Vehicle energy consumption query method, device, equipment, medium and product

Abstract

The application discloses a vehicle energy consumption query method, device, equipment, medium and product. The method comprises the following steps: performing intention analysis on a voice wake-up instruction of a user to obtain a user intention type and slot information; if the user intention type is energy consumption query, a query data type is obtained; according to the query data type, a target data source matched with the query data type is selected from a plurality of preset data sources, and multi-dimensional energy consumption data of a vehicle is obtained from the target data source; and the multi-dimensional energy consumption data of the vehicle is fused and encapsulated, and an energy consumption query result is output in a natural language. The application improves the convenience, accuracy and driving safety of energy consumption query, and improves the intelligent interaction experience of a new energy vehicle.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and more specifically, to a method, device, equipment, medium, and product for querying vehicle energy consumption. Background Technology

[0002] With the popularization of new energy vehicles, users' demand for vehicle energy consumption query, power consumption analysis and abnormal alarm is increasing. However, the current in-vehicle voice interaction system is mostly focused on basic control functions such as multimedia, navigation and air conditioning.

[0003] Currently, for energy consumption queries, users can initiate various vehicle-related questions via voice commands. The system performs semantic parsing on the voice commands to identify the user's query intent and generates corresponding text or voice responses based on a preset knowledge base.

[0004] However, the above methods have low query accuracy, which affects user experience. Summary of the Invention

[0005] The main purpose of this application is to provide a method, device, equipment, medium and product for querying vehicle energy consumption, which can improve the accuracy of the query and thus improve the user experience.

[0006] To achieve the above objectives, firstly, this application provides a method for querying vehicle energy consumption, including: The user's voice wake-up command is parsed to obtain the user's intent type and slot information; If the user's intent type is energy consumption query, retrieve the query data type; Based on the query data type, select the target data source that matches the query data type from multiple preset data sources, and obtain the vehicle's multidimensional energy consumption data from the target data source; The system integrates and encapsulates multidimensional energy consumption data of vehicles and outputs energy consumption query results in natural language.

[0007] In one embodiment, the user's voice wake-up command is parsed to obtain the user's intent type and slot information, including: Semantic recognition is performed on the voice wake-up command to obtain the semantic parsing result; The semantic parsing results are classified into intent types to obtain user intent types and slot information. Among them, user intent types include energy consumption query, energy consumption analysis, or energy consumption execution. Slot extraction is performed on the semantic parsing results to obtain slot information, which includes vehicle identifier, query time range, and real-time data identifier.

[0008] In one embodiment, slot extraction is performed on the semantic parsing results to obtain slot information, including: When the slot information includes the query time range, the semantic parsing results are detected to obtain the time-limited word detection results; When the time-limited keyword detection result indicates the existence of a time range, the time range will be used as the query time range. When the time-limited keyword detection result indicates that no time interval exists, a value is assigned to the time interval, and the assigned time interval is used as the query time interval.

[0009] In one embodiment, based on the query data type, a target data source matching the query data type is selected from multiple preset data sources, and multidimensional energy consumption data of the vehicle is obtained from the target data source, including: Determine the data type of the query and obtain the determination result; When the judgment result is real-time data, the real-time signal from the vehicle end is streamed and statistically analyzed according to the vehicle identifier in the slot information to obtain the real-time energy consumption index. When the determination result is historical data, the energy consumption data of the corresponding time interval in the cloud database is queried according to the query time interval in the slot information to obtain historical energy consumption statistics. When the judgment result is abnormal analysis data, the analysis result of the offline batch processing task is called according to the vehicle identifier in the slot information to obtain the energy consumption abnormality cause analysis result.

[0010] In one embodiment, real-time energy consumption indicators are obtained by streaming and statistically analyzing the vehicle-side signals based on the vehicle identifier in the slot information, including: Based on the vehicle identifier in the slot information, obtain the vehicle-side signal retrieval address; A preset streaming computing engine is used to retrieve the real-time signals from the vehicle-side signal retrieval address to obtain the real-time signal sequence from the vehicle-side. Energy consumption statistics are performed on the real-time signal sequence of the vehicle to obtain real-time energy consumption indicators including current battery level, instantaneous power consumption, and driving range.

[0011] In one embodiment, energy consumption data for the corresponding time interval in the cloud database is queried based on the query time interval in the slot information to obtain historical energy consumption statistics, including: Identify the query time interval and obtain the time interval length information; When no preset time interval is specified for the query time interval, the energy consumption data within the target time interval is queried to obtain historical energy consumption statistics. When the query time interval is specified as a preset time interval, the energy consumption data of the preset time interval in the cloud database is queried to obtain historical energy consumption statistics.

[0012] In one embodiment, the analysis results of the offline batch processing task are invoked based on the vehicle identifier in the slot information to obtain the energy consumption anomaly cause analysis results, including: Based on the vehicle identifier in the slot information, obtain the location information from the offline analysis results; The analysis results of the offline batch processing task corresponding to the location information of the offline analysis results are called to obtain the analysis results of the cause of energy consumption anomalies.

[0013] In one embodiment, the analysis results of the offline batch processing task corresponding to the offline analysis result location information are invoked to obtain the energy consumption anomaly cause analysis results, including: The vehicle tire pressure signal is evaluated to obtain the tire pressure evaluation result; The driving mode signal is evaluated to obtain the driving mode evaluation result; The air conditioner setting signal is evaluated to obtain the air conditioner setting evaluation result; A quantitative analysis of the rate of change of accelerator pedal and the rate of change of braking torque during vehicle operation was conducted to obtain the evaluation results of abnormal driving behavior. The results of tire pressure evaluation, driving mode evaluation, air conditioning setting evaluation, and abnormal driving behavior evaluation are integrated to obtain the analysis results of abnormal energy consumption.

[0014] In one embodiment, a quantitative analysis of the rate of change of the accelerator pedal and the rate of change of the braking torque during vehicle operation is performed to obtain an abnormal driving behavior evaluation result, including: The accelerator pedal opening, braking torque, vehicle speed and acceleration are collected during vehicle operation to obtain time-series driving data; The accelerator pedal change rate in the time-series driving data is processed to obtain the normalized accelerator pedal change rate. Kalman filtering is used to filter the braking torque signal, and the rate of change of the filtered braking torque signal is calculated to obtain the rate of change of the filtered braking torque. The aggressive driving threshold is dynamically adjusted based on vehicle speed and acceleration to obtain adaptive throttle threshold and adaptive braking threshold. The number of frames with normalized accelerator pedal change rate exceeding the adaptive throttle threshold and the number of frames with filtered brake torque change rate exceeding the adaptive braking threshold are statistically analyzed, and confidence weighting is introduced to calculate the proportion of driving aggression, thus obtaining a quantitative index of driving aggression. The energy consumption impact of driving behavior is evaluated based on quantitative indicators of driving aggression, resulting in an evaluation of abnormal driving behavior.

[0015] In one embodiment, multi-dimensional energy consumption data of the vehicle is fused and encapsulated, and energy consumption query results are output in natural language, including: Multi-source data fusion is performed on real-time energy consumption indicators, historical energy consumption statistics and energy consumption anomaly analysis results in multi-dimensional energy consumption data to obtain a structured energy consumption dataset; Natural language generation processing is performed on the structured energy consumption dataset to obtain natural language query result text; The natural language query results text is processed by speech synthesis to output the energy consumption query results in the form of a voice broadcast.

[0016] In one embodiment, multi-source data fusion is performed on real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly cause analysis results from multi-dimensional energy consumption data to obtain a structured energy consumption dataset, including: Timestamp alignment and dimension unification are performed on real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly cause analysis results to obtain aligned multi-source data; The aligned multi-source data is aggregated and ranked by importance to obtain a structured energy consumption dataset that includes the current energy consumption status, historical energy consumption trends, and priority of anomaly causes.

[0017] In one embodiment, the method further includes: If the user intent type is energy consumption execution, then the vehicle control command is generated based on the energy consumption anomaly cause analysis results and sent to the vehicle actuator to perform the corresponding energy-saving operation.

[0018] Secondly, embodiments of this application provide a vehicle energy consumption query device, comprising: The intent parsing module is used to parse the user's voice wake-up command to obtain the user's intent type and slot information; The data type acquisition module is used to obtain the query data type if the user's intent type is energy consumption query. The energy consumption data acquisition module is used to select a target data source that matches the query data type from multiple preset data sources, and to obtain multi-dimensional energy consumption data of the vehicle from the target data source. The energy consumption result output module is used to fuse and encapsulate multi-dimensional energy consumption data of vehicles and output energy consumption query results in natural language.

[0019] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0020] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the methods described above.

[0021] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the methods described above.

[0022] This application provides a vehicle energy consumption query method, device, equipment, medium, and product, including: parsing the user's voice wake-up command to obtain the user's intent type and slot information; if the user's intent type is energy consumption query, obtaining the query data type; according to the query data type, selecting a target data source that matches the query data type from multiple preset data sources, and obtaining multi-dimensional energy consumption data of the vehicle from the target data source; fusing and encapsulating the multi-dimensional energy consumption data of the vehicle, and outputting the energy consumption query result in natural language. This application achieves accurate understanding and targeted data addressing of energy consumption query needs by parsing the user's voice wake-up command, accurately identifying the user's intent type (energy consumption query) and extracting key slot information (such as vehicle identification, time interval, etc.). This avoids the problem of general voice systems not understanding energy consumption terminology. Furthermore, it obtains corresponding multi-dimensional energy consumption data according to the query data type (real-time, historical, or anomaly analysis) and scenarios, which can simultaneously meet the user's multi-level information needs for instantaneous energy consumption, historical statistics, and anomaly attribution. This allows for the fusion and structured encapsulation of multi-source energy consumption data, and outputs the query results in natural language. Users can obtain clear and intuitive energy consumption information and anomaly diagnosis conclusions directly through voice without manually operating the vehicle's menu, thereby significantly improving the convenience, accuracy, and driving safety of energy consumption queries and enhancing the intelligent interactive experience of new energy vehicles. Attached Figure Description

[0023] The accompanying drawings, which form part of this application, are used to provide a further understanding of the application and to make other features, objects, and advantages of the application more apparent. The illustrative embodiments and descriptions of this application are used to explain the application and do not constitute an undue limitation of the application. In the drawings: Figure 1 This is a flowchart illustrating a vehicle energy consumption query method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a vehicle energy consumption query device provided in an embodiment of this application; Figure 3 This is a schematic diagram of the computer device provided in the embodiments of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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 some embodiments of this application, and not all 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.

[0025] The terms first, second, third, fourth, etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein.

[0026] In the embodiments of this application, the term "module" or "unit" refers to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0027] It should be understood that in the various embodiments of this application, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0028] It should be understood that in this application, inclusion and having, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product or device.

[0029] It should be understood that in this application, "multiple" refers to two or more. "And / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, "and / or B" can represent: A existing alone, A and B existing simultaneously, and B existing alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship. "Contains A, B, and C," "Contains A, B, and C" means that all three A, B, and C are contained; "Contains A, B, or C" means that one of A, B, and C is contained; "Contains A, B, and / or C" means that any one, two, or three of A, B, and C are contained.

[0030] It should be understood that in this application, B corresponding to A, B corresponding to A, A and B being mutually related, or B and A being mutually related, indicates that B is associated with A, and B can be determined based on A. Determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information. Matching A and B is defined as a similarity between A and B that is greater than or equal to a preset threshold.

[0031] Depending on the context, as used here, "if" can be interpreted as "when" or "when" or "in response to determination" or "in response to detection".

[0032] The data involved in this application may be data authorized by the tester or fully authorized by all parties. The collection, dissemination, and use of the data shall comply with the relevant laws, regulations and standards of the relevant countries and regions. The implementation methods / executives of this application may be combined with each other.

[0033] It is understood that in the specific embodiments of this application, data such as user information are involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0034] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0035] The present application will now be described in conjunction with the accompanying drawings and specific embodiments.

[0036] With the popularization of new energy vehicles, users' demand for vehicle energy consumption query, power consumption analysis and abnormal alarm is increasing. However, the current in-vehicle voice interaction system is mostly concentrated on basic control functions such as multimedia, navigation and air conditioning. The intelligent interaction capability for the professional scenario of vehicle energy consumption is still relatively weak. The existing voice solution has not been adapted for the energy consumption scenario, making it difficult to support safe and convenient energy consumption voice query and fault cause analysis, thus restricting the intelligent experience of the whole vehicle.

[0037] The most similar existing implementation to this invention is the question-and-answer interaction technology and vehicle interface control technology of in-vehicle voice assistants. Specifically, it is based on the intelligent cockpit voice systems of mainstream car manufacturers and related published patents (such as the visible-to-speak control method adapted to the vehicle system and the voice interaction system for controlling motor vehicles). This solution allows users to initiate various vehicle-related questions through voice commands. The system performs semantic parsing on the voice commands to identify the user's query intent and generates corresponding text or voice reply results based on a preset knowledge base. Some solutions can also realize voice control of basic vehicle functions such as air conditioning and navigation.

[0038] Existing technologies cannot directly query real-time vehicle energy consumption, electricity usage statistics, and energy consumption trends via voice. In particular, they are unable to quickly diagnose the underlying causes of abnormal energy consumption. Users need to manually navigate through multiple levels of in-vehicle menus, which is cumbersome, inefficient, and can easily distract attention while driving, posing a safety hazard. At the same time, voice recognition has poor understanding of energy consumption-related terms (including abnormal operating conditions and energy consumption fluctuation indicators), and its accuracy and response speed are insufficient in the complex in-vehicle environment. Furthermore, it lacks a dedicated low-power processing mechanism, making it impossible to achieve safe, efficient, and natural voice interaction and abnormal diagnosis of energy consumption, which seriously affects the user experience and driving safety.

[0039] To address the aforementioned issues, this application proposes a method for querying vehicle energy consumption.

[0040] Please see Figure 1 , Figure 1 This is a flowchart illustrating a vehicle energy consumption query method provided in an embodiment of this application. Figure 1 As shown, it includes the following steps: Step S101: Perform intent parsing on the user's voice wake-up command to obtain the user intent type and slot information.

[0041] In one embodiment, the user's voice wake-up command is parsed to obtain the user intent type and slot information, including: performing semantic recognition on the voice wake-up command to obtain semantic parsing results; classifying the semantic parsing results to obtain the user intent type and slot information, wherein the user intent type includes energy consumption query, energy consumption analysis, or energy consumption execution; and extracting slots from the semantic parsing results to obtain slot information, wherein the slot information includes vehicle identifier, query time interval, and real-time data identifier.

[0042] In this application, the user intent type is constructed based on a three-level mutually exclusive classification tree constructed according to the progressive logic of cognition, diagnosis, and action. Specifically, the first level is the query class, which is used to handle the user's need to obtain objective energy consumption facts, including subclasses such as real-time energy consumption query, historical energy consumption query, and energy consumption trend query. The system only retrieves and presents data without changing the vehicle status. When the user query is not satisfied with the factual level and further explores the deeper reasons for abnormal energy consumption or fluctuations, it is classified into the second level analysis class, including subclasses such as energy consumption anomaly attribution, driving behavior evaluation, and energy consumption comparison analysis. The system calls offline or online diagnostic logic to output multi-factor attribution conclusions. If the user puts forward an execution request based on the aforementioned analysis conclusions, it is classified into the third level control class, including subclasses such as energy-saving mode switching, air conditioning parameter adjustment, and tire pressure warning linkage. The system automatically generates and executes vehicle control commands according to the analysis results, realizing a closed loop from diagnosis to execution. The above three first-level intents are mutually exclusive. Any user command is uniquely classified into one of the three classes after semantic parsing. Together, they constitute a complete service link of obtaining information, diagnosing causes, and taking action.

[0043] The process involves extracting slots from the semantic parsing results to obtain slot information. This includes: when the slot information includes a query time interval, detecting the semantic parsing results to obtain time-limited word detection results; when the time-limited word detection results indicate the existence of a time interval, using that time interval as the query time interval; and when the time-limited word detection results indicate the absence of a time interval, assigning a value to the time interval and using the assigned time interval as the query time interval.

[0044] Specifically, in one embodiment, when a user wakes up the vehicle via voice and says, "Please check the car's recent battery consumption," the voice command is first subjected to front-end acoustic feature extraction and speech recognition, converting it into semantic parsing results in text form. Then, a pre-deployed semantic understanding model in the cloud or locally is invoked to classify the intent of the text. This model, specifically trained for energy consumption scenarios, can distinguish whether the user's core request is an energy consumption query or a more in-depth energy consumption analysis. In this example, "checking battery consumption" is categorized as an energy consumption query intent. Simultaneously, a slot extraction operation is performed on the semantic parsing results to extract key parameters: for example, the vehicle identifier (if not specifically specified, it defaults to the currently bound vehicle), the recent corresponding query time range (converted to a date range of the last 7 days using a natural language time parsing engine), and the battery consumption, which implicitly requests real-time or statistical data, is assigned a real-time data flag. If the user explicitly states the current real-time battery consumption, the real-time data flag is set to true; if only "recent" is stated, the flag is set to false, indicating a query for historical statistical data. The slot information also includes vehicle identifiers, which are used to locate specific target vehicles in multi-vehicle scenarios.

[0045] In another specific embodiment, the extraction of the query time interval slot involves detecting time-limiting words in the semantic parsing results. For example, if a user says they want to check energy consumption from last Monday to last Friday, the system detects two specific time points, last Monday and last Friday, calculates the complete time interval, and fills it into the query time interval slot. If the user only says they want to check energy consumption without providing any time-limiting words, the slot is automatically assigned the default value of the past 7 days. This default assignment strategy ensures that the query can continue to execute even if the instruction information is incomplete.

[0046] Step S102: If the user intent type is energy consumption query, obtain the query data type.

[0047] Specifically, after completing intent parsing and obtaining the user intent type, it is determined whether the intent is an energy consumption query. If it is determined to be an energy consumption query, the data type required for this query is further determined. Data types include real-time data, historical data, and anomaly analysis data. Specifically, the real-time data identifier and query time range in the slot information are checked: if the real-time data identifier is true, or the query time range includes the current time (e.g., the user asks "now"), the data type is determined to be real-time data; if the query time range is entirely before the current time and the range length is greater than zero, the data type is determined to be historical data; if the user intent type is energy consumption analysis or the user actively asks why energy consumption is high during the query process, the data type is determined to be anomaly analysis data. In this embodiment, since the user only inquires about recent electricity consumption, the real-time data identifier is false and the time range is the last 7 days (excluding the current time), therefore the data type is determined to be historical data.

[0048] Step S103: Based on the query data type, select the target data source that matches the query data type from multiple preset data sources, and obtain the vehicle's multidimensional energy consumption data from the target data source.

[0049] In one embodiment, based on the query data type, a target data source matching the query data type is selected from multiple preset data sources, and multidimensional energy consumption data of the vehicle is obtained from the target data source. This includes: determining the query data type to obtain a determination result; when the determination result is real-time data, performing streaming retrieval and statistics on the vehicle-side real-time signal based on the vehicle identifier in the slot information to obtain real-time energy consumption indicators; when the determination result is historical data, querying the energy consumption data of the corresponding time interval in the cloud database based on the query time interval in the slot information to obtain historical energy consumption statistics; and when the determination result is anomaly analysis data, calling the analysis results of the offline batch processing task based on the vehicle identifier in the slot information to obtain the energy consumption anomaly cause analysis results.

[0050] In one embodiment, real-time energy consumption metrics are obtained by streaming and statistically analyzing the vehicle-side real-time signals based on the vehicle identifier in the slot information. This includes: obtaining the vehicle-side signal retrieval address based on the vehicle identifier in the slot information; using a preset streaming computing engine to retrieve the vehicle-side real-time signals corresponding to the retrieval address to obtain a vehicle-side real-time signal sequence; and performing energy consumption statistics on the vehicle-side real-time signal sequence to obtain real-time energy consumption metrics including current battery level, instantaneous power consumption, and driving range. The preset streaming computing engine includes, but is not limited to, the Flink streaming computing engine.

[0051] Specifically, assuming the user command is for the vehicle's current real-time power consumption, the data type determination result is real-time data. Based on the vehicle identifier in the slot information (e.g., the last six digits of the VIN or vehicle device ID), a vehicle-side signal retrieval address is constructed. Subsequently, the Flink streaming engine is activated, sending a real-time signal request to the vehicle's electronic control unit via the in-vehicle communication gateway to retrieve the raw signal sequences of the current battery voltage, current, remaining power, and instantaneous power. The Flink engine performs sliding window statistical calculations on these signals, filtering noise and aggregating them to obtain the current battery percentage, instantaneous power consumption (unit: kWh / 100 km), and estimated driving range, ultimately generating real-time energy consumption metrics.

[0052] In one embodiment, the energy consumption data for the corresponding time interval in the cloud database is queried according to the query time interval in the slot information to obtain historical energy consumption statistics. This includes: identifying the query time interval to obtain time interval length information; when the query time interval does not specify a preset time interval, the energy consumption data within the target time interval is queried to obtain historical energy consumption statistics; when the query time interval is specified as a preset time interval, the energy consumption data for the preset time interval in the cloud database is queried to obtain historical energy consumption statistics.

[0053] Specifically, based on the recent power consumption example in the above embodiments, the data type is historical data. The query time range (last 7 days) and vehicle identifier are extracted from the slot information. Then, a cloud-based time-series database is connected to construct a structured query statement, such as a daily aggregated energy consumption statistics query. After executing the query, historical energy consumption statistics such as daily mileage, total power consumption, and average power consumption for the vehicle over the past 7 days are obtained. If the user specifies a specific date range (e.g., April 1st to April 5th), the data for that range is directly queried; if the user does not specify a date range, the summary statistics for the last 7 days are returned by default.

[0054] In one embodiment, the analysis results of the offline batch processing task are invoked based on the vehicle identifier in the slot information to obtain the energy consumption anomaly cause analysis results, including: obtaining the offline analysis result location information based on the vehicle identifier in the slot information; and invoking the analysis results of the offline batch processing task corresponding to the offline analysis result location information to obtain the energy consumption anomaly cause analysis results.

[0055] Specifically, the analysis results of the offline batch processing tasks corresponding to the location information of the offline analysis results are called to obtain the energy consumption anomaly cause analysis results, including: evaluating the vehicle tire pressure signal to obtain tire pressure evaluation results; evaluating the driving mode signal to obtain driving mode evaluation results; evaluating the air conditioning setting signal to obtain air conditioning setting evaluation results; performing quantitative analysis of the accelerator pedal change rate and braking torque change rate during vehicle driving to obtain driving behavior anomaly evaluation results; and fusing the tire pressure evaluation results, driving mode evaluation results, air conditioning setting evaluation results, and driving behavior anomaly evaluation results to obtain the energy consumption anomaly cause analysis results.

[0056] The analysis of aggressive driving behavior, including the quantitative analysis of accelerator pedal change rate and braking torque change rate during vehicle operation, yields anomaly evaluation results. This includes: collecting time-series driving data on accelerator pedal opening, braking torque, vehicle speed, and acceleration; processing the accelerator pedal change rate in the time-series data to obtain a normalized accelerator pedal change rate; filtering the braking torque signal using Kalman filtering and calculating the change rate of the filtered braking torque signal to obtain a filtered braking torque change rate; dynamically adjusting the aggressive driving threshold based on vehicle speed and acceleration to obtain adaptive accelerator and adaptive braking thresholds; statistically analyzing the number of frames with normalized accelerator pedal change rates exceeding the adaptive accelerator threshold and the number of frames with filtered braking torque change rates exceeding the adaptive braking threshold, and introducing confidence-weighted calculations to determine the proportion of aggressive driving, resulting in a quantitative index of aggressive driving; and evaluating the energy consumption impact of driving behavior based on this quantitative index to obtain anomaly evaluation results.

[0057] Specifically, assuming a user asks why energy consumption has been so high recently, the user intent is determined to be energy consumption analysis or an unusual question phrase like "why" has been detected. Therefore, the data type is determined to be anomaly analysis data. Based on the vehicle identifier, pre-calculated energy consumption anomaly analysis results are retrieved from the result storage location of the offline batch processing task (such as HDFS or a cloud data warehouse). This offline task is typically executed hourly or daily, analyzing data based on the user's historical driving habits, vehicle status, and environmental dimensions. The retrieved results are generated strictly according to the following logic: Tire pressure assessment: Obtain the tire pressure signal values ​​of the four tires. If the tire pressure of any tire is greater than 300 kPa, the tire pressure assessment result is output as high (code 1); if the tire pressure of any tire is less than 200 kPa, the output is low (code 2); if the tire pressure of all four tires is between 200 kPa and 300 kPa (inclusive), the output is normal (code 0).

[0058] Driving mode evaluation: The percentage of time during which the driving mode signal DrvModReq is 1 (representing ECO mode) is counted in the trip segment. If the ECO mode percentage is greater than 50%, the driving mode evaluation result is output as "ECO percentage is high" (code 1); otherwise, the output is "ECO percentage is low" (code 0).

[0059] Air Conditioning Setting Evaluation: Check if any one of the following three conditions is met: ① The set temperature on the large screen (left side HmiCmptmtTSpForRowFirstLe or right side HmiCmptmtTSpForRowFirstRi) is less than 20℃ or greater than 24℃; ② The estimated air conditioning mass flow rate HvacAirMFlowEstimd is greater than or equal to 350; ③ The external circulation ratio (duration of RecircBtn=0 divided by the total trip duration) is greater than 50%. If any of the above conditions are met, the air conditioning setting evaluation for this trip will output "Air Conditioning Setting Too Extreme" (code 1); otherwise, it will output "Normal" (code 0).

[0060] Anomaly evaluation of driving behavior (quantitative analysis of aggressive driving behavior): The driving behavior during vehicle operation is calculated frame by frame. The specific steps are as follows: a) Collect the following data for each frame during the vehicle's journey in real time to form a time sequence (frame index i=1,2,…,N, where N is the total number of frames in the trip): A i : Accelerator pedal opening in frame i (0~100% or normalized value, unit: % or dimensionless) B i Braking torque in frame i (unit: Nm) v i Vehicle speed in frame i (unit: km / h or m / s) a i Vehicle acceleration in frame i (unit: m / s²) t i : Timestamp corresponding to the i-th frame (unit: seconds) b) Calculate the rate of change of the accelerator pedal in each frame. :

[0061] Among them, w iAssign sliding window weights and ensure >0; A i A i 1: The change in throttle opening between the current frame and the previous frame (absolute value or with sign, usually the magnitude of the change). t i t i 1: The time interval between the current frame and the previous frame (sampling period, unit: s).

[0062] c) Normalize the rate of change of the accelerator pedal to obtain:

[0063] in, The maximum value of the difference in accelerator pedal opening between two adjacent frames during the entire stroke (reflecting the maximum abrupt change in accelerator pedal depressing during that stroke) is used as the normalization benchmark.

[0064] d) Use Kalman filtering to process the braking torque signal Denoising was performed, and the rate of change of braking torque in each frame was calculated. :

[0065] in, For Kalman filter gain, >0.

[0066] B i B i 1: Change in braking torque between the current frame and the previous frame (absolute value, only the magnitude of change is relevant). t i t i 1: The time interval between the current frame and the previous frame (sampling period, unit: s).

[0067] e) Dynamically adjust the aggressive driving threshold based on vehicle speed and acceleration:

[0068]

[0069] Where, θ A Threshold for throttle aggression, ranging from [0.5, 0.8]. The higher the speed, the higher the threshold; that is, a larger pedal change rate is required at high speeds to be considered aggressive. θ B : Braking intensity threshold (dimensionless), with a value range of [0.5, 0.7]. The larger the absolute value of acceleration, the higher the threshold, that is, a larger braking torque change rate is required to be regarded as intense during sharp acceleration and deceleration; v i : Vehicle speed of the current frame (m / s or km / h, units shall be kept consistent); v max : The maximum vehicle speed in this trip (or the preset maximum vehicle speed, such as 180 km / h), used for normalizing the vehicle speed; a i : Acceleration of the current frame (m / s², which can be positive or negative, and the absolute value is taken here); a max : The maximum absolute value of acceleration in this trip (or the preset maximum acceleration, such as 5 m / s²), used for normalizing acceleration.

[0070] f) Count the number of frames that satisfy > and > , and introduce a multi-dimensional confidence level .

[0071] g) Calculate the aggressive driving proportion R according to the total number of frames N of the trip: R

[0072] where, : Indicator function, which takes a value of 1 when the i-th frame satisfies > , otherwise takes a value of 0; : Indicator function, which takes a value of 1 when the i-th frame satisfies > , otherwise takes a value of 0; For each frame, if the throttle operation is aggressive, it contributes ; if the braking operation is aggressive, it contributes 1- ; if both conditions are satisfied at the same time, it contributes +(1 )=1, that is, this frame is fully counted; if neither condition is satisfied, the contribution is 0. Finally, the sum is divided by the total number of frames to obtain the proportion R.

[0073] h) Quantitatively evaluate the aggressive driving behavior of the vehicle based on the aggressive driving proportion R, and output the evaluation result of abnormal driving behavior (for example: R≤0.2 is normal, 0.2<R≤0.5 is moderately aggressive, R>0.5 is highly aggressive). ​​

[0074] The results of the four assessments—tire pressure evaluation, driving mode evaluation, air conditioning setting evaluation, and abnormal driving behavior evaluation—are combined to obtain the analysis results of abnormal energy consumption. For example, if the tire pressure evaluation is low (code 2), the driving mode evaluation is low ECO ratio (code 0), the air conditioning setting evaluation is overly aggressive (code 1), and the driving behavior evaluation is moderately aggressive, then the main reasons for high energy consumption are determined to be overly aggressive air conditioning settings and aggressive driving behavior, and corresponding optimization suggestions are given.

[0075] Step S104: Integrate and encapsulate the multidimensional energy consumption data of the vehicle, and output the energy consumption query results in natural language.

[0076] In one embodiment, the multidimensional energy consumption data of the vehicle is fused and encapsulated, and the energy consumption query results are output in natural language. This includes: fusing real-time energy consumption indicators, historical energy consumption statistics and energy consumption anomaly analysis results from the multidimensional energy consumption data to obtain a structured energy consumption dataset; performing natural language generation processing on the structured energy consumption dataset to obtain natural language query result text; and performing speech synthesis on the natural language query result text to output the energy consumption query results in the form of voice broadcast.

[0077] This involves fusing real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly cause analysis results from multi-dimensional energy consumption data to obtain a structured energy consumption dataset. This includes: aligning real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly cause analysis results with timestamps and unifying dimensions to obtain aligned multi-source data; and aggregating and ranking the aligned multi-source data to obtain a structured energy consumption dataset containing current energy consumption status, historical energy consumption trends, and anomaly cause priority.

[0078] Specifically, after acquiring multidimensional energy consumption data (which may be one or more of real-time energy consumption indicators, historical energy consumption statistics, and anomaly analysis results), the process enters the fusion and output stage.

[0079] First, multi-source data is fused. For example, in a typical interaction, a user repeatedly asks about their recent week's electricity consumption and why it's high, simultaneously obtaining historical statistical data and anomaly analysis results. First, this data is aligned by timestamps, and historical electricity consumption trend data is correlated with driving behavior, tire pressure, and other data from the corresponding time periods in the anomaly analysis results. Then, the fused data is prioritized: first, the difference between the current electricity consumption level and the historical average is highlighted; second, possible causes of anomalies are listed and ranked from highest to lowest confidence; finally, energy-saving suggestions are provided. The prioritized data is packaged into a structured dataset containing the current energy consumption status, historical energy consumption trend curves, and a priority list of anomaly causes.

[0080] Next, the structured dataset is fed into the natural language generation module. This module selects an appropriate response template based on the user's intent type and output scenario. For example, for historical data queries, the template could be "Average power consumption over the past X days is Y kWh / 100km, with the highest power consumption on [date], reaching Z." For anomaly analysis, the template could be "High energy consumption may be caused by the following reasons: 1. Low tire pressure (left front tire 190kPa), 2. Aggressive driving accounts for 35%." It is recommended to fill the corresponding positions in the template with the values ​​from the structured dataset to generate fluent natural language text.

[0081] Finally, the natural language text is sent to the speech synthesis engine to generate the query results in voice broadcast format, which is then played to the user through the in-vehicle speakers. Simultaneously, the vehicle's infotainment screen displays text information and charts, with the cause of the anomaly highlighted or indicated by a warning icon. This completes a full voice query interaction for vehicle energy consumption.

[0082] In one embodiment, the method further includes: if the user intent type is energy consumption execution, then generating a vehicle control command based on the energy consumption anomaly cause analysis result, and sending it to the vehicle actuator to perform the corresponding energy-saving operation.

[0083] Specifically, if the user intent type parsed in step S101 is energy consumption analysis rather than energy consumption query, the query data type determination in step S102 is skipped, and professional energy consumption analysis is executed directly. At this time, an offline batch processing task is triggered or the latest analysis results are called. Multiple recently completed trip segments are analyzed according to the aforementioned tire pressure, driving mode, air conditioning settings, and driving behavior quantification logic, and specific values ​​for each abnormality are output (e.g., actual tire pressure, ECO percentage, air conditioning set temperature and external circulation ratio, and the R value for aggressive driving). After structured fusion, the analysis results are fed back to the user in natural language, along with actionable optimization suggestions (e.g., suggesting inflating the left front tire to 250 kPa, or trying to set the air conditioning temperature to 22°C and turn on external circulation). If the user agrees, navigation or vehicle settings functions can be further linked to achieve a closed loop from diagnosis to execution.

[0084] This application provides a vehicle energy consumption query method, including: parsing the user's voice wake-up command to obtain the user's intent type and slot information; if the user's intent type is energy consumption query, obtaining the query data type; according to the query data type, selecting a target data source that matches the query data type from multiple preset data sources, and obtaining multi-dimensional energy consumption data of the vehicle from the target data source; fusing and encapsulating the multi-dimensional energy consumption data of the vehicle, and outputting the energy consumption query result in natural language. This application achieves accurate understanding and targeted data addressing of energy consumption query needs by parsing the user's voice wake-up command, accurately identifying the user's intent type (energy consumption query) and extracting key slot information (such as vehicle identification, time interval, etc.). This avoids the problem of general voice systems not understanding energy consumption terminology. Furthermore, it obtains corresponding multi-dimensional energy consumption data according to the query data type (real-time, historical, or anomaly analysis) and scenarios, which can simultaneously meet the user's multi-level information needs for instantaneous energy consumption, historical statistics, and anomaly attribution. This allows for the fusion and structured encapsulation of multi-source energy consumption data, and outputs the query results in natural language. Users can obtain clear and intuitive energy consumption information and anomaly diagnosis conclusions directly through voice without manually operating the vehicle's menu, thereby significantly improving the convenience, accuracy, and driving safety of energy consumption queries and enhancing the intelligent interactive experience of new energy vehicles.

[0085] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0086] The following are device embodiments of this application. For details not described in detail, please refer to the corresponding method embodiments described above.

[0087] Figure 2 This diagram illustrates the structure of a vehicle energy consumption query device according to an embodiment of this application. For ease of explanation, only the parts relevant to this embodiment are shown. The vehicle energy consumption query device includes: The intent parsing module 201 is used to parse the user's voice wake-up command to obtain the user's intent type and slot information; The data type acquisition module 202 is used to acquire the query data type if the user intent type is energy consumption query. The energy consumption data acquisition module 203 is used to select a target data source that matches the query data type from multiple preset data sources according to the query data type, and to acquire multi-dimensional energy consumption data of the vehicle from the target data source. The energy consumption result output module 204 is used to fuse and encapsulate the multi-dimensional energy consumption data of the vehicle and output the energy consumption query results in natural language.

[0088] In one embodiment, the intent parsing module 201 is further configured to perform semantic recognition on the voice wake-up command to obtain a semantic parsing result; The semantic parsing results are classified into intent types to obtain user intent types and slot information. Among them, user intent types include energy consumption query, energy consumption analysis, or energy consumption execution. Slot extraction is performed on the semantic parsing results to obtain slot information, which includes vehicle identifier, query time range, and real-time data identifier.

[0089] In one embodiment, the intent parsing module 201 is further configured to detect the semantic parsing results when the slot information includes a query time interval, and obtain the time-limited word detection results; When the time-limited keyword detection result indicates the existence of a time range, the time range will be used as the query time range. When the time-limited keyword detection result indicates that no time interval exists, a value is assigned to the time interval, and the assigned time interval is used as the query time interval.

[0090] In one embodiment, the energy consumption data acquisition module 203 is further configured to determine the type of query data and obtain a determination result; When the judgment result is real-time data, the real-time signal from the vehicle end is streamed and statistically analyzed according to the vehicle identifier in the slot information to obtain the real-time energy consumption index. When the determination result is historical data, the energy consumption data of the corresponding time interval in the cloud database is queried according to the query time interval in the slot information to obtain historical energy consumption statistics. When the judgment result is abnormal analysis data, the analysis result of the offline batch processing task is called according to the vehicle identifier in the slot information to obtain the energy consumption abnormality cause analysis result.

[0091] In one embodiment, the energy consumption data acquisition module 203 is further configured to acquire the vehicle-end signal pull address based on the vehicle identifier in the slot information; A preset streaming computing engine is used to retrieve the real-time signals from the vehicle-side signal retrieval address to obtain the real-time signal sequence from the vehicle-side. Energy consumption statistics are performed on the real-time signal sequence of the vehicle to obtain real-time energy consumption indicators including current battery level, instantaneous power consumption, and driving range.

[0092] In one embodiment, the energy consumption data acquisition module 203 is further used to identify the query time interval and obtain the time interval length information; When no preset time interval is specified for the query time interval, the energy consumption data within the target time interval is queried to obtain historical energy consumption statistics. When the query time interval is specified as a preset time interval, the energy consumption data of the preset time interval in the cloud database is queried to obtain historical energy consumption statistics.

[0093] In one embodiment, the energy consumption data acquisition module 203 is further configured to acquire offline analysis result location information based on the vehicle identifier in the slot information; The analysis results of the offline batch processing task corresponding to the location information of the offline analysis results are called to obtain the analysis results of the cause of energy consumption anomalies.

[0094] In one embodiment, the energy consumption data acquisition module 203 is also used to evaluate the vehicle tire pressure signal and obtain a tire pressure evaluation result; The driving mode signal is evaluated to obtain the driving mode evaluation result; The air conditioner setting signal is evaluated to obtain the air conditioner setting evaluation result; A quantitative analysis of the rate of change of accelerator pedal and the rate of change of braking torque during vehicle operation was conducted to obtain the evaluation results of abnormal driving behavior. The results of tire pressure evaluation, driving mode evaluation, air conditioning setting evaluation, and abnormal driving behavior evaluation are integrated to obtain the analysis results of abnormal energy consumption.

[0095] In one embodiment, the energy consumption data acquisition module 203 is also used to collect the accelerator pedal opening, braking torque, vehicle speed and acceleration during vehicle driving to obtain time-series driving data. The accelerator pedal change rate in the time-series driving data is processed to obtain the normalized accelerator pedal change rate. Kalman filtering is used to filter the braking torque signal, and the rate of change of the filtered braking torque signal is calculated to obtain the rate of change of the filtered braking torque. The aggressive driving threshold is dynamically adjusted based on vehicle speed and acceleration to obtain adaptive throttle threshold and adaptive braking threshold. The number of frames with normalized accelerator pedal change rate exceeding the adaptive throttle threshold and the number of frames with filtered brake torque change rate exceeding the adaptive braking threshold are statistically analyzed, and confidence weighting is introduced to calculate the proportion of driving aggression, thus obtaining a quantitative index of driving aggression. The energy consumption impact of driving behavior is evaluated based on quantitative indicators of driving aggression, resulting in an evaluation of abnormal driving behavior.

[0096] In one embodiment, the energy consumption result output module 204 is also used to perform multi-source data fusion on the real-time energy consumption index, historical energy consumption statistics and energy consumption anomaly cause analysis results in the multi-dimensional energy consumption data to obtain a structured energy consumption dataset. Natural language generation processing is performed on the structured energy consumption dataset to obtain natural language query result text; The natural language query results text is processed by speech synthesis to output the energy consumption query results in the form of a voice broadcast.

[0097] In one embodiment, the energy consumption result output module 204 is also used to perform timestamp alignment and dimension unification on real-time energy consumption indicators, historical energy consumption statistics and energy consumption anomaly cause analysis results to obtain aligned multi-source data; The aligned multi-source data is aggregated and ranked by importance to obtain a structured energy consumption dataset that includes the current energy consumption status, historical energy consumption trends, and priority of anomaly causes.

[0098] In one embodiment, the device further includes an energy consumption analysis module, which is used to generate a vehicle control command based on the energy consumption anomaly cause analysis result if the user intent type is energy consumption execution, and send it to the vehicle actuator to perform the corresponding energy-saving operation.

[0099] This application provides a vehicle energy consumption query device, specifically used for: parsing the user's voice wake-up command to obtain the user's intent type and slot information; if the user's intent type is energy consumption query, obtaining the query data type; according to the query data type, selecting a target data source that matches the query data type from multiple preset data sources, and obtaining multi-dimensional energy consumption data of the vehicle from the target data source; fusing and encapsulating the multi-dimensional energy consumption data of the vehicle, and outputting the energy consumption query result in natural language. This application achieves accurate understanding and targeted data addressing of energy consumption query needs by parsing the user's voice wake-up command, accurately identifying the user's intent type (energy consumption query) and extracting key slot information (such as vehicle identification, time interval, etc.). This avoids the problem of general voice systems not understanding energy consumption terminology. Furthermore, it obtains corresponding multi-dimensional energy consumption data according to the query data type (real-time, historical, or anomaly analysis) and scenarios, which can simultaneously meet the user's multi-level information needs for instantaneous energy consumption, historical statistics, and anomaly attribution. This allows for the fusion and structured encapsulation of multi-source energy consumption data, and outputs the query results in natural language. Users can obtain clear and intuitive energy consumption information and anomaly diagnosis conclusions directly through voice without manually operating the vehicle's menu, thereby significantly improving the convenience, accuracy, and driving safety of energy consumption queries and enhancing the intelligent interactive experience of new energy vehicles.

[0100] This application Figure 3 A schematic diagram of a computer device is provided. (Example) Figure 3As shown, the computer device 3 in this embodiment includes a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, it implements the steps described in the various vehicle energy consumption query method embodiments above, for example... Figure 1 Steps 101 to 104 are shown. Alternatively, when processor 301 executes computer program 303, it implements the functions of each module / unit in the above-described vehicle energy consumption query device embodiments, for example... Figure 2 The functions of modules / units 201 to 204 shown.

[0101] This application also provides a readable storage medium storing a computer program, which, when executed by a processor, is used to implement the vehicle energy consumption query method provided in the various embodiments described above.

[0102] The readable storage medium can be a computer storage medium or a communication medium. A communication medium includes any medium that facilitates the transfer of computer programs from one location to another. A computer storage medium can be any available medium accessible to a general-purpose or special-purpose computer. For example, a readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application-Specific Integrated Circuit (ASIC). Alternatively, the ASIC can be located in a user equipment. Of course, the processor and the readable storage medium can also exist as discrete components in a communication device. The readable storage medium can be a read-only memory (ROM), random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0103] This application also provides a program product including executable instructions stored in a readable storage medium. At least one processor of the device can read the executable instructions from the readable storage medium, and the at least one processor executes the executable instructions to cause the device to implement the vehicle energy consumption query method provided in the various embodiments described above.

[0104] In the embodiments of the above-described device, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0105] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for querying vehicle energy consumption, characterized in that, include: The user's voice wake-up command is parsed to obtain the user's intent type and slot information; If the user intent type is energy consumption query, obtain the query data type; Based on the query data type, a target data source matching the query data type is selected from multiple preset data sources, and multidimensional energy consumption data of the vehicle is obtained from the target data source. The multidimensional energy consumption data of the vehicle is fused and encapsulated, and the energy consumption query results are output in natural language.

2. The vehicle energy consumption query method as described in claim 1, characterized in that, The process of parsing the user's voice wake-up command to obtain the user's intent type and slot information includes: The voice wake-up command is semantically recognized to obtain semantic parsing results; The semantic parsing results are classified into intent types to obtain user intent types and slot information, wherein the user intent types include energy consumption query, energy consumption analysis, or energy consumption execution. The semantic parsing results are subjected to slot extraction to obtain slot information, wherein the slot information includes vehicle identifier, query time interval and real-time data identifier.

3. The vehicle energy consumption query method as described in claim 2, characterized in that, The process of extracting slots from the semantic parsing results to obtain slot information includes: If the slot information includes the query time interval, the semantic parsing result is detected to obtain the time-limited word detection result; When the detection result of the time-limited keyword indicates the existence of a time interval, the time interval is used as the query time interval; When the time-limited keyword detection result indicates that the time interval does not exist, a value is assigned to the time interval, and the assigned time interval is used as the query time interval.

4. The vehicle energy consumption query method as described in claim 1, characterized in that, The step of selecting a target data source matching the query data type from multiple preset data sources, and obtaining multidimensional energy consumption data of the vehicle from the target data source, includes: The data type of the query is determined, and a determination result is obtained; When the determination result is real-time data, the real-time signal of the vehicle end is streamed and statistically analyzed according to the vehicle identifier in the slot information to obtain the real-time energy consumption index. When the determination result is historical data, the energy consumption data of the corresponding time interval in the cloud database is queried according to the query time interval in the slot information to obtain historical energy consumption statistics. When the determination result is abnormal analysis data, the analysis result of the offline batch processing task is called according to the vehicle identifier in the slot information to obtain the energy consumption abnormality cause analysis result.

5. The vehicle energy consumption query method as described in claim 4, characterized in that, The step of streaming and statistically analyzing real-time signals from the vehicle end based on the vehicle identifier in the slot information to obtain real-time energy consumption indicators includes: Based on the vehicle identifier in the slot information, obtain the vehicle-side signal retrieval address; A preset streaming computing engine is used to retrieve the real-time vehicle signal corresponding to the vehicle signal retrieval address to obtain the real-time vehicle signal sequence. Energy consumption statistics are performed on the real-time signal sequence of the vehicle to obtain real-time energy consumption indicators including current battery level, instantaneous power consumption, and driving range.

6. The vehicle energy consumption query method as described in claim 4, characterized in that, The step of querying energy consumption data for the corresponding time interval in the cloud database based on the query time interval in the slot information to obtain historical energy consumption statistics includes: The query time interval is identified to obtain the time interval length information; When no preset time interval is specified for the query time interval, the energy consumption data within the target time interval is queried to obtain the historical energy consumption statistics. When the query time interval is specified as a preset time interval, the energy consumption data of the preset time interval in the cloud database is queried to obtain the historical energy consumption statistics.

7. The vehicle energy consumption query method as described in claim 4, characterized in that, The step of calling the analysis results of the offline batch processing task based on the vehicle identifier in the slot information to obtain the energy consumption anomaly cause analysis results includes: Based on the vehicle identifier in the slot information, obtain the offline analysis result location information; The analysis results of the offline batch processing task corresponding to the location information of the offline analysis results are called to obtain the analysis results of the cause of the energy consumption anomaly.

8. The vehicle energy consumption query method as described in claim 7, characterized in that, The step of calling the analysis results of the offline batch processing task corresponding to the offline analysis result location information to obtain the energy consumption anomaly cause analysis results includes: The vehicle tire pressure signal is evaluated to obtain the tire pressure evaluation result; The driving mode signal is evaluated to obtain the driving mode evaluation result; The air conditioner setting signal is evaluated to obtain the air conditioner setting evaluation result; A quantitative analysis of the rate of change of accelerator pedal and the rate of change of braking torque during vehicle operation was conducted to obtain the evaluation results of abnormal driving behavior. The tire pressure evaluation results, driving mode evaluation results, air conditioning setting evaluation results, and abnormal driving behavior evaluation results are integrated to obtain the energy consumption abnormality cause analysis results.

9. The vehicle energy consumption query method as described in claim 8, characterized in that, The quantitative analysis of the rate of change of accelerator pedal and the rate of change of braking torque during vehicle operation yields abnormal driving behavior evaluation results, including: The accelerator pedal opening, braking torque, vehicle speed and acceleration are collected during vehicle operation to obtain time-series driving data; The accelerator pedal change rate in the time-series driving data is processed to obtain the normalized accelerator pedal change rate. Kalman filtering is used to filter the braking torque signal, and the rate of change of the filtered braking torque signal is calculated to obtain the rate of change of the filtered braking torque. The aggressive driving threshold is dynamically adjusted based on vehicle speed and acceleration to obtain adaptive throttle threshold and adaptive braking threshold. The number of frames with normalized accelerator pedal change rate exceeding the adaptive throttle threshold and the number of frames with filtered brake torque change rate exceeding the adaptive braking threshold are statistically analyzed, and a confidence-weighted calculation of the proportion of driving aggression is introduced to obtain a quantitative index of driving aggression. The energy consumption impact of driving behavior is evaluated based on the quantitative index of driving aggression, resulting in an evaluation of abnormal driving behavior.

10. The vehicle energy consumption query method as described in claim 1, characterized in that, The process of fusing and encapsulating the multidimensional energy consumption data of the vehicle and outputting the energy consumption query results in natural language includes: The real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly analysis results in the multidimensional energy consumption data are fused from multiple sources to obtain a structured energy consumption dataset. Natural language generation processing is performed on the structured energy consumption dataset to obtain natural language query result text; The natural language query result text is processed by speech synthesis and output to obtain the energy consumption query result in the form of a voice broadcast.

11. The vehicle energy consumption query method as described in claim 10, characterized in that, The process involves fusing real-time energy consumption indicators, historical energy consumption statistics, and energy consumption anomaly analysis results from the multi-dimensional energy consumption data to obtain a structured energy consumption dataset, including: The real-time energy consumption indicators, the historical energy consumption statistics, and the energy consumption anomaly cause analysis results are time-stamp aligned and dimension unified to obtain aligned multi-source data. The aligned multi-source data is aggregated and sorted by importance to obtain a structured energy consumption dataset that includes the current energy consumption status, historical energy consumption trends, and priority of abnormal causes.

12. The vehicle energy consumption query method as described in claim 8, characterized in that, The method further includes: If the user intent type is energy consumption execution, then a vehicle control command is generated based on the energy consumption anomaly cause analysis result and sent to the vehicle actuator to perform the corresponding energy-saving operation.

13. A vehicle energy consumption query device, characterized in that, include: The intent parsing module is used to parse the user's voice wake-up command to obtain the user's intent type and slot information; The data type acquisition module is used to acquire the query data type if the user intent type is energy consumption query. The energy consumption data acquisition module is used to select a target data source that matches the query data type from multiple preset data sources, and to acquire multidimensional energy consumption data of the vehicle from the target data source; The energy consumption result output module is used to fuse and encapsulate the multi-dimensional energy consumption data of the vehicle and output the energy consumption query results in natural language.

14. A computer device, characterized in that, Includes a memory, and one or more processors communicatively connected to the memory; The memory stores instructions that can be executed by the one or more processors to enable the one or more processors to implement the vehicle energy consumption query method as described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, Includes a program or instructions that, when run on a computer, implement the vehicle energy consumption query method according to any one of claims 1-12.

16. A computer program product, characterized in that, The system includes a computer program that, when executed by a processor, implements the vehicle energy consumption query method according to any one of claims 1-12.

Images