-vehicle prediction model and a scenario-defined prediction model. The single-vehicle prediction model is trained using data collected from a single vehicle, while the scenario-defined prediction model is trained using data collected from multiple vehicles. The scenario-defined prediction model primarily refers to driving scenarios with distinct energy consumption characteristics, such as warm / cold city roads, highways, and mountainous terrain. After generating the models, the cloud sends the single-vehicle prediction model to the in-vehicle terminal and stores the scenario-defined prediction model itself.

[0118] Furthermore, to ensure model accuracy, the cloud platform can also update its own model. Specifically, the bicycle prediction model is updated when at least one of the following conditions is met:

[0119] The adjusted model parameters sent by the vehicle terminal are received.

[0120] Based on the driving status information sent by the vehicle terminal, it is determined that the newly added mileage of the vehicle at the location of the vehicle terminal is greater than the first value.

[0121] Based on the driving status information sent by the vehicle terminal, it is determined that the number of new charge / discharge cycles for the vehicle at which the vehicle terminal is located is greater than the second value.

[0122] The prediction model for the defined scenario is updated when at least one of the following conditions is met:

[0123] A new driving scenario has been identified;

[0124] The existing driving scenario data is greater than the third value.

[0125] The first, second, and third values ​​can be set as needed. A new scenario refers to a scenario that differs from existing scenarios.

[0126] In this embodiment, when the predicted driving range obtained using the first prediction model does not meet the first preset requirement, the first prediction model is adjusted to obtain a second prediction model, and the second prediction model is used to predict the driving range. Since the first prediction model can be adjusted when the prediction does not meet the first preset requirement, the solution using this embodiment can obtain a prediction model that better meets the preset requirement by adjusting the prediction model, thereby improving the accuracy of the driving range prediction.

[0127] See Figure 4 , Figure 4 This is a flowchart of the driving range prediction method provided in the embodiments of this application. For example... Figure 4 As shown, the method may include:

[0128] Step 401: Use machine learning methods in the cloud to train the model, resulting in a single-vehicle model and a model for defining the scenario.

[0129] Step 402: The cloud will publish the generated bicycle model to the corresponding vehicle terminal.

[0130] Step 403: While using the single-vehicle model to infer the remaining driving range, the vehicle terminal determines whether the model needs to be updated based on certain conditions, and initiates the model adjustment process when an update is needed.

[0131] Model adjustments may include local adjustments or adjustments requested from the cloud (such as requesting a defined scenario model from the cloud). After the new model is generated, the vehicle updates the model and uses the new model to predict driving range.

[0132] Step 404: The vehicle sends the adjusted new model parameters from the local model to the cloud.

[0133] Step 405: Record the model adjustment parameters in the cloud for optimization of the vehicle model and the corresponding scene model.

[0134] like Figure 5 The diagram illustrates the process of training the model in the cloud. The cloud uses historical driving data collected from the vehicle to train both a single-vehicle prediction model and a scenario-defined prediction model.

[0135] Specifically, the process of training a bicycle prediction model may include:

[0136] Step 501: Select data from a single vehicle for range prediction, generate machine learning features, and begin model training.

[0137] Step 502: Generate a driving range prediction model for the current vehicle and send it to the current vehicle via OTA.

[0138] Specifically, the vehicle can receive the model through its onboard terminal.

[0139] Step 503: When any of the following conditions are met, the cloud updates the model, adds the new data to the dataset, or records the parameters of the model that have been adjusted manually:

[0140] The current vehicle has had its model automatically adjusted; the current vehicle has added more than N tens of thousands of kilometers of mileage; the current vehicle has added more than M charge-discharge cycles. M and N are integers greater than 0.

[0141] Specifically, the process of training a prediction model for a defined scenario may include:

[0142] Step 504: Based on data selected from multiple vehicles, start model training for different defined driving scenarios.

[0143] Step 505: Generate driving range prediction models for each scenario and send them to the designated vehicles as needed.

[0144] Step 506: When any of the following conditions are met, the model is updated in the cloud, and the new data is added to the dataset:

[0145] A new scene was identified; the amount of newly added data to the existing scene exceeded L tens of thousands. L is an integer greater than 0.

[0146] like Figure 6 The diagram shown illustrates the process of adjusting the prediction model on the vehicle side, including:

[0147] Step 601: The vehicle receives the range prediction model published by the cloud via OTA.

[0148] Step 602: The driving range prediction service initiates SOA service registration and establishes service subscription / publishing relationships with related services in other domains.

[0149] Related services mainly refer to services related to vehicle energy consumption characteristics, such as services in the vehicle body domain (air conditioning and window control), driver assistance domains (vehicle speed and location-related), and computing services with abundant computing resources, such as computing services in the driver assistance domain and infotainment / smart cockpit domain. Based on a well-established SOA service subscription, the range prediction service can obtain energy consumption, vehicle speed, location, or computing resource information published by these related services.

[0150] Step 603: Estimate the SoC (State of Charge) based on the battery status, and estimate the driving range based on the SoC and other relevant vehicle parameters, and display it on the screen.

[0151] Step 604: The vehicle determines whether the driving range prediction model needs to be updated based on certain conditions. For example, it can determine whether an update is needed by judging whether the prediction accuracy has dropped below a threshold. If it is below the threshold, an update is required.

[0152] For example, if the accuracy of the range prediction is less than 80% within a cumulative driving range of 30km (or within a range of 5kWh battery drain), the model needs to be updated.

[0153] If yes, proceed with the subsequent process. Otherwise, continue to determine based on the settings.

[0154] Step 605: Determine if the vehicle-side computing resources are sufficient.

[0155] If there are sufficient resources on the vehicle side, proceed to step 606; otherwise, proceed to step 613.

[0156] Step 606: Obtain computing resources for domains such as assisted driving and infotainment through the subscribed service to support model training in the vehicle and save the original model parameters.

[0157] Step 607: The vehicle uses the data that occurred in the current period to train the model and starts the timer.

[0158] Step 608: Determine whether a new model is generated when the timer expires.

[0159] Step 609: If yes, update the model on the vehicle side. Otherwise, proceed to step 613.

[0160] Step 610: Determine whether the accuracy of the new model is higher than the threshold.

[0161] Step 611: If yes, the vehicle will continue to use the updated model and send the new model parameters to the cloud according to the defined conditions.

[0162] The defined conditions may include, for example, stable communication between the vehicle and the cloud.

[0163] Step 612: If not, if the vehicle is currently using a cloud-based model, then revert to the original model.

[0164] If the vehicle is not currently using a cloud-based model, proceed to step 613.

[0165] Step 613: The vehicle sends a request to the cloud to update the model.

[0166] Step 614: Determine if there is a matching model in the cloud.

[0167] Step 615: If yes, the cloud will send the parameters of the matching model to the vehicle via the communication network.

[0168] Step 616: If not, the cloud sends a response message to the terminal, instructing the vehicle to revert to the original model.

[0169] The original model can refer to the model used by the vehicle before the model update.

[0170] like Figure 7 The diagram illustrates the process of a power domain subscribing to services from other domains. Figure 7 The service subscription process shown may include:

[0171] Step 701: The driving range estimation service shall be designed and implemented as needed.

[0172] Step 702: The driving range estimation service, driver assistance domain related services, and infotainment domain related services send service registration requests to the service management middleware.

[0173] Step 703: The service management middleware performs service registration management based on the service registration request.

[0174] Step 704: The driving range estimation service sends a subscription request to the service management middleware for assisted driving services and other related services, which are used to subscribe to driving status information and computing resource information of other domains.

[0175] Step 705: The service management middleware performs service subscription management.

[0176] Step 706: The service management middleware sends service subscription establishment requests to services related to the driver assistance domain and infotainment domain.

[0177] In steps 705 and 706, the service management middleware can query the management rules of the corresponding service based on the subscription requests of the assisted driving service and other related services, agree to forward the subscription request from step 704 to the corresponding service, or send a response rejecting the subscription request to the range estimation service.

[0178] If the forwarding is approved, the service management middleware can send a service subscription establishment request to the corresponding service. If the service management middleware receives a response from the corresponding service agreeing to the service subscription establishment request, it updates the service subscription status and sends a service subscription success response to the range estimation service.

[0179] Step 707: If there is dynamic service information to be published, the assisted driving domain related services and infotainment domain related services send a service publication response to the range estimation service, including information on their respective available computing resources.

[0180] Step 708: When the driving range estimation service needs to utilize computing resources from other domains to train the model, select an appropriate service based on the service release response information. In this embodiment, for example, services related to the assisted driving domain are selected for model training.

[0181] Step 709: The driving range estimation service sends training data for model training to the assisted driving domain related services and the infotainment domain related services.

[0182] Step 710: Train the model for services related to the assisted driving domain.

[0183] Step 711: The assisted driving domain related services send the model training results to the driving range estimation service.

[0184] In the above process, machine learning can typically achieve an accuracy of 90% in predicting driving range. When an electric vehicle displays a driving range of 400 kilometers at the start, and the actual accumulated driving distance from the start is 30 kilometers (which can be obtained from the vehicle's odometer), if the displayed driving range becomes between 367 and 373 kilometers, the accuracy can be considered to be above 90%. If we define the threshold for the decrease in prediction accuracy as 80%, then when the actual accumulated driving distance is 30 kilometers, if the displayed mileage is >376 or <364, the accuracy can be considered to be below 80%, i.e., below the threshold.

[0185] Assuming an electric vehicle primarily used in city A has a VID of 2, and the driving area is changed to a northern region during winter, the range prediction will be inaccurate, affecting charging planning and the travel experience. The main differences between geographical scenarios and daily driving scenarios are: significant changes in ambient temperature and ambient energy consumption. In this case, energy consumption increases significantly, and the accuracy of the original range prediction deteriorates significantly, requiring readjustment of the accuracy.

[0186] In practical applications, assuming an electric vehicle primarily used for urban commuting has a VID of 1, in scenarios where highway travel is predominant during holidays, the predicted driving range becomes inaccurate, impacting charging planning and the travel experience. The main differences between highway and urban commuting scenarios are: a significantly higher proportion of high speeds leading to increased wind resistance, higher overall vehicle weight, and potential changes in altitude. In these situations, energy consumption increases significantly, and the accuracy of the original driving range prediction deteriorates markedly, requiring readjustment. The vehicle detects that the driving range estimate accuracy is approximately 70%, below the threshold (defined as 90%), and initiates a model adjustment. Since the vehicle is currently using assisted driving functions at high speed, the computational resources in the assisted driving domain are being consumed, and the computational resources in the infotainment domain are temporarily insufficient to support driving range model training. Therefore, the vehicle interacts with the cloud to request model adjustment. After receiving parameters from the vehicle, such as location, speed, and total energy consumption over a certain number of kilometers (reflecting nearby energy consumption), the cloud sends similar model parameters to the vehicle via the communication network. After receiving the new model parameters, the vehicle backs up the original model and uses the new model to infer the remaining driving range. When the accuracy reaches 90%, the new model is maintained, and the vehicle user is notified via screen display and voice prompts that "Based on current usage, the driving range has been updated by A kilometers," reminding the vehicle to plan a charging session. Furthermore, it can assist in automatically adjusting power-intensive accessory functions such as air conditioning and seat ventilation, for example, by incorporating user interaction to achieve automatic adjustments or adjustments based on user preferences.

[0187] In practical applications, assuming the vehicle detects that the accuracy of the estimated driving range is approximately 80%, which is below the threshold (defined as 90%), a model adjustment is initiated. The driving range prediction service receives computing resource information from related services in the assisted driving domain and deems the computing resources sufficient to support driving range model training. The vehicle then attempts to initiate the model adjustment process locally (the current vehicle status may be parked or driving with assisted driving functions disabled). The vehicle uses recent driving data as training data to begin training the driving range estimation model. After obtaining the new model, the original model is backed up and the new model is used to infer the driving range. When the accuracy reaches 90%, the new model is used, and the vehicle user is notified via screen display and voice prompts that "Based on the current energy consumption in the area, the current driving range has been updated to A kilometers," reminding the vehicle to update its charging plan. Alternatively, at the end of the current driving trip, the user can be reminded that "Based on the current energy consumption in the area, the current model's fully charged driving range has been updated to B kilometers; please increase the charging frequency accordingly."

[0188] As can be seen from the above description, the solution using the embodiments of this application can not only ensure the accuracy of mileage prediction, but also, because it establishes service publishing and subscription relationships between domains based on SOA, the terminal can adjust the model not only locally, but also through the cloud. Therefore, this solution improves the efficiency of model updates, saves time, and is more flexible and convenient.

[0189] This application also provides a driving range prediction device, applied to an in-vehicle terminal. See also Figure 8 , Figure 8 This is a structural diagram of the driving range prediction device provided in the embodiments of this application. Figure 8 As shown, the driving range prediction device 800 includes:

[0190] The first acquisition module 801 is used to acquire a first prediction model; the first prediction module 802 is used to obtain a first driving range prediction value using the first prediction model; the second acquisition module 803 is used to adjust the first prediction model to obtain a second prediction model when the first driving range prediction value does not meet the first preset requirement; the second prediction module 804 is used to predict the driving range using the second prediction model.

[0191] Optional, such as Figure 9 As shown, the device further includes:

[0192] The first module 805 is used to establish service subscription and publication relationships with the target domain based on SOA.

[0193] The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain.

[0194] Optional, such as Figure 10 As shown, the first establishment module 805 may include:

[0195] The first sending submodule 8051 is used to send an assisted driving service subscription request to the service management middleware after registration is completed. The assisted driving service subscription request is used to subscribe to information on the available computing resources of the target domain from the target domain. The first receiving submodule 8052 is used to receive a service publishing message sent by the target domain. The service publishing message includes information on the available computing resources of the target domain.

[0196] Optional, such as Figure 11 As shown, the second acquisition module 803 may include:

[0197] The first acquisition submodule 8031 ​​is used to acquire available computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain;

[0198] The first processing submodule 8032 is used to train the model using the available computing resources if the available computing resources meet the second preset requirements, so as to obtain the second prediction model.

[0199] The second processing submodule 8033 is used to obtain the second prediction model from the cloud if the available computing resources do not meet the second preset requirements.

[0200] Optional, such as Figure 12 As shown, the second processing submodule 8033 may include:

[0201] The first sending unit 80331 is used to send a model acquisition request to the cloud.

[0202] The first receiving unit 80332 is used to receive model parameters sent by the cloud according to the model acquisition request, and update the model according to the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the vehicle terminal is located.

[0203] Optionally, the first processing submodule 8032 is configured to, if the available computing resources of the local domain meet the second preset requirement, use the available computing resources of the local domain to train the model and obtain the second prediction model; if the available computing resources of the target domain meet the second preset requirement, send training data to the target domain and obtain the second prediction model from the target domain, wherein the target domain trains the model based on the training data to obtain the second prediction model.

[0204] Optionally, the second prediction module 804 is used to predict the driving range using the second prediction model if the second prediction model is successfully obtained within a first preset time.

[0205] Optional, such as Figure 13 As shown, the device may further include:

[0206] The third acquisition module 806 is used to acquire the second prediction model from the cloud if the second prediction model is not successfully acquired within the first preset time.

[0207] Optional, such as Figure 14 As shown, the device may further include:

[0208] The first determining module 807 is used to determine whether the second driving range prediction value obtained by the second prediction model meets the first preset requirement.

[0209] The first processing module 808 is used to predict the remaining range using the second prediction model when the second predicted driving range value meets the first preset requirement.

[0210] The second processing module 809 is used to predict the driving range using the first prediction model if the first prediction model is obtained from the cloud when the second predicted driving range does not meet the first preset requirement; and to obtain the prediction model from the cloud if the first prediction model is not obtained from the cloud when the second predicted driving range does not meet the first preset requirement.

[0211] Optional, such as Figure 15 As shown, the device may further include:

[0212] The first sending module 810 is used to send the model parameters of the second prediction model to the cloud when the second driving range prediction value meets the first preset requirement.

[0213] The apparatus provided in this application embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, so it will not be described again here.

[0214] This application also provides a driving range prediction device, applied in the cloud. See also Figure 16 , Figure 16 This is a structural diagram of the driving range prediction device provided in the embodiments of this application. Figure 16 As shown, the driving range prediction device 1600 includes:

[0215] The first receiving module 1601 is used to receive the model acquisition request sent by the vehicle terminal;

[0216] The first sending module 1602 is used to send model parameters to the vehicle terminal according to the model acquisition request. The model parameters are used by the vehicle terminal to adjust the first prediction model to obtain the second prediction model.

[0217] The model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

[0218] Optional, such as Figure 17 As shown, the device further includes:

[0219] The generation module 1603 is used to generate a single-vehicle prediction model and a prediction model that defines the scenario;

[0220] The second sending module 1604 is used to send the single-vehicle prediction model to the vehicle terminal.

[0221] Optional, such as Figure 18 As shown, the device further includes:

[0222] The first update module 1605 is used to update the single-vehicle prediction model when at least one of the following conditions is met:

[0223] The adjusted model parameters sent by the vehicle terminal are received.

[0224] Based on the driving status information sent by the vehicle terminal, it is determined that the newly added mileage of the vehicle at the location of the vehicle terminal is greater than the first value.

[0225] Based on the driving status information sent by the vehicle terminal, it is determined that the number of new charge / discharge cycles for the vehicle at which the vehicle terminal is located is greater than the second value.

[0226] The first update module 1605 is used to update the prediction model of the defined scenario when at least one of the following conditions is met:

[0227] A new driving scenario has been identified;

[0228] The existing driving scenario data is greater than the third value.

[0229] The apparatus provided in this application embodiment can execute the above method embodiment, and its implementation principle and technical effect are similar, so it will not be described again here.

[0230] It should be noted that the division of units in the embodiments of this application is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0231] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0232] See Figure 19 , Figure 19 This is a structural diagram of an electronic device provided in an embodiment of the present invention. The electronic device may include a memory 1901, a processor 1902, and a program stored in the memory and executable on the processor. The processor 1902 is used to read the program in the memory to implement the steps in the aforementioned driving range prediction method.

[0233] This application also provides a readable storage medium storing a program. When executed by a processor, this program implements the various processes of the above-described driving range prediction method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here. The readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to magnetic storage (e.g., floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.), optical storage (e.g., CD, DVD, BD, HVD, etc.), and semiconductor storage (e.g., ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD), etc.).

[0234] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0235] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0236] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for predicting driving range, applied to an in-vehicle terminal, characterized in that, include: Obtain the first prediction model; The first predicted driving range is obtained using the first prediction model; When the first predicted driving range does not meet the first preset requirement, the first prediction model is adjusted to obtain a second prediction model, wherein the first preset requirement is the prediction accuracy requirement. The second prediction model is used to predict the driving range; Before adjusting the first prediction model to obtain the second prediction model, the method further includes: Based on the service-oriented architecture (SOA), a service subscription and publish relationship is established with the target domain. The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain. The establishment of service subscription and publication relationships with the target domain includes: After registration is completed in the service management middleware, an assisted driving service subscription request is sent to the service management middleware. The assisted driving service subscription request is used to subscribe to information on the available computing resources of the target domain from the target domain. Receive a service publishing message sent by the target domain, the service publishing message including information on the available computing resources of the target domain; The step of adjusting the first prediction model to obtain the second prediction model includes: Obtain available idle computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain; If the available idle computing resources meet the second preset requirement, then the model is trained using the available idle computing resources to obtain the second prediction model; wherein the second preset requirement is set according to whether the computing space is sufficient; and If the available idle computing resources do not meet the second preset requirement, then the second prediction model is obtained from the cloud. The step of obtaining the second prediction model from the cloud includes: sending a model acquisition request to the cloud; receiving model parameters sent by the cloud according to the model acquisition request, and updating the model according to the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

2. The method according to claim 1, characterized in that, If the available idle computing resources meet the second preset requirement, then the model is trained using the available idle computing resources to obtain the second prediction model, including: If the available computing resources in this domain meet the second preset requirement, then the available computing resources in this domain are used to train the model to obtain the second prediction model. If the available computing resources of the target domain meet the second preset requirement, then training data is sent to the target domain and the second prediction model is obtained from the target domain, wherein the target domain trains the model based on the training data to obtain the second prediction model.

3. The method according to claim 2, characterized in that, If the available computing resources of this domain are used for model training to obtain the second prediction model, or if training data is sent to the target domain and the second prediction model is obtained from the target domain, the step of using the second prediction model to predict the driving range includes: If the second prediction model is successfully obtained within the first preset time, the remaining driving range is predicted using the second prediction model.

4. The method according to claim 3, characterized in that, The method further includes: If the second prediction model is not successfully obtained within the first preset time, the second prediction model is obtained from the cloud.

5. The method according to claim 1, characterized in that, After predicting the driving range using the second prediction model, the method further includes: Determine whether the second predicted driving range obtained using the second prediction model meets the first preset requirement; When the second predicted driving range meets the first preset requirement, the second prediction model is used to predict the driving range. When the second predicted driving range does not meet the first preset requirement, if the first prediction model is obtained from the cloud, then the first prediction model is used to predict the driving range; when the second predicted driving range does not meet the first preset requirement, if the first prediction model is not obtained from the cloud, then the prediction model is obtained from the cloud.

6. The method according to claim 5, characterized in that, The method further includes: When the second predicted driving range meets the first preset requirement, the model parameters of the second prediction model are sent to the cloud.

7. A method for predicting driving range, applied in the cloud, characterized in that, include: Receive model acquisition requests sent by the vehicle-mounted terminal; According to the model acquisition request, the vehicle terminal sends model parameters, which are used to: when the first predicted driving range obtained by the first prediction model does not meet the first preset requirement, the vehicle terminal adjusts the first prediction model to obtain a second prediction model, wherein the first preset requirement is a prediction accuracy requirement. Before adjusting the first prediction model to obtain the second prediction model, the method further includes: Based on the service-oriented architecture (SOA), a service subscription and publish relationship is established with the target domain. The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain. The establishment of service subscription and publication relationships with the target domain includes: After registration is completed in the service management middleware, an assisted driving service subscription request is sent to the service management middleware. The assisted driving service subscription request is used to subscribe to information on the available computing resources of the target domain from the target domain. Receive a service publishing message sent by the target domain, the service publishing message including information on the available computing resources of the target domain; The step of adjusting the first prediction model to obtain the second prediction model includes: Obtain available idle computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain; If the available idle computing resources meet the second preset requirement, then the model is trained using the available idle computing resources to obtain the second prediction model; wherein the second preset requirement is set according to whether the computing space is sufficient; and If the available idle computing resources do not meet the second preset requirement, then the second prediction model is obtained from the cloud. The step of obtaining the second prediction model from the cloud includes: sending a model acquisition request to the cloud; receiving model parameters sent by the cloud according to the model acquisition request, and updating the model according to the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

8. The method according to claim 7, characterized in that, The method further includes: Generate a single-vehicle prediction model and define a scenario prediction model; The single-vehicle prediction model is sent to the vehicle terminal.

9. The method according to claim 8, characterized in that, The method further includes: The single-vehicle prediction model is updated when at least one of the following conditions is met: The adjusted model parameters sent by the vehicle terminal are received. Based on the driving status information sent by the vehicle terminal, it is determined that the newly added mileage of the vehicle at the location of the vehicle terminal is greater than the first value. Based on the driving status information sent by the vehicle terminal, it is determined that the number of new charge / discharge cycles for the vehicle at which the vehicle terminal is located is greater than the second value.

10. The method according to claim 8, characterized in that, The method further includes: The prediction model for the defined scenario is updated when at least one of the following conditions is met: A new driving scenario has been identified; The existing driving scenario data is greater than the third value.

11. A driving range prediction device, applied to an in-vehicle terminal, characterized in that, include: The first acquisition module is used to acquire the first prediction model; The first prediction module is used to obtain a first predicted driving range value using the first prediction model. The second acquisition module is used to adjust the first prediction model to obtain a second prediction model when the first predicted driving range does not meet the first preset requirement, wherein the first preset requirement is a prediction accuracy requirement. The second prediction module is used to predict the driving range using the second prediction model. The device further includes: The first module is used to establish service subscription and publication relationships with the target domain based on SOA. The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain. The first establishment module includes: The first sending submodule is used to send an assisted driving service subscription request to the service management middleware after registration is completed. The assisted driving service subscription request is used to subscribe to information on available computing resources of the target domain from the target domain. The first receiving submodule is configured to receive a service publishing message sent by the target domain, the service publishing message including information on the available computing resources of the target domain; The step of adjusting the first prediction model to obtain the second prediction model includes: Obtain available idle computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain; If the available idle computing resources meet the second preset requirement, then the model is trained using the available idle computing resources to obtain the second prediction model; wherein the second preset requirement is set according to whether the computing space is sufficient; and If the available idle computing resources do not meet the second preset requirement, then the second prediction model is obtained from the cloud. The step of obtaining the second prediction model from the cloud includes: sending a model acquisition request to the cloud; receiving model parameters sent by the cloud according to the model acquisition request, and updating the model according to the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

12. A driving range prediction device, applied in the cloud, characterized in that, include: The first receiving module is used to receive model acquisition requests sent by the vehicle terminal; The first sending module is used to send model parameters to the vehicle terminal according to the model acquisition request. The model parameters are used to: when the first predicted driving range obtained by the first prediction model does not meet the first preset requirement, the vehicle terminal adjusts the first prediction model to obtain a second prediction model, wherein the first preset requirement is a prediction accuracy requirement. The device further includes: The first module is used to establish service subscription and publication relationships with the target domain based on SOA. The target domain includes one or more of the following domains: chassis domain, body domain, driver assistance domain, and infotainment domain. The first establishment module includes: The first sending submodule is used to send an assisted driving service subscription request to the service management middleware after registration is completed. The assisted driving service subscription request is used to subscribe to information on available computing resources of the target domain from the target domain. The first receiving submodule is configured to receive a service publishing message sent by the target domain, the service publishing message including information on the available computing resources of the target domain; The step of adjusting the first prediction model to obtain the second prediction model includes: Obtain available idle computing resources, wherein the available idle computing resources include available computing resources in the current domain or available computing resources in the target domain; If the available idle computing resources meet the second preset requirement, then the model is trained using the available idle computing resources to obtain the second prediction model; wherein the second preset requirement is set according to whether the computing space is sufficient; and If the available idle computing resources do not meet the second preset requirement, then the second prediction model is obtained from the cloud. The step of obtaining the second prediction model from the cloud includes: sending a model acquisition request to the cloud; receiving model parameters sent by the cloud according to the model acquisition request, and updating the model according to the model parameters to obtain the second prediction model, wherein the model parameters correspond to the driving scenario of the vehicle where the in-vehicle terminal is located.

13. The apparatus according to claim 12, characterized in that, The device further includes: The generation module is used to generate prediction models for individual vehicles and prediction models that define scenarios; The second sending module is used to send the single-vehicle prediction model to the vehicle terminal.

Images