New energy vehicle driving behavior safety assessment method based on battery energy consumption data

Abstract

The invention discloses a new energy vehicle driving behavior safety assessment method based on battery energy consumption data, and the method comprises the steps: S1, setting a data collection module, and obtaining original data from various sensors and a bus of a vehicle; s2, setting a data processing and feature extraction module; the data processing module is used for processing original, heterogeneous and high-dimensional data acquired by the data acquisition module, and extracting low-dimensional and high-value characteristics capable of directly representing driving operation essence from the original, heterogeneous and high-dimensional data; s3, setting a driving behavior identification and quantification module; s2, receiving a series of high-value features from S2, converting continuous data streams into discrete and explainable driving operation events through a designed discrimination rule set and a quantitative model, and evaluating the intensity of the driving operation events; s4, setting a safety evaluation and early warning module; s5, setting an early warning and output module; physical signals of a battery system are directly collected and deeply fused with time-space information of a vehicle, so that more scientific and more accurate safety risk assessment and early warning of driving behaviors are realized.

Description

Technical Field

[0001] This invention relates to the field of active safety technology for new energy vehicles, specifically to a method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data. Background Technology

[0002] Currently, GPS-based driving behavior analysis technology has been widely used in the field of vehicle safety monitoring. These systems typically rely on GPS positioning information to obtain vehicle kinematic parameters such as speed and acceleration through time-difference calculations, and then use these parameters to identify and evaluate driving behaviors such as rapid acceleration, rapid deceleration, and sharp turns.

[0003] However, such GPS-based analysis methods have several inherent limitations. First, regarding accuracy and real-time performance, due to the inherent limitations in the update frequency and positioning errors of GPS signals, the acceleration obtained through position differential analysis exhibits significant delays and biases, making it difficult to accurately capture instantaneous, dramatic driving actions, such as short-term high-power demands or emergency braking. Second, these systems can only reflect the external motion state of the vehicle and cannot perceive the internal responses of the vehicle's powertrain caused by driving behavior. Especially for new energy vehicles, rapid acceleration and deceleration directly cause significant current, voltage, and power surges in the battery system. These changes in electrical parameters are direct physical signals reflecting the intensity of driving behavior, but current technologies have failed to effectively collect and utilize these signals for analysis, resulting in a fundamental deficiency in the evaluation dimensions.

[0004] Furthermore, existing methods generally lack the ability to comprehensively assess vehicle operating conditions. For example, when encountering complex road conditions such as long uphill or downhill sections, changes in vehicle speed and acceleration may be a necessary response to road conditions, rather than solely caused by the driver's subjective actions. Because existing technologies fail to integrate information reflecting the relationship between vehicle load and road conditions, such as battery energy consumption data and motor output characteristics, they struggle to distinguish between "dangerous driving behavior" and "reasonable operation under complex road conditions," easily leading to misjudgments and affecting the accuracy and practicality of the assessment results.

[0005] Therefore, there is an urgent need in this field for a driving behavior safety assessment scheme that is more direct, more accurate, and more capable of identifying operating conditions, in order to overcome the aforementioned deficiencies in the existing technology. Summary of the Invention

[0006] The present invention proposes a method and storage medium for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data. Specifically, it is a method and system for safety assessment and early warning of driving behavior by integrating real-time energy consumption data of vehicle batteries and GPS trajectory information, which can at least solve one of the technical problems in the background art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data includes the following steps: S1: Data Acquisition Module. This module is the foundation of the entire system, responsible for acquiring raw data from various vehicle sensors and buses with high precision, high frequency, and high reliability.

[0008] S2: Data Processing and Feature Extraction Module. This module is responsible for cleaning, aligning, and fusing the raw, heterogeneous, and high-dimensional data collected by the data acquisition module, and extracting low-dimensional, high-value features that can directly characterize the essence of driving operations.

[0009] S3: Driving Behavior Recognition and Quantification Module. This module undertakes the core reasoning and judgment functions. It receives a series of high-value features from the S2 module, and through the designed set of discrimination rules and quantification models, transforms the continuous data stream into discrete, interpretable driving operation events, and evaluates their strength.

[0010] S4: Safety Assessment and Early Warning Module. This module receives a structured list of driving operation events from the S3 module, establishes a multi-dimensional, interpretable, and fair assessment system, transforms discrete event sequences into comprehensive safety scores and real-time risk levels, and ultimately triggers corresponding early warning and reporting mechanisms.

[0011] S5: Early Warning and Output Module. This module receives comprehensive scores, risk levels, and structured event lists from the S4 module, and constructs a multi-level, multi-channel, adaptive output and interaction system.

[0012] Further module S1 includes the following units: S101: Battery Data Acquisition Unit. The core of this unit is to directly monitor and parse the vehicle's network bus to obtain first-hand data from the Battery Management System (BMS).

[0013] S102: Positioning Data Acquisition Unit. This unit is responsible for acquiring the vehicle's spatiotemporal information, providing rich scene context for battery data.

[0014] S103: Data synchronization mechanism unit. Resolves the problem of different data sources having different frequencies and latency.

[0015] Further unit S101 includes the following: S1011: Data source acquisition channel. Monitor the Controller Area Network (CAN) bus of the new energy vehicle power network to obtain key battery pack status parameters broadcast by the Battery Management System (BMS) at fixed intervals.

[0016] S1012: Determine the physical interface. Connect to the vehicle's OBD-II diagnostic interface via the in-vehicle intelligent terminal (T-Box) or dedicated data acquisition unit (DAU), or connect directly in parallel to the CAN bus.

[0017] S1013: Data Acquisition Parameters. The hardware terminal acquires the required data by listening to and parsing specific message identifiers. This includes: total current ( ), total voltage ( Instantaneous power ( ).in: 1. Total current ( ): The total current flowing into / out of the battery pack; positive values ​​indicate driving, negative values ​​indicate recycling. It is the most direct signal reflecting driving / braking demand.

[0018] 2. Total voltage ( ): The total voltage of the battery pack. Used to calculate power and reflect the battery's state of load.

[0019] 3. Instantaneous power ( ):Depend on Calculations show that it is a core indicator for measuring instantaneous changes in energy.

[0020] S1014: Data sampling frequency setting. To ensure that instantaneous changes in current can be captured, the sampling frequency is required. .

[0021] Further unit S102 includes the following: S1021: Data source acquisition channel. From Global Navigation Satellite Systems (GNSS), including GPS, BeiDou (BDS), etc.

[0022] S1022: Hardware Configuration. The vehicle-mounted terminal integrates or connects to an external multi-mode GNSS receiver chip, with the following performance requirements: 1. High update rate: Supports at least 5-10Hz data output to match high-frequency CAN data.

[0023] 2. High precision: Supports SBAS (Satellite-based Augmentation System), providing sub-meter level positioning accuracy.

[0024] S1023: Data acquisition parameters. Includes: latitude and longitude coordinates. ,speed UTC timestamp, altitude .in: 1. Latitude and longitude coordinates WGS84 coordinate system.

[0025] 2. Speed : Calculated from Doppler frequency shift.

[0026] 3. UTC timestamp: Provides a high-precision PPS (pulses per second) signal as a reference for absolute time synchronization of the entire system.

[0027] 4. Altitude The required accuracy error is at the centimeter level.

[0028] Further unit S103 includes the following: S1031: Unified Time Standard. The PPS pulse and UTC time output by the GNSS receiver are used as the absolute time standard for the entire system. When the vehicle-mounted terminal receives the PPS pulse, it associates it with the chip timer counter of the vehicle-mounted equipment to add a unified timestamp with microsecond-level precision to all subsequent data. .

[0029] S1032: Data Alignment Model. CAN data has a high frequency, while GPS data has a low frequency. For each GPS data point... Search the cache for the CAN data packet with the closest timestamp. The following requirements must be met:

[0030] in It is a very small tolerance threshold.

[0031] Further module S2 includes the following units: S201: Data Preprocessing and Synchronization Alignment Unit. This unit serves as a prerequisite for all subsequent advanced analyses, ensuring data consistency in both time and space.

[0032] S202: Battery Time-Domain Feature Extraction Unit. This unit directly extracts dynamic time-domain features that reflect the driver's intentions and the intensity of their operation from the battery's current, voltage, and power signals.

[0033] S203: Trajectory Feature Extraction Unit. This unit extracts features reflecting the road environment in which the vehicle is located from GPS trajectory data, providing crucial scene context for driving behavior.

[0034] Further unit S201 includes the following: S2011: Abnormal Data Cleaning. A hybrid algorithm combining amplitude limiting filtering and median filtering is used to remove outliers caused by sensor instantaneous errors, network packet loss, electromagnetic interference, etc.

[0035] S2012: Precise synchronization of multi-source data. An interpolation-based external synchronization model is used to obtain data at this GPS time point. The best matching CAN data.

[0036] Further unit S202 includes the following: S2021: Calculate the rate of change of the index. A discrete approximation calculation is performed using the backward difference method. Let the sampling time interval be... .in: 1. Rate of change of current. The rate of change of current quantifies how fast the current changes.

[0037]

[0038] 2. Power change rate. The power change rate value indicates the severity of driving / braking behavior.

[0039]

[0040] S2022: Statistical Characteristic Analysis. A sliding time window of length N is defined, and the variance within the window is calculated to measure the intensity of signal fluctuations. A larger variance indicates more drastic current changes during that time period, resulting in less smooth driving operation.

[0041] ,

[0042] Calculate the extreme values ​​within the window, and capture the maximum driving and braking currents within the window to determine the operating intensity.

[0043]

[0044] Further unit S203 includes the following: S2031: Calculate the rate of change of elevation to estimate the slope. To avoid single-point calculation errors, this invention proposes a slope calculation model using linear fitting, as follows: Take time point front and back Latitude, longitude, and elevation data for each step Record the starting point ( ) and the last point ( The latitude and longitude coordinates are: ,in and It is a latitude value. and It is the longitude value.

[0045] Calculate horizontal movement distance

[0046] in,

[0047]

[0048]

[0049] In the formula: It is a two-parameter arctangent function.

[0050] Calculate the elevation change as follows The slope of the fitted line This is the estimated average elevation change rate (slope) for that route segment:

[0051] Further module S3 includes the following units: S301: Rapid acceleration recognition unit. Used to identify unnecessary and violent acceleration behaviors initiated by the driver.

[0052] S302: Rapid deceleration detection unit. Used to identify sudden, unnecessary braking actions.

[0053] S303: Long-duration aggressive driving recognition unit. Used to identify prolonged periods of aggressive driving, which can cause the battery to operate under high load continuously, affecting safety and lifespan.

[0054] S304: Output Unit. This unit outputs a structured list of driving operation events, providing the most direct input for module S4 to perform comprehensive scoring and risk assessment. Each event... Include: 1. Event Types ( Rapid acceleration ( ), rapid deceleration ( ), prolonged and intense driving.

[0055] 2. Event start timestamp: .

[0056] 3. Event duration: .

[0057] 4. Event Intensity : Intensity value of rapid acceleration or deceleration.

[0058] 5. Related environmental data: average slope, speed, etc. at the time of the event.

[0059] Further unit S301 includes the following: S3011: Judgment conditions (must be met simultaneously): Power change rate threshold:

[0060] in, It is an empirical threshold. It directly reflects the explosive growth in power demand and is the most essential electrical characteristic of rapid acceleration.

[0061] Absolute power threshold:

[0062] in, The value is a certain percentage of the vehicle's rated power P.

[0063] 3. Eliminate unnecessary operations:

[0064] in, This is the uphill gradient threshold. If the vehicle is currently on a steep uphill slope (gradient ≥ 1), it will be affected. If the first two conditions are met, then a high-power request is necessary. Even if the first two conditions are met, it should not be judged as a "bad" rapid acceleration, thereby reducing the false alarm rate.

[0065] S3012: Calculate the rapid acceleration intensity value:

[0066] Further unit S302 includes the following: S3021: Judgment conditions (must be met simultaneously): Current change rate threshold:

[0067] in, It is a negative threshold. The sharp negative change in current indicates that the drive motor is transforming into a generator, entering a strong energy recovery mode.

[0068] Regenerative braking judgment:

[0069] in, It is a negative threshold. This confirms that the current is clearly negative and the system is in a regenerative braking state, thus eliminating other possible causes for minor current fluctuations.

[0070] Exclude unnecessary operations:

[0071] in, This refers to the downhill threshold. If the vehicle is on a steep downhill slope, then strong energy recovery is a necessary safety measure to control the vehicle speed and must be eliminated.

[0072] S3022: Calculate the rapid deceleration intensity value:

[0073] The intensity of rapid deceleration is directly proportional to the rate of change of current and the magnitude of the braking current. The larger this value, the more rapid the braking.

[0074] Further unit S303 includes the following: S3031: Establish a discriminant model and statistically analyze the proportion of high power within a time window.

[0075] Define a long-lasting sliding window ; Statistics within this window, instantaneous power Exceeding a certain moderate threshold Time percentage .

[0076]

[0077] S3032: Decision criteria:

[0078] The value is 0.7. If the vehicle operates at high power for more than 70% of the time, then that period of time is considered aggressive driving.

[0079] Further module S4 includes the following units: S401: Driving Behavior-Battery Instantaneous Stress Relationship Calculation Unit. Used to assess the safety risks of rapid acceleration and deceleration events. Instantaneous stress. The calculation formula is as follows:

[0080] in, Indicates a driving operation event The instantaneous stress exerted on the battery system is the event intensity. The function.

[0081] S402: Driving Behavior-Battery Cumulative Stress Relationship Calculation Unit. Used to assess the safety risks of prolonged aggressive driving. Its cumulative stress... Defined as a function of average power and duration. Its calculation formula is as follows:

[0082] in, The calculation formula is:

[0083] In the formula: :time average power Long-lasting sliding window Number of sampling points within :time instantaneous power A sliding window (e.g., 60 seconds) for recognizing prolonged aggressive driving. : Represents the thermal sensitivity coefficient, which is a constant.

[0084] S403: Comprehensive Score Calculation Unit. This unit calculates an intuitive safety score for a complete trip or a specific time period.

[0085] S404: Early Warning Strategy and Output Unit. Based on the above evaluation results, this unit designs a multi-level early warning triggering strategy, recovery strategy, and report generation strategy based on real-time risk assessment.

[0086] Further unit S401 includes the following: S4011: For rapid acceleration events, the stress can be estimated as follows:

[0087] S4012: For a rapid deceleration event, the stress can be estimated as follows:

[0088] and This represents the sensitivity coefficient for rapid acceleration and deceleration, and is a constant.

[0089] Further unit S403 includes the following: S4031: Single trip scoring is calculated using a base score deduction system. An initial score is set. =100.

[0090]

[0091]

[0092] In the formula: Weighting coefficients for different event types. These weights are set according to the security importance of different event types and can be dynamically adjusted based on the scenario context output by S3.

[0093] S4032: Real-time risk level assessment calculation. Define the sliding time window. Calculate the real-time risk value within this window. , is used to represent the weighted intensity of all operation events that occur within the cumulative sliding time window, and its output is a continuously changing value.

[0094]

[0095] in: Low risk: <

[0096] Medium risk: ≤ <

[0097] High risk: ≥

[0098] In the formula: : Represents the weighting coefficient at time j.

[0099] : Represents the event intensity at time j This is the low-risk threshold.

[0100] This is a high-risk threshold.

[0101] Further unit S404 includes the following: S4041: Warning Triggering Strategy 1. Hierarchical early warning based on real-time risk values The system continuously monitors and calculates real-time risk values. Different levels of alerts are triggered based on preset thresholds: Level 1 Warning (High-Risk Warning): When the real-time risk value... ≥ This level of warning is triggered when the driver is in a sustained high-risk driving situation. The system will issue a strong audio-visual warning through the in-vehicle human-machine interface, including a continuously flashing red warning icon and a short, sharp warning sound, to remind the driver to adjust their driving behavior immediately.

[0102] Level 2 Warning (Medium Risk Warning): When ≤ < This level of warning is triggered at specific times. It targets moderate-risk driving conditions. The system provides warnings through a flashing orange icon and gentle voice prompts (such as "Please drive smoothly"), alerting the driver while avoiding excessive distraction.

[0103] 2. Special early warning based on the intensity of a single event In addition to risk-value-based early warnings, the system also has a special early warning mechanism for extreme single events: Special event warning: When the intensity of a single driving operation... Exceeding the absolute safety threshold At any given time, regardless of the current real-time risk level, the system will immediately trigger the highest priority warning. This type of warning is for extremely dangerous driving maneuvers; the system will record detailed data before and after the event and issue the strongest level of alert.

[0104] S4042: Early warning execution and recovery strategy.

[0105] After the warning is triggered, the system will continuously monitor the driving status, and when the real-time risk value falls back to a safe range ( After a period of sustained stability, the warning status will be automatically lifted, and the system will return to normal monitoring status.

[0106] S4043: Cloud-based reporting and driver profile generation strategy.

[0107] Feature Vector Construction: Long-term features are extracted for each driver to form a feature vector. The K-Means algorithm is used to classify drivers into profiles, automatically categorizing them into "aggressive," "normal," and "moderate" types, providing accurate data support for refined fleet management and insurance UBI pricing.

[0108]

[0109] Further module S5 includes the following units: S501: Real-time vehicle warning unit. This unit is responsible for providing the driver with immediate and clear feedback when a risk occurs, aiming to correct unsafe driving behaviors.

[0110] S502: Cloud Reporting and Data Analysis Unit. This unit runs on the server side, aggregating and mining massive amounts of historical data to generate in-depth analysis reports for different roles.

[0111] S503: Data Interaction Interface Unit. This unit is responsible for exposing the system's output capabilities to external third-party systems.

[0112] Further unit S501 includes the following: S5011: Constructing an early warning strategy engine. This is a rule-based state machine whose decision-making logic is based on the real-time risk level output by module S4. and event intensity Its core workflow is described as follows: 1. Initial state: Monitoring status After the system starts, the early warning strategy engine enters monitoring mode. In this mode, the engine continuously receives real-time data streams from the security assessment module, including real-time risk values. Intensity of a single driving operation event Contextual information.

[0113] 2. State transition and early warning triggering mechanism Level 1 warning trigger path When real-time risk value Reaching or exceeding the high-risk threshold Upon this, the system immediately switches from monitoring mode to Level 1 warning mode. This transition indicates that the vehicle is in a sustained high-risk driving mode, requiring immediate and strong intervention.

[0114] Special warning trigger path When the event intensity of a single driving operation Reaching or exceeding the absolute safety threshold At this time, regardless of the current real-time risk value, the system will immediately switch from monitoring mode to special warning mode. This path is specifically designed for sudden, extremely dangerous operations, ensuring that no major safety risks are overlooked.

[0115] Level 2 warning trigger path When real-time risk value Reaching or exceeding the low-risk threshold (But not yet reached) When a certain level of risk accumulates, the system transitions from monitoring to a Level 2 early warning state. This state provides appropriate early warning intervention for moderate levels of risk accumulation.

[0116] 3. Early warning execution and status recovery Once any warning state is entered, the system will immediately execute the corresponding warning action: Level 1 Warning Status: Triggers a strong audible and visual warning, including a full-screen red warning and a rapid warning sound; Special alert status: Executes the highest priority alert and automatically records event-related data; Level 2 Warning Status: Activates a flashing icon and gentle voice prompts to avoid excessive interference with driving. All warning statuses have self-recovery capabilities. When a risk indicator is detected to remain below the corresponding threshold for a sustained period and remains stable, the system will automatically return from the warning status to the monitoring status, completing a full warning cycle.

[0117] 4. State Priority and Conflict Resolution Mechanism The early warning strategy engine uses a priority coverage mechanism to handle situations where multiple conditions are triggered simultaneously: Special warnings (based on the intensity of a single event) have the highest priority; The next level is a Level 1 warning (based on persistently high risk); Level II warning has the lowest priority.

[0118] When multiple triggering conditions are met simultaneously, the system only executes the highest priority warning action to ensure the clarity and effectiveness of the warning response.

[0119] S5012: Human-Computer Interaction (HMI) Execution Strategy.

[0120] 1. Visual channel Level 1 Warning: The in-vehicle central control screen or the dedicated area of ​​the instrument panel continuously flashes red, displaying the text or icon of "Rapid Acceleration / Rapid Deceleration Warning", and may also display the current safety score.

[0121] Level 2 Warning: The icon flashes orange without interrupting the current navigation or entertainment interface.

[0122] Special warning: Screen recording prompt.

[0123] 2. Auditory channel Level 1 warning: Issues a short, sharp, non-verbal warning sound.

[0124] Level 2 warning: Issues a gentle voice reminder.

[0125] Further unit S502 includes the following: S5021: Driving behavior analysis report generation.

[0126] Following step S4043, perform in-depth analysis of long-term historical data on the cloud platform to generate a comprehensive report. The comprehensive report includes: 1. Trip Summary: Trip distance, average speed, total energy consumption, safety score ( ).

[0127] 2. Event statistics: Bar charts or trend graphs showing the number of times, frequency (times / 100 km), and average intensity of risky operations such as rapid acceleration and deceleration.

[0128] 3. Scenario Analysis: Distribution of risky operations across different road types, time periods, and weather conditions.

[0129] 4. Battery Health Correlation: Based on long-term data, estimate the percentage potential impact of poor driving habits on battery life.

[0130] 5. Report Generation and Push: Reports are automatically generated periodically and pushed to users via email or mobile app. Fleet administrators can view reports for all their vehicles through a web portal.

[0131] Further unit S503 includes the following: S5031: API (Application Programming Interface) Design.

[0132] 1. Interface type: Provides a RESTful API, using the standard HTTP protocol and JSON data format, making it easy for developers to integrate.

[0133] 2. Security Authentication: OAuth 2.0 or API Key mechanism is used to authenticate the identity and manage permissions of the caller.

[0134] S5032: Data interaction scenario construction.

[0135] 1. Insurance UBI Platform: Regularly retrieves users' driving behavior scores via API, which serve as the basis for adjusting the next year's premium.

[0136] 2. Automaker After-Sales System: When the system detects that a vehicle has been driven aggressively for a long period of time, it sends a reminder via API to the after-sales system to "suggest a deep battery health check", proactively creating service opportunities.

[0137] 3. Fleet Management Platform: Receives high-risk early warning events in real time and displays the safety status of all vehicles in the fleet on a map, facilitating dispatcher intervention.

[0138] 4. Traffic Management Platform: After anonymization and aggregation, it provides regional hotspot maps of rapid acceleration / deceleration to support decision-making for road safety management.

[0139] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0140] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0141] As can be seen from the above technical solution, the new energy vehicle driving behavior safety assessment method and medium based on battery energy consumption data of the present invention directly collects the physical signals of the battery system and deeply integrates them with the vehicle's spatiotemporal information, thereby achieving a more scientific and accurate safety risk assessment and early warning of driving behavior.

[0142] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: 1. Introducing a direct measurement mechanism based on battery electrophysical signals significantly improves the accuracy of behavior recognition. Traditional methods rely on GPS differential calculations to determine vehicle motion parameters, which are subject to delays and biases, making it difficult to accurately capture sudden, intense driving behaviors. This invention, by acquiring high-frequency electrical signals from the battery system, such as current, voltage, and power, directly reflects the impact of driver actions on the powertrain. This enables accurate, real-time identification of critical driving behaviors such as rapid acceleration and deceleration, effectively overcoming the shortcomings of traditional indirect calculation methods in terms of response speed and accuracy.

[0143] 2. By integrating information from multiple sources to construct a scene-aware discrimination model, the false positive rate is significantly reduced. When recognizing driving behavior, the system innovatively incorporates GPS elevation data and a slope estimation model, combined with real-time road conditions, to make necessity judgments and effectively distinguish between "dangerous operations" and "reasonable responses under complex road conditions." This mechanism significantly reduces false alarms caused by road conditions such as inclines and declines, making the assessment results more scenario-based and practical.

[0144] 3. Establish a quantifiable driving behavior-battery stress mapping relationship to provide an objective and unified evaluation benchmark. By constructing a mathematical model of driving behavior and the instantaneous and cumulative stresses of the battery system, the impact of driving operations is transformed into calculable physical indicators, which are further converted into intuitive safety scores and risk levels. This system provides a scientific and consistent basis for the quantitative evaluation of driving behavior, overcoming the shortcomings of traditional methods, such as strong subjectivity and inconsistent standards.

[0145] 4. Achieve a closed-loop security management system covering the entire lifecycle and multiple levels, balancing real-time intervention with long-term value extraction. The system covers the entire process from real-time vehicle warnings to long-term driving reports, featuring multi-level risk triggering and adaptive recovery mechanisms. It can provide effective safety intervention at critical moments and generate driver profiles and analysis reports based on historical data. This capability not only enhances the immediacy of driving safety but also provides a reliable data foundation for derivative applications such as insurance UBI pricing, refined fleet management, and battery health prediction. Attached Figure Description

[0146] Figure 1 This is a diagram illustrating the architecture of a new energy vehicle driving behavior safety assessment system based on battery energy consumption data, as per the present invention. Figure 2 This is a flowchart of the multi-level early warning triggering strategy for risk assessment based on real-time and single-event scenarios, as presented in this invention. Detailed Implementation

[0147] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0148] like Figure 1 As shown in this embodiment, the method for assessing the driving behavior of new energy vehicles based on battery energy consumption data specifically includes the following steps: S1: Data Acquisition Module. This module is the foundation of the entire system, responsible for acquiring raw data from various vehicle sensors and buses with high precision, high frequency, and high reliability. Its design goal is to ensure the integrity, accuracy, and synchronization of information at the data source, laying a solid data foundation for all subsequent advanced analyses.

[0149] S101: Battery Data Acquisition Unit. The core of this unit is to directly monitor and parse the vehicle's network bus to obtain first-hand data from the Battery Management System (BMS).

[0150] S1011: Data source acquisition channel. The power network of new energy vehicles adopts the Controller Area Network (CAN) bus, which follows standard protocols such as SAE J1939 and GB / T 27930. The BMS, as a node on the network, broadcasts key status parameters of the battery pack at fixed intervals.

[0151] S1012: Determine the physical interface. This invention connects to the vehicle's OBD-II diagnostic interface or directly to the CAN bus via an in-vehicle intelligent terminal (T-Box) or a dedicated data acquisition unit (DAU); the hardware terminal, such as the in-vehicle intelligent terminal or the dedicated data acquisition unit, needs to have a dual-channel or multi-channel CAN controller to simultaneously monitor the power CAN and the body CAN.

[0152] S1013: Data Acquisition Parameters. The hardware terminal acquires the required data by listening to and parsing specific message identifiers. This includes: total current ( ), total voltage ( Instantaneous power ( ).in: 1. Total current ( ): The total current flowing into / out of the battery pack; positive values ​​indicate driving, negative values ​​indicate recycling. It is the most direct signal reflecting driving / braking demand.

[0153] 2. Total voltage ( ): The total voltage of the battery pack. Used to calculate power and reflect the battery's state of load.

[0154] 3. Instantaneous power ( ):Depend on Calculations show that it is a core indicator for measuring instantaneous changes in energy.

[0155] S1014: Data sampling frequency setting. To ensure that instantaneous changes in current can be captured, this invention requires a specific sampling frequency. The recommended frequency is 20-50Hz.

[0156] S102: Positioning Data Acquisition Unit. This unit is responsible for acquiring the vehicle's spatiotemporal information, providing rich scene context for battery data.

[0157] S1021: Data source acquisition channel. From Global Navigation Satellite Systems (GNSS), including GPS, BeiDou (BDS), etc.

[0158] S1022: Hardware Configuration. The vehicle-mounted terminal integrates or connects to an external multi-mode GNSS receiver chip, with the following performance requirements: 1. High update rate: Supports at least 5-10Hz data output to match high-frequency CAN data.

[0159] 2. High precision: Supports SBAS (Satellite-based Augmentation System), providing sub-meter level positioning accuracy.

[0160] S1023: Data acquisition parameters. Includes: latitude and longitude coordinates. ,speed UTC timestamp, altitude .

[0161] 1. Latitude and longitude coordinates: WGS84 coordinate system.

[0162] 2. Speed : km / h or m / s, calculated from Doppler frequency shift, which is more accurate and smoother than the speed calculated by position difference.

[0163] 3. UTC timestamp: Provides a high-precision PPS (pulses per second) signal as a reference for absolute time synchronization of the entire system.

[0164] 4. Altitude The required accuracy error is at the centimeter level.

[0165] S103: Data Synchronization Mechanism This solves the problem of different data sources having different frequencies and different latency.

[0166] S1031: Unified Time Standard. The PPS pulse and UTC time output by the GNSS receiver are used as the absolute time standard for the entire system. When the vehicle-mounted terminal receives the PPS pulse, it associates it with the chip timer counter of the vehicle-mounted equipment to add a unified timestamp with microsecond-level precision to all subsequent data. .

[0167] S1032: Data Alignment Model. CAN data has a high frequency, while GPS data has a low frequency. For each GPS data point... Search the cache for the CAN data packet with the closest timestamp. The following requirements must be met:

[0168] in: It is a very small tolerance threshold (10ms in this invention). In this way, each GPS point corresponds to a set of CAN data that is precisely matched in time, forming a high-frequency data segment with spatiotemporal context.

[0169] S2: Data Processing and Feature Extraction Module. This module is responsible for cleaning, aligning, and fusing the raw, heterogeneous, and high-dimensional data collected by the data acquisition module, and extracting low-dimensional, high-value features that directly characterize the essence of driving operations. Its output is a highly engineered feature vector, providing direct input for subsequent recognition and evaluation modules.

[0170] S201: Data Preprocessing and Synchronization Alignment Unit. This unit is a prerequisite for all subsequent advanced analyses, ensuring data consistency in time and space.

[0171] S2011: Abnormal Data Cleaning. A hybrid algorithm combining amplitude limiting filtering and median filtering is used to remove outliers caused by sensor instantaneous errors, network packet loss, electromagnetic interference, etc. This algorithm can effectively smooth random fluctuations while preserving the data's variation characteristics, making it suitable for capturing the starting points of rapid acceleration / deceleration. Taking total current data as an example: Set total current limit parameters Values ​​outside the range will be considered invalid.

[0172]

[0173] in, and These are the theoretical maximum and minimum values ​​of the total current data. Based on the physical limits of the vehicle's power system, the values ​​used in this invention are... =500A, =-500A, to eliminate spike pulse interference. express The source data value output at a given time point.

[0174] Set the median parameter of total current. For a length of The invention sorts and takes the median value of data within a sliding window containing (single-value) data. The value is 5).

[0175]

[0176] S2012: Precise synchronization of multi-source data. An interpolation-based external synchronization model is used to find the synchronization point with this GPS time. The most suitable CAN data solves the time phase asynchrony problem between CAN data (high frequency) and GPS data (low frequency).

[0177] At the point of time Acquire a GPS data point. Estimate. CAN data values ​​at any given time. Total current at... The interpolated estimate at time t is:

[0178] in, and These are two adjacent CAN sampling time points, and .

[0179] S202: Battery Time-Domain Feature Extraction Unit. This unit directly extracts dynamic time-domain features that reflect the driver's intentions and the intensity of their operation from the battery's current, voltage, and power signals.

[0180] S2021: Calculate the rate of change of indicators. The rate of change is a core indicator for measuring "urgency." The rate of change of current directly corresponds to the severity of torque demand, and the rate of change of power directly corresponds to the abrupt change in energy flow. A discrete approximation calculation is performed using the backward difference method. Let the sampling time interval be... .

[0181] 1. Rate of Change of Current. The rate of change of current quantifies how fast the current changes. A very large positive peak indicates a rapid acceleration, while a very large negative trough indicates a rapid deceleration or a strong energy recovery operation.

[0182]

[0183] 2. Power change rate. The power change rate value indicates the severity of driving / braking behavior, avoiding the influence of voltage fluctuations.

[0184]

[0185] S2022: Statistical Feature Analysis. This section analyzes fluctuations in statistical features within a short time window to identify sustained high-intensity driving behaviors (such as aggressive driving or long-distance hill climbing).

[0186] Set a length of A sliding time window is used to calculate the variance within the window, which is used to measure the intensity of signal fluctuations. The larger the variance, the more drastic the current changes within that time period, and the less smooth the driving operation.

[0187] ,

[0188] Calculate the extreme values ​​within the window, and capture the maximum driving and braking currents within the window to determine the operating intensity.

[0189]

[0190] S203: Trajectory Feature Extraction Unit. This unit extracts features reflecting the road environment in which the vehicle is located from GPS trajectory data, providing crucial scene context for driving behavior.

[0191] S2031: Calculate the rate of change of elevation to estimate the slope. The rate of change of elevation is key to determining the "necessity" of driving behavior. Rapid acceleration on flat roads is dangerous, while high-power demands on uphill roads are necessary. To avoid single-point calculation errors, this invention proposes a slope calculation model using linear fitting.

[0192] Take time point front and back Each step size (the recommended value range in this invention is...) Latitude, longitude, and elevation data of each point Record the starting point ( ) and the last point ( The latitude and longitude coordinates are: ,in and It is a latitude value. and It is the longitude value.

[0193] Calculate horizontal movement distance

[0194] in,

[0195]

[0196]

[0197] In the formula: It is a two-parameter arctangent function.

[0198] Calculate the elevation change as follows The slope of the fitted line This is the estimated average elevation change rate (slope) for that route segment:

[0199] S3: Driving Behavior Recognition and Quantification Module. This module is the "brain" of the system, undertaking the core reasoning and judgment functions. It receives a series of high-value features from the S2 module, and through a designed set of discrimination rules and a quantification model, transforms the continuous data stream into discrete, interpretable driving operation events, and assesses their intensity. Its innovation lies in the deep fusion and judgment of battery electrical stress signals and environmental operating condition signals, thereby achieving accurate and scenario-based recognition.

[0200] This module employs a hybrid discrimination model based on multi-dimensional thresholds and fuzzy logic. Its core idea is that a driving operation event must simultaneously exceed thresholds in both the battery electrical signal dimension and the social vehicle motion dimension, and be comprehensively judged in conjunction with the scene context before it can be finally confirmed. This effectively avoids false alarms from single sensors and misjudgments of the scene. The overall decision-making process of this module involves the collaborative judgment of multiple sub-units.

[0201] S301: Rapid acceleration recognition unit. Used to identify unnecessary and violent acceleration behaviors initiated by the driver.

[0202] S3011: Judgment conditions (must be met simultaneously): 1. Power change rate threshold:

[0203] This is the primary trigger condition. It is an empirical threshold (recommended range 15 kW / s ~ 20 kW / s). It directly captures the explosive growth in power demand and is the most essential electrical characteristic of rapid acceleration.

[0204] 2. Absolute power threshold:

[0205] This is an auxiliary condition used to filter out minor fluctuations at low power levels. The value is a certain percentage of the vehicle's rated power P (the recommended value in this invention). =P amount * 0.3).

[0206] 3. Eliminate unnecessary operations:

[0207] This is the key to contextual judgment. This is the uphill gradient threshold (recommended range is 3%-5%). If the vehicle is currently on a steep uphill slope (gradient ≥ 0.5%), If the first two conditions are met, then a high-power request is necessary, and even if the first two conditions are met, it should not be judged as a "bad" rapid acceleration. This condition greatly reduces the false alarm rate.

[0208] S3012: Calculate the rapid acceleration intensity value:

[0209] S302: Rapid deceleration detection unit. Identifies sudden, unnecessary braking actions (energy recovery).

[0210] S3021: Judgment conditions (must be met simultaneously): 1. Current change rate threshold:

[0211] in, It is a negative threshold (recommended range is -30 A / s to -20 A / s). A sharp negative change in current indicates that the drive motor is transforming into a generator, entering a strong energy recovery mode.

[0212] 2. Regenerative braking judgment:

[0213] in, It is a negative threshold (recommended value is -10A). This confirms that the current is clearly negative and the system is in regenerative braking mode, thus eliminating other possible causes for minor current fluctuations.

[0214] 3. Eliminate unnecessary operations:

[0215] in, This is the downhill threshold (recommended range is -5% to -2%). If the vehicle is on a steep downhill slope, then strong energy recovery is a necessary safety measure to control speed and should not result in penalty points.

[0216] S3022: Calculate the rapid deceleration intensity value:

[0217] The intensity of rapid deceleration is directly proportional to the rate of change of current and the magnitude of the braking current. The larger this value, the more rapid the braking.

[0218] S303: Long-duration aggressive driving recognition unit. Used to identify prolonged periods of aggressive driving, which can cause the battery to operate under high load continuously, affecting safety and lifespan.

[0219] S3031: Establish a discriminant model and statistically analyze the proportion of high power within a time window.

[0220] Define a long-lasting sliding window (e.g., 60 seconds); Statistics within this window, instantaneous power Exceeding a certain moderate threshold (This value is taken as 50% of the rated power) time percentage .

[0221]

[0222] S3032: Decision criteria:

[0223] The value is 0.7. If the vehicle operates at high power for more than 70% of the time, then that period of time is considered aggressive driving.

[0224] S304: Output Unit. This unit outputs a structured list of driving operation events, providing the most direct input for module S4 to perform comprehensive scoring and risk assessment. Each event... Include: 1. Event Types ( Rapid acceleration ( ), rapid deceleration ( ), prolonged and intense driving.

[0225] 2. Event start timestamp: .

[0226] 3. Event duration: .

[0227] 4. Event Intensity : Intensity value of rapid acceleration or deceleration.

[0228] 5. Related environmental data: average slope, speed, etc. at the time of the event.

[0229] S4: Safety Assessment and Early Warning Module. This module receives a structured list of driving operation events from module S3. Its core task is to establish a multi-dimensional, interpretable, and fair assessment system, transforming discrete event sequences into comprehensive safety scores and real-time risk levels, and ultimately triggering corresponding early warning and reporting mechanisms. Its design goal is not only to record behavior but also to assess its risk implications.

[0230] S401: Driving Behavior-Battery Instantaneous Stress Relationship Calculation Unit. It serves as a bridge connecting driving operations with their potential physical impacts, providing a theoretical basis for quantitative assessment.

[0231]

[0232] in, Indicates a driving operation event The instantaneous stress exerted on the battery system is the event intensity. The function of. For example, a high-intensity rapid acceleration event will produce higher battery stress (greater current surge and thermal load) than a moderate-intensity rapid deceleration event.

[0233] S4011: For rapid acceleration events, the stress can be estimated as follows:

[0234] S4012: For a rapid deceleration event, the stress can be estimated as follows:

[0235] and This represents the sensitivity coefficient for rapid acceleration and deceleration, and is a constant.

[0236] S402: Driving Behavior-Battery Cumulative Stress Relationship Calculation Unit. Used to assess the safety risks of prolonged aggressive driving. Its cumulative stress... Defined as a function of average power and duration. Its calculation formula is as follows:

[0237] in, The calculation formula is:

[0238] In the formula: :time average power Long-lasting sliding window Number of sampling points within :time instantaneous power A sliding window (e.g., 60 seconds) for recognizing prolonged aggressive driving. : Represents the thermal sensitivity coefficient, which is a constant.

[0239] Table 1: Values ​​of relevant parameters

[0240] S403: Comprehensive Scoring Calculation Unit. This unit is designed to calculate an intuitive safety score for a complete trip or a specific time period.

[0241] S4031: Single trip scoring is calculated using a base score deduction system. An initial score is set. =100.

[0242]

[0243]

[0244] In the formula: Weighting coefficients for different event types. These weights are set according to the security importance of different event types and can be dynamically adjusted based on the scenario context output by S3.

[0245] S4032: Real-time risk level assessment calculation. Define the sliding time window. Calculate the real-time risk value within this window. , is used to represent the weighted intensity of all operation events that occur within the cumulative sliding time window, and its output is a continuously changing value.

[0246]

[0247] in: Low risk: <

[0248] Medium risk: ≤ <

[0249] High risk: ≥

[0250] In the formula: : Represents the weighting coefficient at time j.

[0251] : Represents the event intensity at time j This is the low-risk threshold.

[0252] This is a high-risk threshold.

[0253] S404: Early Warning Strategy and Output Unit. Based on the above evaluation results, this unit designs a multi-level early warning triggering strategy, recovery strategy, and report generation strategy based on real-time risk assessment; such as... Figure 2 As shown; S4041: Warning Triggering Strategy 1. Hierarchical early warning based on real-time risk values The system continuously monitors and calculates real-time risk values. Different levels of alerts are triggered based on preset thresholds: Level 1 Warning (High-Risk Warning): When the real-time risk value... ≥ This level of warning is triggered when the driver is in a sustained high-risk driving situation. The system will issue a strong audio-visual warning through the in-vehicle human-machine interface, including a continuously flashing red warning icon and a short, sharp warning sound, to remind the driver to adjust their driving behavior immediately.

[0254] Level 2 Warning (Medium Risk Warning): When ≤ < This level of warning is triggered at specific times. It targets moderate-risk driving conditions. The system provides warnings through a flashing orange icon and gentle voice prompts (such as "Please drive smoothly"), alerting the driver while avoiding excessive distraction.

[0255] 2. Special early warning based on the intensity of a single event In addition to risk-value-based early warnings, the system also has a special early warning mechanism for extreme single events: Special event warning: When the intensity of a single driving operation... Exceeding the absolute safety threshold At any given time, regardless of the current real-time risk level, the system will immediately trigger the highest priority warning. This type of warning is for extremely dangerous driving maneuvers; the system will record detailed data before and after the event and issue the strongest level of alert.

[0256] S4042: Early warning execution and recovery strategy.

[0257] After the warning is triggered, the system will continuously monitor the driving status, and when the real-time risk value falls back to a safe range ( After a period of sustained stability, the warning status will be automatically lifted, and the system will return to normal monitoring status.

[0258] S4043: Cloud-based reporting and driver profile generation strategy.

[0259] Feature Vector Construction: Long-term features are extracted for each driver to form a feature vector. The K-Means algorithm is used to classify drivers into profiles, automatically categorizing them into "aggressive," "normal," and "moderate" types, providing accurate data support for refined fleet management and insurance UBI pricing.

[0260]

[0261] S5: Early Warning and Output Module. This module receives the comprehensive score, risk level, and structured event list from the S4 module. Its core task is to build a multi-level, multi-channel, adaptive output and interaction system to ensure that the evaluation results can reach the target object in the most effective way and realize the closed-loop value flow of data.

[0262] S501: Real-time vehicle warning unit. This unit is responsible for providing the driver with immediate and clear feedback when a risk occurs, aiming to correct poor driving behavior, and is a form of active safety intervention.

[0263] S5011: Constructing an early warning strategy engine. This is a rule-based state machine whose decision-making logic is based on the real-time risk level output by module S4. and event intensity Its core workflow is described as follows: 1. Initial state: Monitoring status After the system starts, the early warning strategy engine enters monitoring mode. In this mode, the engine continuously receives real-time data streams from the security assessment module, including real-time risk values. Intensity of a single driving operation event Contextual information.

[0264] 2. State transition and early warning triggering mechanism Level 1 warning trigger path When real-time risk value Reaching or exceeding the high-risk threshold Upon this, the system immediately switches from monitoring mode to Level 1 warning mode. This transition indicates that the vehicle is in a sustained high-risk driving mode, requiring immediate and strong intervention.

[0265] Special warning trigger path When the event intensity of a single driving operation Reaching or exceeding the absolute safety threshold At this time, regardless of the current real-time risk value, the system will immediately switch from monitoring mode to special warning mode. This path is specifically designed for sudden, extremely dangerous operations, ensuring that no major safety risks are overlooked.

[0266] Level 2 warning trigger path When real-time risk value Reaching or exceeding the low-risk threshold (But not yet reached) When a certain level of risk accumulates, the system transitions from monitoring to a Level 2 early warning state. This state provides appropriate early warning intervention for moderate levels of risk accumulation.

[0267] 3. Early warning execution and status recovery Once any warning state is entered, the system will immediately execute the corresponding warning action: Level 1 Warning Status: Triggers a strong audible and visual warning, including a full-screen red warning and a rapid warning sound; Special alert status: Executes the highest priority alert and automatically records event-related data; Level 2 Warning Status: Activates a flashing icon and gentle voice prompts to avoid excessive interference with driving. All warning statuses have self-recovery capabilities. When a risk indicator is detected to remain below the corresponding threshold for a sustained period and remains stable, the system will automatically return from the warning status to the monitoring status, completing a full warning cycle.

[0268] 4. State Priority and Conflict Resolution Mechanism The early warning strategy engine uses a priority coverage mechanism to handle situations where multiple conditions are triggered simultaneously: Special warnings (based on the intensity of a single event) have the highest priority; The next level is a Level 1 warning (based on persistently high risk); Level II warning has the lowest priority.

[0269] When multiple triggering conditions are met simultaneously, the system only executes the highest priority warning action to ensure the clarity and effectiveness of the warning response.

[0270] S5012: Human-Computer Interaction (HMI) Execution Strategy.

[0271] 1. Visual channel Level 1 Warning: The in-vehicle central control screen or the dedicated area of ​​the instrument panel continuously flashes red, displaying the text or icon of "Rapid Acceleration / Rapid Deceleration Warning", and may also display the current safety score.

[0272] Level 2 Warning: The icon flashes orange without interrupting the current navigation or entertainment interface.

[0273] Special warning: Screen recording prompt (e.g., "Dangerous event has been recorded").

[0274] 2. Auditory channel Level 1 warning: Issues a short, sharp non-verbal warning sound (such as "beep beep beep").

[0275] Level 2 warning: Issues a gentle voice reminder (such as "Please drive smoothly").

[0276] S502: Cloud Reporting and Data Analysis Unit. This unit runs on the server side, aggregating and mining massive amounts of historical data to generate in-depth analysis reports for different roles.

[0277] S5021: Driving behavior analysis report generation.

[0278] Following step S4043, perform in-depth analysis of long-term historical data on the cloud platform to generate a comprehensive report. The comprehensive report includes: 1. Trip Summary: Trip distance, average speed, total energy consumption, safety score ( ).

[0279] 2. Event statistics: Bar charts or trend graphs showing the number of times, frequency (times / 100 km), and average intensity of risky operations such as rapid acceleration and deceleration.

[0280] 3. Scenario Analysis: Distribution of risky operations across different road types, time periods, and weather conditions.

[0281] 4. Battery Health Correlation: Based on long-term data, estimate the percentage potential impact of poor driving habits on battery life.

[0282] 5. Report Generation and Push: Reports are automatically generated periodically and pushed to users via email or mobile app. Fleet administrators can view reports for all their vehicles through a web portal.

[0283] S503: Data Interaction Interface Unit. This unit is responsible for opening up the system's output capabilities to external third-party systems, expanding the application ecosystem and value of the patent.

[0284] S5031: API (Application Programming Interface) Design.

[0285] 1. Interface type: Provides a RESTful API, using the standard HTTP protocol and JSON data format, making it easy for developers to integrate.

[0286] 2. Security Authentication: OAuth 2.0 or API Key mechanism is used to authenticate the identity and manage permissions of the caller.

[0287] S5032: Data interaction scenario construction.

[0288] 1. Insurance UBI Platform: Regularly retrieves users' driving behavior scores via API, which serve as the basis for adjusting the next year's premium.

[0289] 2. Automaker After-Sales System: When the system detects that a vehicle has been driven aggressively for a long period of time, it sends a reminder via API to the after-sales system to "suggest a deep battery health check", proactively creating service opportunities.

[0290] 3. Fleet Management Platform: Receives high-risk early warning events in real time and displays the safety status of all vehicles in the fleet on a map, facilitating dispatcher intervention.

[0291] 4. Traffic Management Platform: After anonymization and aggregation, it provides regional hotspot maps of rapid acceleration / deceleration to support decision-making for road safety management.

[0292] The following are examples: Example: Verification of driving behavior safety assessment based on battery energy consumption data Scene settings Vehicle parameters: Simulates an electric SUV (B-class car) with a rated power P = 150kW and a battery capacity of 60kWh.

[0293] Road Scenario: Design a 5-minute journey that includes the following road segments: Flat section (0-60 seconds): The vehicle travels at a constant speed and then suddenly performs a rapid acceleration (simulating a dangerous operation).

[0294] Uphill section (60-120 seconds): Gradient = 5%, vehicle goes uphill at medium power (simulating necessary operation).

[0295] Downhill section (120-180 seconds): Gradient = 4%, the vehicle performs energy recovery braking (simulating necessary operation).

[0296] Flat section (180-300 seconds): The vehicle performs a sudden deceleration (simulating a dangerous operation) and has a long period of intense driving (continuous high power).

[0297] Data sampling frequency: The method of this invention uses battery data (sampling frequency) =20Hz) and GPS data (sampling frequency 5Hz); the traditional method only uses GPS data (sampling frequency 1Hz, simulating typical GPS delay).

[0298] Key threshold Rapid acceleration: power change rate threshold =18kW / s, absolute power threshold =0.3×P_rated=45kW.

[0299] Rapid deceleration: Current change rate threshold = 25A / s, regenerative braking current threshold = 10A.

[0300] Prolonged aggressive driving: Window =60s, power threshold =0.5×P_rated=75kW, time percentage threshold =0.7.

[0301] Slope threshold: Uphill =5%, downhill = 3%.

[0302] Stress coefficient: Rapid acceleration sensitivity coefficient =0.15s / A, rapid deceleration sensitivity coefficient =0.18s / A, thermal susceptibility coefficient =0.003K / kJ.

[0303] Table 2 compares the simulation results of vehicle safe driving behavior, showing the simulated data at key time points and comparing the detection results of the traditional method (GPS acceleration only) and the method of this invention. GPS acceleration is calculated through position difference, which has a delay and error of about 1 second; the method of this invention uses the battery power change rate and real-time slope. The simulation results show that the technology proposed in this invention significantly reduces the false alarm rate and false negative rate, demonstrating the characteristics and advantages of the technology.

[0304] Table 2 Comparison of Simulation Experiment Results of Vehicle Safe Driving Behavior

[0305] .

[0306] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0307] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0308] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the new energy vehicle driving behavior safety assessment methods based on battery energy consumption data in the above embodiments.

[0309] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0310] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0311] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0312] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0313] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 the present invention.

Claims

1. A method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data, characterized in that, Includes the following steps, S1. Set up a data acquisition module, which is responsible for acquiring raw data from various vehicle sensors and buses; S2. Set up the data processing and feature extraction module; Responsible for cleaning, aligning, and fusing the raw, heterogeneous, and high-dimensional data collected by the data acquisition module, and extracting low-dimensional, high-value features that can directly characterize the essence of driving operations; S3. Set up a driving behavior recognition and quantification module; Used to receive a series of high-value features from S2, and through the designed set of discrimination rules and quantization model, transform the continuous data stream into discrete, interpretable driving operation events and evaluate their strength; S4. Set up a safety assessment and early warning module to receive a structured list of driving operation events from S3, establish an assessment system, transform discrete event sequences into comprehensive safety scores and real-time risk levels, and ultimately trigger corresponding early warning and reporting mechanisms. S5: Set up an early warning and output module to receive security scores, risk levels and structured event lists from S4, and build a multi-level, multi-channel, adaptive output and interaction system.

2. The method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data according to claim 1, characterized in that: S1 includes, S101: Set up a battery data acquisition unit to directly monitor and parse the vehicle's network bus to obtain first-hand data from the battery management system; S102: Set up a positioning data acquisition unit, responsible for acquiring the vehicle's spatiotemporal information and providing scene context for battery data; S103: Set up a data synchronization mechanism unit to solve the problem of different data sources having different frequencies and different delays; S101 includes the following: S1011: Data source acquisition channel, by listening to the CAN bus of the controller area network of the new energy vehicle power network to obtain the key status parameters of the battery pack broadcast by the battery management system at fixed intervals; S1012: Determine the physical interface; connect to the vehicle's OBD-II diagnostic interface via the vehicle-mounted intelligent terminal T-Box or dedicated data acquisition unit DAU, or connect directly to the CAN bus; S1013: Data acquisition parameters; the hardware terminal obtains the required data by listening to and parsing the specified message identifier; including: total current. Total voltage Instantaneous power ; S1014: Data sampling frequency setting; Required sampling frequency .

3. The method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data according to claim 2, characterized in that: S103 includes the following: S1031: Unified time base; Using the PPS pulse and UTC time output by the GNSS receiver as the absolute time reference for the entire system, when the vehicle-mounted terminal receives the PPS pulse, it associates it with the chip timer counter of the vehicle-mounted equipment to add a unified timestamp with microsecond precision to all subsequent data. ; S1032: Data alignment model; CAN data frequency is high, GPS data frequency is low, for each GPS data point Search the cache for the CAN data packet with the closest timestamp. The following requirements must be met: It is the tolerance threshold.

4. The method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data according to claim 1, characterized in that: S2 includes the following: S201. Set up a data preprocessing and synchronization alignment unit to ensure data consistency in time and space; S202. A battery time-domain feature extraction unit is set up to directly extract dynamic time-domain features that can reflect the driver's intention and the intensity of operation from the battery's current, voltage, and power signals. S203: Set up a trajectory feature extraction unit to extract features reflecting the road environment where the vehicle is located from GPS trajectory data; S202 specifically includes, S2021: Calculate the rate of change of the index; use the backward difference method for discrete approximation calculation; let the sampling time interval be... ;in: Rate of change of current: Power change rate: The power change rate value determines the severity of driving / braking behavior; S2022: Statistical Characteristic Analysis; Set a sliding time window of length N, calculate the variance within the window, and use it to measure the intensity of signal fluctuations. , Calculate the extreme values ​​within the window, and capture the maximum driving and braking currents within the window to determine the intensity of operation. 。 5. The method for assessing the driving behavior safety of new energy vehicles based on battery energy consumption data according to claim 4, characterized in that: S203 includes the following: S2031. Calculate the rate of change of elevation and estimate the slope; The slope calculation model using linear fitting is as follows: Take time point front and back Latitude, longitude, and elevation data for each step Record the starting point and end point The latitude and longitude coordinates are: ,in and It is a latitude value. and It is the longitude value; Calculate horizontal movement distance ; in, ； ； In the formula: It is a two-parameter arctangent function; Calculate the elevation change as follows The slope of the fitted line This refers to the average rate of change of elevation of the path segment, i.e., the estimated slope value. 。 6. The method for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data according to claim 1, characterized in that: S3 includes the following steps: S301. A rapid acceleration recognition unit is set up to identify abnormal acceleration behavior initiated by the driver. S302. A rapid deceleration identification unit is set up to identify abnormal braking behavior; S303: Equipped with a long-term aggressive driving recognition unit to identify aggressive driving states that last for more than a set time; S304: Set the output unit to output a structured list of driving operation events, providing the most direct input for S4 to perform comprehensive scoring and risk assessment; Each event Include: Event Type Rapid acceleration Rapid deceleration Prolonged and aggressive driving; Event start timestamp: ; Event duration: ; Event intensity : Intensity value of rapid acceleration or deceleration; Related environmental data: average slope and speed at the time of the event.

7. The method for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data according to claim 6, characterized in that: S301 includes the following: S3011: Decision criteria must be met simultaneously: Power change rate threshold: in, It is an empirical threshold; it directly reflects the explosive growth in power demand and is the most essential electrical characteristic of rapid acceleration. Absolute power threshold: ; in, The value is a certain percentage of the vehicle's rated power P. Exclude unnecessary operations: ; in, This is the uphill gradient threshold; if the vehicle is currently on a steep uphill slope, i.e., the gradient is ≥ If a high-power request is made, even if the first two conditions are met, it will not be judged as a "bad" rapid acceleration, thereby reducing the false alarm rate. S3012: Calculate the rapid acceleration intensity value: S302 includes the following: S3021: Decision criteria must be met simultaneously: Current change rate threshold: ; in, It is a negative threshold; the sharp negative change in current indicates that the drive motor is turning into a generator and entering a strong energy recovery mode. Regenerative braking judgment: ; in, It is a negative threshold; confirming that the current is clearly negative and in a regenerative braking state, thus eliminating other possible situations that could cause minor fluctuations in the current. Exclude unnecessary operations: in, It is the downhill threshold; if the vehicle is on a steep downhill slope, then strong energy recovery is a necessary safety behavior to control the vehicle speed and needs to be eliminated. S3022: Calculate the rapid deceleration intensity value: The intensity of rapid deceleration is directly proportional to the rate of change of current and the magnitude of the braking current; the larger the value, the more rapid the braking. S303 includes the following: S3031: Establish a discriminant model and statistically analyze the proportion of high power within a time window; Define a long-lasting sliding window ; Statistics within this window, instantaneous power Exceeding a certain moderate threshold Time percentage ; S3032: Decision criteria: The value is 0.

7. If the vehicle operates at high power for more than 70% of the time, then that period of time is considered aggressive driving.

8. The method for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data according to claim 1, characterized in that: S4 includes the following units: S401: Calculation unit for driving behavior-battery instantaneous stress relationship; used to assess the safety risks of rapid acceleration and deceleration events; instantaneous stress. The calculation formula is as follows: in, Indicates a driving operation event The instantaneous stress exerted on the battery system is the event intensity. The function; S402: Calculation unit for driving behavior-battery cumulative stress relationship; used to assess the safety risks of prolonged aggressive driving; its cumulative stress... Defined as a function of average power and duration; its calculation formula is as follows: in, The calculation formula is: In the formula: :time Average power; Long-lasting sliding window The number of sampling points within; :time Instantaneous power; A sliding window for recognizing prolonged aggressive driving; : Represents the thermal sensitivity coefficient, which is a constant; S403: Set up a comprehensive scoring calculation unit to calculate a safety score for a complete trip or a specified time period; S404: Set up early warning strategies and output units. Based on the evaluation results, a set of multi-level early warning triggering strategies, recovery strategies and report generation strategies based on real-time risk assessment were designed.

9. The method for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data according to claim 8, characterized in that: S401 includes the following: S4011: For rapid acceleration events, the stress estimation is as follows: S4012: For a rapid deceleration event, the stress estimate is as follows: and This represents the sensitivity coefficient for rapid acceleration and deceleration, and is a constant. S403 includes the following: S4031: Single trip scoring calculation uses a base score deduction system; initial score is set. =100; In the formula: Weighting coefficients for different event types; these weights are set according to the security importance of different event types and can be dynamically adjusted based on the scenario context output by S3. S4032: Real-time risk level assessment calculation; defining the sliding time window Calculate the real-time risk value within this window. This is used to represent the weighted strength of all operation events that occur within the accumulated sliding time window, and its output is a continuously changing numerical value. in: Low risk: < Medium risk: ≤ < High risk: ≥ In the formula: : Represents the weighting coefficient at time j; : Indicates the intensity of the event at time j; This is the low-risk threshold. This is a high-risk threshold. S404 includes the following: S4041: Warning Triggering Strategy The tiered early warning system based on real-time risk values ​​is as follows: The system continuously monitors and calculates real-time risk values. Different levels of alerts are triggered based on preset thresholds: Level 1 warning, also known as high-risk warning: when the real-time risk value... ≥ This level of warning is triggered when the driver is in a high-risk driving state. The system will issue a strong audio and visual warning through the in-vehicle human-machine interface, including a continuously flashing red warning icon and a short, sharp warning sound, to remind the driver to adjust their driving behavior immediately. Level 2 warning, also known as medium-risk warning: when ≤ < Triggered at a certain time; this level of warning is for medium-risk driving conditions. The system will issue a warning through a flashing orange icon and a gentle voice prompt, reminding the driver while avoiding excessive interference. The following are special warnings based on the intensity of a single event: In addition to risk-value-based early warnings, the system also has a special early warning mechanism for extreme single events: Special event warning: When the intensity of a single driving operation... Exceeding the absolute safety threshold At any time, regardless of the current real-time risk value, the system will immediately trigger the highest priority warning; this type of warning is for extremely dangerous driving operations, the system will record detailed data before and after the event, and issue the strongest level of warning; S4042: Early Warning Execution and Recovery Strategy After the warning is triggered, the system will continuously monitor the driving status, and will stop when the real-time risk value falls back to a safe range. After remaining stable for a period of time, the warning status will be automatically lifted, and the system will return to normal monitoring status. S4043: Cloud-based reporting and driver profile generation strategy; Constructing Feature Vectors: Extracting long-term features for each driver to form a feature vector; using the K-Means algorithm to classify drivers into profiles, automatically classifying drivers into "aggressive," "normal," and "mild" categories, providing accurate data support for refined fleet management and insurance UBI pricing; 。 10. The method for safety assessment of driving behavior of new energy vehicles based on battery energy consumption data according to claim 1, characterized in that: S5 includes the following units: S501. A real-time vehicle warning unit is set up to provide feedback to the driver when a risk occurs and correct bad driving behavior. S502, the cloud-based reporting and data analysis unit, runs on the server side and is used to aggregate and mine massive amounts of historical data to generate in-depth analysis reports for different roles. S503, Set up a data interaction interface unit, which is responsible for opening up the system's output capabilities to external third-party systems.

Images