A scene self-adaption based new energy vehicle hierarchical wake-up method

Abstract

This invention discloses a scenario-adaptive, graded wake-up method for new energy vehicles, relating to the field of new energy vehicle control technology. The method acquires vehicle status and environmental data, identifies vehicle scenarios and calculates travel intention confidence; calculates the remaining battery capacity and health, and predicts the average static current; calculates the battery's sustainability and determines the vehicle's risk level; extracts the current scenario's preset target sustainability and target average current, and calculates energy consumption health by combining the sustainability and predicted average static current; determines a basic wake-up level based on the risk level and energy consumption health, and corrects it with travel intention confidence to obtain the final wake-up level; based on the final wake-up level, the complete wake-up set is pruned to generate a minimum wake-up set, and the baseline sleep delay time is corrected with travel intention confidence to determine the final sleep delay time, executing a wake-up or sleep command. This invention effectively reduces the risk of ineffective battery consumption and depletion when the vehicle is not in motion.

Description

Technical Field

[0001] This invention relates to a scenario-adaptive hierarchical wake-up method for new energy vehicles, belonging to the field of new energy vehicle control technology. Background Technology

[0002] As the intelligence level of new energy vehicles continues to improve, the number of domain controllers and overall power consumption of the vehicles have increased significantly. During prolonged parking or standby, vehicles typically periodically power on all domain controllers to collect relevant data to assess the risk of battery depletion, maintain a fixed standby time, and then re-enter sleep mode. However, this strategy has significant shortcomings: existing vehicle wake-up strategies are often rather crude, employing a one-time full-domain wake-up that also wakes up high-power domain controllers such as those for cockpit entertainment, resulting in significant unnecessary power waste. Furthermore, the standby time after wake-up is fixed and cannot be dynamically adjusted according to vehicle scenarios or user travel intentions. When the battery level is low, this can easily exacerbate depletion and, in severe cases, affect the vehicle's normal starting. Summary of the Invention

[0003] The purpose of this invention is to provide a scenario-adaptive graded wake-up method for new energy vehicles, which aims to solve the technical problems of existing vehicle wake-up strategies being crude and maintaining a fixed standby time after wake-up, resulting in a large amount of unnecessary power consumption waste and battery depletion when the vehicle is not in motion.

[0004] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution.

[0005] This invention provides a scenario-adaptive hierarchical wake-up method for new energy vehicles, comprising:

[0006] Obtain structured vehicle status data and environmental data, wherein the vehicle status data includes basic battery operating data and vehicle static current;

[0007] Based on vehicle status data and environmental data, the current vehicle scenario is identified through a preset scenario classifier, and the user's travel intention confidence is calculated through a preset inference model.

[0008] The remaining capacity and health of the battery are calculated based on the basic operating data of the battery. Based on the current vehicle scenario, the vehicle static current, environmental data, the basic operating data of the battery, and the calculated battery health, the average static current prediction value within a preset time window is predicted using the static current prediction model.

[0009] By combining the remaining battery capacity, battery health, and average static current prediction, the current battery lifespan is calculated, and the vehicle risk level is determined.

[0010] Extract the preset target maintenance duration and preset target average current corresponding to the current vehicle scenario, and calculate the energy health status by combining the maintenance duration and the predicted average static current value.

[0011] The basic wake-up level is determined based on the vehicle risk level and energy consumption health, and the basic wake-up level is offset and corrected based on the confidence level of travel intention to obtain the final wake-up level.

[0012] Based on the final wake-up level, the pre-stored complete wake-up set is retrieved and dynamically trimmed to generate a minimum wake-up set. The baseline sleep delay time corresponding to the current vehicle scenario is then corrected based on the travel intention confidence level to determine the final sleep delay time. Finally, the corresponding wake-up or sleep command is executed on the vehicle.

[0013] Optionally, based on vehicle status data and environmental data, the current vehicle scene is identified using a preset scene classifier, including:

[0014] Extract feature parameters from vehicle status data and environmental data, including at least: power start-stop status, charging status, door lock status, remote APP command status, network command status, occupancy status of occupants, system time information, and ambient light intensity.

[0015] The extracted feature parameters are input into a pre-built decision tree-based scene classifier, and the current vehicle scene is output according to the preset feature combination judgment rules.

[0016] The current vehicle scenarios include fast charging scenarios, slow charging scenarios, remote control scenarios, preset departure scenarios, parking and waiting scenarios, and nighttime scenarios.

[0017] Optionally, the confidence level of the user's travel intention can be calculated using a preset inference model, including:

[0018] Extract system time information, user historical behavior data, remote APP command status, in-vehicle occupancy status and Bluetooth key signal strength from vehicle status data and environmental data to construct a multi-dimensional intent feature vector;

[0019] The multidimensional intent feature vector is input into a pre-built inference model based on logistic regression algorithm, and the corresponding probability value is calculated and output as the confidence level of the travel intent.

[0020] Based on the preset confidence threshold, the confidence level of travel intention is divided into three levels: high confidence, medium confidence, and low confidence.

[0021] Optionally, the remaining battery capacity and battery health can be calculated based on the battery's basic operating data, including:

[0022] Obtain the battery voltage, battery internal resistance, battery temperature, charging and discharging current, and current maximum usable capacity of the battery from the basic operating data of the battery;

[0023] The remaining battery capacity is obtained by performing coulomb integration calculation based on the charging and discharging current, and by combining the battery voltage and battery temperature calibration integration results.

[0024] Battery health is calculated using the following formula. :

[0025] ;

[0026] in, These are the weighting coefficients obtained from the offline calibration of the entire vehicle; This is the preset internal resistance at the end of the battery's lifespan. This represents the current internal resistance of the battery. The preset nominal internal resistance of the new battery; This represents the battery's current maximum usable capacity. This is the preset nominal capacity of the battery.

[0027] Optionally, the static current prediction model is an XGBoost regression model, whose input features include:

[0028] Scene category features generated based on the current vehicle scene;

[0029] Static current hysteresis characteristics, static current rolling statistical characteristics, and static current trend characteristics obtained from the static current conversion of the whole vehicle;

[0030] Environmental state characteristics obtained from environmental data transformation;

[0031] The battery status characteristics are composed of the battery voltage from the basic battery operating data and the calculated battery health status.

[0032] Optionally, calculate the current battery life and determine the vehicle risk level, including:

[0033] The remaining battery life can be calculated using the following formula. :

[0034] ;

[0035] in, This refers to the remaining capacity of the battery. For battery health; The average static current is the predicted value; the coefficient 0.7 is the preset safety margin for the depth of discharge.

[0036] Duration and average static current prediction value The risk level of the vehicle is determined by comparing it with the preset safe time threshold and safe current threshold within a range.

[0037] Optionally, the energy consumption health status Calculated using the following formula:

[0038] ;

[0039] in, The preset target duration for the current vehicle scenario. The battery's lifespan; The preset target average static current corresponding to the current vehicle scenario. This is the predicted value of the average static current.

[0040] Optionally, the process of determining the base wake-up level and the final wake-up level includes:

[0041] The basic wake-up level is determined by matching the energy consumption health status within a preset energy consumption health threshold range and combining this with the vehicle risk level assessment. ;

[0042] The corresponding integer level correction step size is determined based on the confidence level of the travel intention. The correction step size is 1 for high confidence, 0 for medium confidence, and -1 for low confidence.

[0043] The final wake-up level is calculated using the following boundary constraint formula. :

[0044] ;

[0045] in, This is the preset lower limit for the wake-up level; This is the preset maximum wake-up level.

[0046] Optionally, all vehicle scenarios are pre-configured with a unified complete wake-up set, which consists of the cockpit domain controller, powertrain domain controller, body domain controller, and BMS domain controller. Based on the final wake-up level, this complete wake-up set is dynamically trimmed to generate a minimum wake-up set, including:

[0047] Based on the final wake-up level value, it is divided into high priority level, medium priority level and low priority level;

[0048] For high-priority levels, the cockpit domain controller is switched to a preset low-power standby mode, while the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set.

[0049] For medium priority levels, the cockpit domain controllers in the complete wake-up set are removed, and only the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set.

[0050] For low-priority levels, the cockpit domain controller and powertrain domain controller are removed from the complete wake-up set, and only the body domain controller and BMS domain controller are retained as the minimum wake-up set.

[0051] Optionally, the baseline sleep delay time corresponding to the current vehicle scenario is adjusted based on the travel intention confidence level to determine the final sleep delay time, including:

[0052] Obtain the baseline sleep delay duration corresponding to the current vehicle scenario;

[0053] Based on the confidence level of travel intention, a corresponding dormancy correction coefficient is matched, where the higher the confidence level of travel intention, the larger the corresponding dormancy correction coefficient;

[0054] The final sleep delay is calculated by multiplying the baseline sleep delay duration by the sleep correction coefficient.

[0055] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0056] This invention quantifies wake-up demands into specific vehicle scenarios and travel intention confidence levels through a scenario classifier and inference model, providing a basis for subsequent dynamic decision-making and avoiding crude wake-up from the outset. A static current prediction model predicts future average static current values, thereby deriving the battery's sustainability and vehicle risk level, enabling proactive perception of power consumption trends. A basic wake-up level is determined by integrating "vehicle risk level" and "energy health," and then offset using "travel intention confidence" to obtain the final wake-up level, achieving fine-grained control and avoiding misjudgments based on a single indicator. The pre-stored complete wake-up set is dynamically pruned based on the final wake-up level to generate a minimum wake-up set, changing the traditional full-domain wake-up model. Simultaneously, the baseline sleep delay corresponding to the current vehicle scenario is corrected using travel intention confidence to determine the final sleep delay duration, thereby dynamically adjusting the standby time of the woken domain controllers. This invention, while ensuring a good user experience, eliminates unnecessary domain controllers to the greatest extent possible, effectively solving the problems of unnecessary static power consumption waste and accelerated battery depletion when the vehicle is not in motion. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating the scenario-adaptive graded wake-up method for new energy vehicles provided in an embodiment of the present invention. Detailed Implementation

[0058] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0059] Example 1

[0060] This embodiment introduces a scenario-adaptive hierarchical wake-up method for new energy vehicles. This method operates in the vehicle's non-real-time control layer, and references... Figure 1 This method includes the following steps:

[0061] S1. Multi-source data acquisition and preprocessing;

[0062] When the vehicle is off and not in motion, each domain controller is in sleep mode. During a preset operating cycle, it acquires structured vehicle status and environmental data for the current moment, including the following characteristic parameters: power start / stop status, charging status (charging gun connection status, charging power status), door lock status, remote APP command status, network command status, occupant status, Bluetooth key signal strength, user historical behavior data, system time information, ambient light intensity, vehicle static current, and basic battery operating data. The basic battery operating data includes battery voltage, battery internal resistance, battery temperature, charging / discharging current, and the battery's current maximum available capacity. The structured vehicle status and environmental data within the vehicle are derived from raw data reported by various sensors and buses, obtained by aligning timestamps uniformly on the timeline and filling in missing data using linear interpolation.

[0063] To eliminate abrupt changes and noise in specific physical parameters of structured data caused by signal transmission, this method employs a data protection mechanism to perform anomaly filtering on the corresponding specific feature parameters in the structured vehicle state data and environmental data: if the rate of change of the vehicle static current at three consecutive sampling points exceeds 200% / s, it is determined to be an invalid abrupt change, and the valid value at the previous sampling time is used for replacement; if the battery voltage exceeds the preset first voltage threshold and second voltage threshold range, it is marked as invalid, and the previous valid value is used for filtering and replacement.

[0064] S2. Vehicle scenario classification and user travel confidence calculation;

[0065] Based on vehicle status data and environmental data, the current vehicle scenario is identified through a preset scenario classifier, and the user's travel intention confidence is calculated through a preset inference model.

[0066] S2.1 Vehicle scenario classification;

[0067] The vehicle's start / stop status, charging status, door lock status, remote APP command status, network command status, occupant status, system time information, and ambient light intensity are input into a pre-built decision tree scene classifier. The current vehicle scene is determined based on the following feature combination rules:

[0068] If the engine is off in the start-stop state, the charging status shows that the charging gun is connected, and the charging power is higher than the preset first power threshold, then the vehicle is determined to be in a fast charging scenario.

[0069] If the engine is off in the start-stop state, the charging status shows that the charging gun is connected, and the charging power is lower than the preset second power threshold, then the vehicle is determined to be in a slow charging scenario.

[0070] If the engine is off, the doors are locked, and the remote APP command status shows that an APP control command has been received and is pending, then the vehicle is determined to be in a remote control scenario.

[0071] If the engine is off when the start-stop system is on, and the user has preset a departure time, and the current system time information is within the preset departure time window (within 30 minutes before departure), then the vehicle is determined to be in the preset departure scenario.

[0072] If the engine is off, the doors are unlocked, and there are people inside the vehicle, then the vehicle is considered to be in a parking waiting scenario.

[0073] If the engine is off, the doors are locked, there are no commands in the current network command status and remote APP command status, the ambient light intensity is below the threshold, and the current system time information is in the nighttime period, then the current scene is determined to be nighttime.

[0074] If none of the above rules are met, the current vehicle scene is determined to be in the default scene of normal static.

[0075] S2.2 Calculation of confidence level of travel intention;

[0076] Based on system time information, user historical behavior data, remote APP command status, vehicle occupancy status, and Bluetooth key signal strength from vehicle status data and environmental data, a multi-dimensional intent feature vector is constructed. The multi-dimensional intent feature vector is then input into a pre-built inference model based on a logistic regression algorithm, and the corresponding probability value is calculated and output as the confidence level of the travel intent.

[0077] Multidimensional intent feature vectors include:

[0078] First characteristic The absolute value of the difference between the current system time and the historical travel habit time (from user historical behavior data), expressed by a function. Mapped to The interval, where, For calibrable coefficients, This is the absolute value of the difference.

[0079] Second feature : Whether it is a remote APP command status (such as a remote air conditioner start command). If yes, it is 1; otherwise, it is 0.

[0080] Third characteristic : Whether the driver's seat in the vehicle occupancy status changes from empty to occupied. If it remains occupied for 10 seconds, it is judged as 1; otherwise, it is 0.

[0081] Fourth feature The first-order differential signal strength of the Bluetooth key is evaluated by whether it increases beyond a certain range within 5 seconds; if yes, it is evaluated as 1, otherwise as 0.

[0082] Fifth feature : Is the current system time within 30 minutes before the preset departure time? If yes, it is judged as 1; otherwise, it is 0.

[0083] Calculate weighted scores , These are the weight coefficients for the logistic regression algorithm; in order to adapt to the computing power limitations of the vehicle's underlying controller and reduce computational power consumption, the weight coefficients are represented and calculated using the Q16 fixed-point number format; This is the intercept term constant obtained during offline training and calibration of the logistic regression algorithm, used to characterize the basic intent noise of the vehicle when all multidimensional intent feature vectors are 0. Then, the weighted scores are calculated using a pre-computed sigmoid lookup table. The confidence level mapped to a range of 0 to 1 is used as the confidence level of the travel intention.

[0084] Based on the preset confidence threshold, the confidence level of travel intention is divided into three levels: high confidence, medium confidence, and low confidence.

[0085] S3. Battery status calculation and static current trend prediction;

[0086] S3.1 Calculate the remaining capacity and health of the battery based on the basic operating data of the battery;

[0087] The remaining battery capacity is calculated by integrating the battery charging and discharging current using a coulomb counter, and then calibrating the integration results using the battery voltage and temperature. ;

[0088] The battery health is calculated using the collected battery voltage and the battery's current maximum usable capacity, according to the following formula. :

[0089] ;

[0090] in, These are the weighting coefficients obtained from the offline calibration of the entire vehicle; This is the preset internal resistance at the end of the battery's lifespan. This represents the current internal resistance of the battery. The preset nominal internal resistance of the new battery; This represents the battery's current maximum usable capacity. This is the preset nominal capacity of the battery.

[0091] S3.2 Calculate the predicted average static current value;

[0092] Based on the current vehicle scenario, vehicle static current, environmental data, basic battery operating data, and calculated battery health, the average static current prediction value within a preset time window is predicted using a static current prediction model.

[0093] In this method, the static current prediction model is an XGBoost regression model, and its input features include the following 23-dimensional feature vector:

[0094] The scene category features generated based on the current vehicle scene are in 7 dimensions: the current vehicle scene (6 categories and default scene) identified by S2 is one-hot encoded to obtain 7-dimensional binary features.

[0095] Based on the static current hysteresis characteristics, static current rolling statistical characteristics, and static current trend characteristics obtained from the vehicle static current conversion:

[0096] Static current hysteresis characteristics, in 8 dimensions: vehicle static current at the 1st, 2nd, 3rd, 4th, 5th, 6th, 10th, and 15th minutes in the past;

[0097] The static current rolling statistical characteristics consist of four dimensions: the moving mean, standard deviation, maximum value, and minimum value of the vehicle's static current over the past 5 minutes.

[0098] Static current trend characteristics, 1 dimension: linear fitting of the vehicle static current value over the past 3 minutes, and taking the slope of the fitted line.

[0099] The environmental state features obtained from the transformation of environmental data consist of one dimension: the current ambient temperature value.

[0100] The battery state characteristics consist of two dimensions, namely the battery voltage from the basic battery operation data and the calculated battery health status (SOH).

[0101] The static current prediction model performs regression calculations on the 23-dimensional feature vector constructed above, and outputs the average static current prediction value within a preset time window (e.g., the next 10 minutes). .

[0102] S4. Calculate the current battery life and determine the vehicle risk level;

[0103] The remaining battery life can be calculated using the following formula. :

[0104] ;

[0105] in, This refers to the remaining capacity of the battery. For battery health; The average static current is the predicted value; the coefficient 0.7 is the preset safety margin for the depth of discharge.

[0106] Duration and average static current prediction value The risk level of the vehicle is determined by comparing it with the preset safe time threshold and safe current threshold within a range.

[0107] For example: according to and Define risk level :when More than 24 hours and A reading less than 30mA (this threshold can be adjusted according to actual conditions) is a Level 0 warning; when... Between 12 and 24 hours or A level 1 warning is issued when the value is between 30 and 60 mA; when Less than 12 hours or A value greater than 60 mA triggers a Level 2 warning.

[0108] S5. Extract the preset target maintenance duration and preset target average current corresponding to the current vehicle scenario, and calculate the energy health status by combining the maintenance duration and the predicted average static current value.

[0109] Energy health Calculated using the following formula:

[0110] ;

[0111] in, The preset target duration for the current vehicle scenario. The battery's lifespan; The preset target average static current corresponding to the current vehicle scenario. This is the predicted value of the average static current.

[0112] The calculation benchmark for energy health varies across different scenarios. This method predefines the target maintenance duration for different vehicle scenarios. Target average static current The target maintenance duration for each scenario reflects the battery safety redundancy requirements for that scenario, while the target average quiescent current reflects the stringency of energy consumption control.

[0113] S6. Determine the basic wake-up level based on the vehicle risk level and energy consumption health, and make an offset correction on the basic wake-up level based on the travel intention confidence to obtain the final wake-up level.

[0114] S6.1 Determine the basic wake-up level;

[0115] The basic wake-up level is determined by matching the energy consumption health status within a preset energy consumption health threshold range and combining this with the vehicle risk level assessment. ;

[0116] The basic wake-up levels include 0, 1, and 2, as detailed below:

[0117] Greater than or equal to the first health threshold and When the value is 0, the base awakening level is 2;

[0118] Less than the first health threshold and greater than or equal to the second health threshold, or When the value is 1, the base awakening level is 1;

[0119] Less than the second health threshold, or When the value is 2, the base awakening level is 0.

[0120] The basic wake-up level represents the quality of the battery's physical condition.

[0121] S6.2, Offset correction yields the final wake-up level;

[0122] The corresponding integer level correction step size is determined based on the confidence level of the travel intention. The correction step size is 1 for high confidence, 0 for medium confidence, and -1 for low confidence.

[0123] The final wake-up level is calculated using the following boundary constraint formula. :

[0124] ;

[0125] in, This is the preset lower limit for the wake-up level; in this example, it is set to 0. This is the preset maximum wake-up level value; in this example, it is set to 3.

[0126] This offset correction mechanism enables fine-grained control by "determining the basic wake-up level based on the scene and making fine adjustments when the intent is displayed".

[0127] S7. Based on the final wake-up level, retrieve the pre-stored complete wake-up set for dynamic trimming to generate the minimum wake-up set, and combine the travel intention confidence to correct the baseline sleep delay time corresponding to the current vehicle scenario, determine the final sleep delay time, and execute the corresponding wake-up or sleep command on the vehicle.

[0128] S7.1 Generate the minimum wake-up set;

[0129] All vehicle scenarios are pre-configured with a unified complete wake-up set, which consists of the cockpit domain controller, powertrain domain controller, body domain controller, and BMS domain controller. Based on the final wake-up level, this complete wake-up set is dynamically trimmed to generate a minimum wake-up set, including:

[0130] Based on the final wake-up level value, it is divided into high priority level, medium priority level and low priority level;

[0131] For high-priority levels, the cockpit domain controller is switched to a preset low-power standby mode, while the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set.

[0132] Specifically, the low-power standby mode is manifested by removing the enable permissions for peripheral loads attached to the cockpit domain controller, such as disabling peripherals like displays and speakers.

[0133] For medium priority levels, the cockpit domain controllers in the complete wake-up set are removed, and only the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set.

[0134] The removal of the cockpit domain controller is manifested in that the system no longer sends network wake-up commands to it.

[0135] For low-priority levels, the cockpit domain controller and powertrain domain controller are removed from the complete wake-up set, and only the body domain controller and BMS domain controller are retained as the minimum wake-up set.

[0136] S7.2 Determine the final hibernation delay duration;

[0137] The baseline sleep delay duration corresponding to the current vehicle scenario is obtained by looking up a table.

[0138] Based on the confidence level of travel intention, a corresponding dormancy correction coefficient is matched, with the specific rules as follows:

[0139] A correction factor of 1.5 corresponds to a high confidence level, 1.0 corresponds to a medium confidence level, and 0.5 corresponds to a low confidence level.

[0140] The final sleep delay is calculated by multiplying the baseline sleep delay duration by the sleep correction coefficient.

[0141] For example, in a slow charging scenario, multiplying 15 minutes by a low confidence level of 0.5 results in a sleep latency reduction to 7.5 minutes.

[0142] S7.3 Execute the corresponding wake-up or sleep command;

[0143] Based on the minimum wake-up set and the final sleep delay duration, during this wake-up cycle, only domain controllers within the set will be woken up, and they will remain in standby mode for the final sleep delay duration; domain controllers removed from the set will not be woken up and will remain in sleep mode.

[0144] When the final sleep delay ends, a sleep command is uniformly issued to the domain controllers currently in standby mode, causing them to enter low-power sleep mode.

[0145] In summary, the present invention obtains the minimum wake-up set through dynamic pruning and corrects the sleep delay time using travel intention confidence. While ensuring the user's driving experience, it eliminates the invalid standby power consumption of unnecessary domain controllers to the greatest extent and reasonably adjusts the standby waiting time after wake-up. This fundamentally realizes the fine-grained control of static power consumption in non-driving states and effectively avoids unnecessary power waste and the problem of aggravated battery depletion.

[0146] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0148] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention 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 the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A scenario-adaptive hierarchical wake-up method for new energy vehicles, characterized in that, include: Obtain structured vehicle status data and environmental data, wherein the vehicle status data includes basic battery operating data and vehicle static current; Based on vehicle status data and environmental data, the current vehicle scenario is identified through a preset scenario classifier, and the user's travel intention confidence is calculated through a preset inference model. Calculate the remaining battery capacity and battery health based on the basic operating data of the battery. Based on the current vehicle scenario, vehicle static current, environmental data, basic battery operating data, and calculated battery health, the average static current prediction value within a preset time window is predicted using a static current prediction model. By combining the remaining battery capacity, battery health, and average static current prediction, the current battery lifespan is calculated, and the vehicle risk level is determined. Extract the preset target maintenance duration and preset target average current corresponding to the current vehicle scenario, and calculate the energy health status by combining the maintenance duration and the predicted average static current value. The basic wake-up level is determined based on the vehicle risk level and energy consumption health, and the basic wake-up level is offset and corrected based on the confidence level of travel intention to obtain the final wake-up level. Based on the final wake-up level, the pre-stored complete wake-up set is retrieved and dynamically trimmed to generate the minimum wake-up set. The baseline sleep delay time corresponding to the current vehicle scenario is then corrected based on the travel intention confidence level to determine the final sleep delay time. Finally, the corresponding wake-up or sleep command is executed on the vehicle.

2. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, Based on vehicle status data and environmental data, the current vehicle scene is identified using a pre-defined scene classifier, including: Extract feature parameters from vehicle status data and environmental data, including at least: power start-stop status, charging status, door lock status, remote APP command status, network command status, occupancy status of occupants, system time information, and ambient light intensity. The extracted feature parameters are input into a pre-built decision tree-based scene classifier, and the current vehicle scene is output according to the preset feature combination judgment rules. The current vehicle scenarios include fast charging scenarios, slow charging scenarios, remote control scenarios, preset departure scenarios, parking and waiting scenarios, and nighttime scenarios.

3. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, The confidence level of a user's travel intention is calculated using a pre-defined inference model, including: Extract system time information, user historical behavior data, remote APP command status, in-vehicle occupancy status and Bluetooth key signal strength from vehicle status data and environmental data to construct a multi-dimensional intent feature vector; The multidimensional intent feature vector is input into a pre-built inference model based on logistic regression algorithm, and the corresponding probability value is calculated and output as the confidence level of the travel intent. Based on the preset confidence threshold, the confidence level of travel intention is divided into three levels: high confidence, medium confidence, and low confidence.

4. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, Calculate the remaining battery capacity and battery health based on the battery's basic operating data, including: Obtain the battery voltage, battery internal resistance, battery temperature, charging and discharging current, and current maximum usable capacity of the battery from the basic operating data of the battery; The remaining battery capacity is obtained by performing coulomb integration calculation based on the charging and discharging current, and by combining the battery voltage and battery temperature calibration integration results. Battery health is calculated using the following formula. : ； in, These are the weighting coefficients obtained from the offline calibration of the entire vehicle. This is the preset internal resistance at the end of the battery's lifespan. This represents the current internal resistance of the battery. The preset nominal internal resistance of the new battery; This represents the battery's current maximum usable capacity. This is the preset nominal capacity of the battery.

5. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 4, characterized in that, The static current prediction model is an XGBoost regression model, and its input features include: Scene category features generated based on the current vehicle scene; Static current hysteresis characteristics, static current rolling statistical characteristics, and static current trend characteristics obtained from the static current conversion of the whole vehicle; Environmental state characteristics obtained from environmental data transformation; The battery status characteristics are composed of the battery voltage from the basic battery operating data and the calculated battery health status.

6. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, Calculate the current battery life and determine the vehicle risk level, including: The remaining battery life can be calculated using the following formula. : ； in, This refers to the remaining capacity of the battery. For battery health; The average static current is the predicted value; the coefficient 0.7 is the preset safety margin for the depth of discharge. Duration and average static current prediction value The risk level of the vehicle is determined by comparing it with the preset safe time threshold and safe current threshold within a range.

7. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, Energy consumption health Calculated using the following formula: ； in, The preset target duration for the current vehicle scenario. The battery's lifespan; The preset target average static current corresponding to the current vehicle scenario. This is the predicted value of the average static current.

8. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 3, characterized in that, The process of determining the base wake-up level and the final wake-up level includes: The basic wake-up level is determined by matching the energy consumption health status within a preset energy consumption health threshold range and combining this with the vehicle risk level assessment. ; The corresponding integer level correction step size is determined based on the confidence level of the travel intention. The correction step size is 1 for high confidence, 0 for medium confidence, and -1 for low confidence. The final wake-up level is calculated using the following boundary constraint formula. : ； in, This is the preset lower limit for the wake-up level; This is the preset maximum wake-up level.

9. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, All vehicle scenarios are pre-configured with a unified complete wake-up set, which consists of the cockpit domain controller, powertrain domain controller, body domain controller, and BMS domain controller. Based on the final wake-up level, this complete wake-up set is dynamically trimmed to generate a minimum wake-up set, including: Based on the final wake-up level value, it is divided into high priority level, medium priority level and low priority level; For high-priority levels, the cockpit domain controller is switched to a preset low-power standby mode, while the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set. For medium priority levels, the cockpit domain controllers in the complete wake-up set are removed, and only the power domain controller, body domain controller, and BMS domain controller are retained as the minimum wake-up set. For low-priority levels, the cockpit domain controller and powertrain domain controller are removed from the complete wake-up set, and only the body domain controller and BMS domain controller are retained as the minimum wake-up set.

10. The scenario-adaptive hierarchical wake-up method for new energy vehicles according to claim 1, characterized in that, The baseline sleep delay time corresponding to the current vehicle scenario is adjusted based on the travel intention confidence level to determine the final sleep delay time, including: Obtain the baseline sleep delay duration corresponding to the current vehicle scenario; Based on the confidence level of travel intention, a corresponding dormancy correction coefficient is matched, where the higher the confidence level of travel intention, the larger the corresponding dormancy correction coefficient; The final sleep delay duration is calculated by multiplying the baseline sleep delay duration by the sleep correction coefficient.

Images