Electronic control intelligent adjusting method and system for vehicle accessories

Abstract

The invention discloses an electric control intelligent adjustment method and system for vehicle accessories. The method comprises the steps that a data dynamic feature model is constructed to continuously analyze the extreme value change trend of monitoring data of multiple sensors; starting a saturation arithmetic operation mechanism for the calculation process of all numerical values in the control unit; the output instruction is limited within the safe operation range of the actuator; and when an unrecoverable abnormal state is detected, a system protection mechanism is automatically triggered, and switching of the control right and ordered degradation of the system state are executed. According to the method, a complete closed-loop protection system is formed from early warning to protection, from verification to constraint and finally to safety guarantee, whole-process protection from a calculation source to an execution terminal is achieved, the technical problem that numerical value overflow caused by multiplication of multiple limit parameters cannot be dealt with in the prior art is solved, and the safety of the terminal is improved. And control instruction abnormity caused by value overflow is effectively avoided.

Description

Technical Field

[0001] This invention relates to the field of fan noise reduction technology, and more specifically, to an electronically controlled intelligent adjustment method and system for vehicle accessories. Background Technology

[0002] With the rapid development of automotive electronics technology, various vehicle component systems have generally achieved electronically controlled intelligent adjustment. These electronically controlled intelligent adjustment systems for vehicle components, such as active suspension systems and electric steering systems, generally rely on embedded microcontrollers for real-time numerical calculations. By collecting sensor data and executing control algorithms, they ultimately drive actuators to complete precise adjustments.

[0003] Under normal driving conditions, this type of system can operate stably and reliably. When the control algorithm processes sensor data, all input variables are within normal ranges, and the calculation results are entirely within the effective representation range of the microcontroller's data type. Taking a typical stress estimation algorithm for a suspension system as an example, the calculated stress value is equal to the product of the current load, the slope compensation coefficient, and the dynamic safety factor. Under normal operating conditions, the values ​​of each multiplier factor are reasonable, and their product can be correctly represented and processed by the control system.

[0004] However, when a vehicle faces complex scenarios with multiple extreme operating conditions, such as a fully loaded vehicle with heavy luggage racks climbing a steep mountain road with a gradient significantly greater than that of ordinary roads, the sensor data collected by the system under these conditions is already at a high level. While the load values ​​detected by the weight sensor and the slope values ​​measured by the tilt sensor are within normal ranges, they are close to their upper limits. In this situation, when the control system executes the aforementioned stress estimation algorithm, the product of the extremely high current load value and the extremely high slope compensation coefficient will increase exponentially, easily exceeding the maximum representation range of the integer or fixed-point data type used by the microcontroller, leading to numerical overflow.

[0005] This type of numerical overflow, caused by the superposition of multiple limiting parameters and amplification through multiplication and other operations, is highly deceptive and destructive. The system will treat the overflowed erroneous value as a legitimate control command, thereby driving actuators such as solenoid valves or motors in the suspension system to perform dangerous actions. For example, it may cause the suspension system to output maximum force or the steering system to make excessive corrections, leading to serious consequences such as mechanical overload of vehicle components and failure of the control system, ultimately resulting in damage to vehicle components or even driving accidents. Summary of the Invention

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention discloses an electronically controlled intelligent adjustment method for vehicle accessories. The vehicle includes an adjustment system, which includes multiple sensors for monitoring the vehicle's operating status, an actuator for executing adjustment commands, and a control unit connecting the sensors and the actuator. The method includes:

[0008] A dynamic data feature model is constructed to continuously analyze the extreme value change trends of data monitored by multiple sensors, thereby identifying in real time whether the vehicle component adjustment system is under extreme conditions.

[0009] When extreme operating conditions are identified, a saturated arithmetic operation mechanism is enabled for the calculation process of all values ​​in the control unit to ensure that all intermediate calculation results and final output values ​​remain within the valid numerical range.

[0010] The calculation results after saturated arithmetic operations are cross-validated with the expected range output by the data dynamic feature model. Based on the principle of consistency with physical laws, abnormal values ​​that do not conform to physical laws are identified and corrected to obtain the verified calculation results.

[0011] A dynamic constraint boundary model is established based on the mechanical performance limits of vehicle parts. The output amplitude is limited based on the verified calculation results, and the output command is restricted to the safe operating range of the actuator.

[0012] When any of the extreme condition identification, cross-validation, or output amplitude limiting processes detects an unrecoverable abnormal state, the system protection mechanism is automatically triggered, resulting in a switch of control and an orderly degradation of the system state.

[0013] Furthermore, the spatial magnetic field vector data is acquired through an array of magnetoresistive sensors deployed on the surface of the metal support.

[0014] Furthermore, the saturation arithmetic operation mechanism is implemented by calling the processor's saturation arithmetic instructions through the compiler's built-in function calls. The saturation arithmetic instructions include the SSAT and USAT instructions of the ARM processor. When the numerical calculation result exceeds the representation range of the target data type, the saturation arithmetic instructions clamp the result to the maximum or minimum value that the data type can represent.

[0015] Furthermore, the construction of the dynamic feature model of the data includes:

[0016] The standard deviation and gradient of the sensor data are calculated within the sliding time window. When the standard deviation exceeds the first threshold and the gradient exceeds the second threshold, it is determined to be an extreme working condition.

[0017] The data dynamic feature model outputs a confidence score that characterizes the hazard level of the working condition.

[0018] Furthermore, cross-validation employs a Bayesian probability model. The Bayesian probability model uses the previous time-to-time's valid calculation result as the prior probability and the current saturated arithmetic operation result as the observation value. The posterior probability is calculated using the Bayesian formula. When the posterior probability is lower than a preset threshold, it is judged as an abnormal value and corrected.

[0019] Furthermore, the correction process of the Bayesian probability model adopts the Kalman filtering algorithm, and the calculation results are optimally estimated through the state space equation, which contains the process noise covariance matrix derived from the data dynamic feature model.

[0020] Furthermore, the dynamic constraint boundary model establishes a safe operating area (SOA) based on the material fatigue limit and structural strength of vehicle components. The SOA is represented by a convex polyhedron in a multidimensional space, where each dimension corresponds to a key mechanical parameter.

[0021] The boundary values ​​of the Safe Operating Area (SOA) are dynamically adjusted by real-time monitoring of component temperature and number of operating cycles, while adaptive boundary conditions that take into account material performance degradation are established.

[0022] Furthermore, the system protection mechanism is triggered using a tiered strategy:

[0023] When a Level 1 anomaly is detected, the system control performance is reduced;

[0024] When a Level 2 anomaly is detected, the system switches to the backup control unit.

[0025] When a Level 3 anomaly is detected, an orderly shutdown procedure is executed.

[0026] Furthermore, methods for implementing an orderly shutdown process include:

[0027] The entire process, from stopping non-critical loads to gradually reducing the control frequency and then maintaining basic safety monitoring functions, is completed within a preset time.

[0028] Secondly, the present invention discloses an electronically controlled intelligent adjustment system for vehicle parts, used to implement the aforementioned electronically controlled intelligent adjustment method for vehicle parts. The electronically controlled intelligent adjustment system for vehicle parts includes: a dynamic range sensing module, a saturation arithmetic processing module, a cross-validation module, a dynamic constraint module, and a system protection module.

[0029] The dynamic range sensing module is used to continuously analyze the extreme value change trends of data monitored by multiple sensors and build a dynamic feature model of the data;

[0030] The saturated arithmetic processing module is used to enable saturated arithmetic operations in the calculation process of all values ​​in the control unit when extreme operating conditions are identified.

[0031] The cross-validation module is used to cross-validate the results of saturated arithmetic operations with the expected range of the data dynamic feature model;

[0032] The dynamic constraint module is used to establish dynamic constraint boundary models based on the mechanical performance limits of vehicle components;

[0033] The system protection module is used to automatically trigger the system protection mechanism when an unrecoverable abnormal state is detected.

[0034] Compared with related technologies, the present invention has the following beneficial effects:

[0035] This invention uses a dynamic feature model to monitor sensor data in real time, identify extreme operating conditions, and provide early warnings for subsequent protection. Upon identifying an extreme condition, a saturated arithmetic operation mechanism is activated to ensure all numerical calculations remain within a valid range, preventing numerical overflow at the source. Next, by cross-validating the calculation results with physical laws, abnormal values ​​are identified and corrected to ensure the rationality of control commands. Then, dynamic constraint boundaries are established based on mechanical performance limits to limit the amplitude of output commands, ensuring the actuator operates within a safe range. Finally, when any link detects an unrecoverable anomaly, the system protection mechanism is triggered, executing control switching and orderly degradation to form the final safety guarantee. These five steps are interconnected, from early warning to protection, from verification to constraint, and finally to safety guarantee, forming a complete closed-loop protection system. This achieves end-to-end protection from the source of calculation to the execution terminal, solving the technical problem that existing technologies cannot handle numerical overflow caused by the multiplication of multiple extreme parameters. It effectively avoids abnormal control commands due to numerical overflow, reducing the probability of vehicle component damage and traffic accidents.

[0036] This invention elevates numerical overflow protection from the software level to the hardware instruction level by calling the processor's saturation arithmetic instructions through compiler built-in functions. By directly calling the processor's saturation arithmetic instructions, nanosecond-level overflow protection response is achieved, ensuring that the system can still meet real-time requirements under the highest load conditions. At the same time, the hardware-level data clamping mechanism avoids the boundary condition omission problem that may occur in traditional software protection.

[0037] This invention upgrades the safe operating area from a traditional fixed boundary to a multi-dimensional dynamic boundary based on materials science. By dynamically adjusting the safety boundary through real-time monitoring of parameters such as component temperature and load history, the system can fully tap the hardware potential while ensuring safety. This avoids performance waste caused by conservative design and prevents mechanical failures caused by overload operation, achieving an optimal balance between safety and performance indicators. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the steps of an electronically controlled intelligent adjustment method for vehicle parts provided by the present invention.

[0039] Figure 2 This invention provides a schematic diagram of the data processing flow of an electronically controlled intelligent adjustment system for vehicle parts. Detailed Implementation

[0040] 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 only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Example 1:

[0042] Please see Figure 1 As shown in Embodiment 1, this invention provides an electronically controlled intelligent adjustment method for vehicle components. This method is mainly applied to vehicles. Specifically, the vehicle is equipped with an adjustment system, which mainly includes multiple sensors for monitoring the vehicle's operating status, actuators for executing adjustment commands, and a control unit connecting the sensors and the actuators. In this embodiment, the sensor group can be a stress sensor group arranged at key stress-bearing parts such as the spring supports and shock absorber connection points of the suspension system, using strain gauge force sensors with a selectable range of 0-5000N and a sampling frequency of 1000Hz. The attitude sensor group can include a triaxial accelerometer (selectable range ±16g) and a triaxial gyroscope (selectable range ±2000° / s) with a sampling frequency of 400Hz. The load sensor can be installed in the air spring air pressure line of the chassis, with a selectable range of 0-2.5MPa and a sampling frequency of 100Hz. All sensors are connected to the control unit via a CAN bus, and the control unit can use a 32-bit ARM-Cortex-M7 microprocessor.

[0043] A method for electronically controlled intelligent adjustment of vehicle parts includes the following steps:

[0044] Step 1: Construct a dynamic data feature model. By continuously analyzing the extreme value change trends of data monitored by multiple sensors, identify in real time whether the vehicle component adjustment system is under extreme operating conditions.

[0045] For example, the construction of the data dynamic feature model includes calculating the standard deviation and gradient of the sensor data within a sliding time window. When the standard deviation exceeds a first threshold and the gradient exceeds a second threshold, it is determined to be an extreme working condition. The data dynamic feature model outputs a confidence score that characterizes the hazard level of the working condition.

[0046] Specifically, when constructing a dynamic feature model of data, the first step is to perform data preprocessing. Three levels of filtering are applied to the raw sensor data: hardware low-pass filtering with a cutoff frequency of 200Hz; software median filtering with a window length of 5 sampling points; and Kalman adaptive filtering to eliminate measurement noise.

[0047] Then, extreme value trend analysis is performed, which can establish a sliding time window mechanism with a fixed window length of 2 seconds. Next, at each sampling time t, the following calculations are performed:

[0048] For each sensor channel k: update the maximum value within the window in real time. and minimum value ;

[0049] Calculate the dynamic range index The calculation formula is as follows:

[0050] ;

[0051] in Let k be the full-scale value of sensor k.

[0052] The extreme value change trend of sensor monitoring data can be represented by the trend strength index. The trend strength index is calculated as follows: :

[0053] ;

[0054] in The forgetting factor has a value of 0.9. , This represents the raw reading of the k-th sensor channel at time t after preprocessing (filtering). This represents the change in the raw reading of the k-th sensor channel within the time interval Δt.

[0055] Finally, there's multi-sensor collaborative recognition. First, we define the extreme condition recognition function. :

[0056] ;

[0057] in, The weighting coefficients for the dynamic range of each sensor are 0.4 for stress sensors, 0.3 for attitude sensors, and 0.3 for load sensors. The weighting coefficients for the trend intensity of each sensor are 0.5 for stress sensors, 0.3 for attitude sensors, and 0.2 for load sensors. This is the trend compensation coefficient, with a value of 1.2.

[0058] when ( When the value is 0.75 (an empirical threshold), the system is determined to have entered an extreme operating condition.

[0059] For the real-time identification mechanism, the control unit executes the following judgment process every 1ms: reads the latest data from each sensor and updates the sliding window; calculates the data of each channel in parallel. and ; Calculate the comprehensive identification function ;Compare With threshold Output the extreme condition flag to the subsequent processing module.

[0060] Here are a few examples of the technical effects in practical application scenarios:

[0061] Scenario 1: For heavy-load climbing conditions, with a slope of 12% and a load of 380kg, the F(t) value increases from 0.35 to 0.82 in 3.2 seconds, identifying the extreme working condition 1.8 seconds in advance.

[0062] Scenario 2: For emergency lane change scenarios, at a speed of 80 km / h, F(t) jumps from 0.28 to 0.91 within 0.5 seconds, providing a warning 0.3 seconds in advance.

[0063] Scenario 3: For continuous bumpy roads, when driving at 60km / h on Belgian roads, F(t) is maintained in the range of 0.68-0.85, accurately identifying the continuous extreme state.

[0064] This implementation method achieves accurate and rapid identification of extreme working conditions through multi-sensor data fusion and dynamic trend analysis, providing a reliable basis for subsequent saturation arithmetic operations and the activation of system protection mechanisms.

[0065] Step two, please combine Figure 1 and Figure 2 When extreme operating conditions are identified, a saturated arithmetic operation mechanism is enabled for the calculation process of all values ​​in the control unit to ensure that all intermediate calculation results and final output values ​​remain within the valid numerical range; if extreme operating conditions are not identified, the normal calculation mode is maintained.

[0066] For example, the control unit may employ a 32-bit microprocessor with a saturated arithmetic instruction set, supporting saturated arithmetic operations on both signed and unsigned numbers. The processor's built-in memory protection unit can be configured to monitor the overflow status of the data bus and trigger a protection mechanism upon detecting an anomaly.

[0067] For example, the saturation arithmetic operation mechanism is implemented by the compiler's built-in function call to the processor's saturation arithmetic instructions. The saturation arithmetic instructions include the SSAT and USAT instructions of the ARM processor. When the numerical calculation result exceeds the representation range of the target data type, the saturation arithmetic instructions clamp the result to the maximum or minimum value that the data type can represent.

[0068] Specifically, upon receiving an extreme condition warning signal from the dynamic range sensing mechanism, the system performs the following configuration operations: enables the processor's saturation operation mode and sets the saturation operation enable flag in the status register; allocates a dedicated saturation operation memory area for critical computing tasks, which has hardware protection functions; and switches the system arithmetic operation library to the saturation operation version, covering all numerical calculation processes.

[0069] To address different computational needs, a data type boundary mapping table is established: stress calculation uses 32-bit signed integers with a value range limited to [-5000, 5000]; position calculation uses 32-bit signed integers with a value range limited to [-1000, 1000]; control output uses 16-bit signed integers with a value range limited to [-4095, 4095]; the boundary values ​​for each data type are pre-set according to their physical meaning and system safety requirements.

[0070] In implementing saturation arithmetic logic, the processor's dedicated saturation arithmetic instructions are first used. These instructions ensure at the hardware level that the operation result will not exceed the representation range of the target data type. When the operation result exceeds the range, the processor automatically clamps the result to the maximum or minimum value that the data type can represent.

[0071] Next, software-level boundaries are constrained, adding a secondary boundary check at the software level on top of instruction-level protection. For each numerical calculation result, the following constraint function is executed. :

[0072] ;

[0073] in, This is the original calculation result. and The upper and lower boundaries are preset. This is the final output value.

[0074] Through comparative testing, this implementation method demonstrates significant advantages in the following aspects:

[0075] Numerical stability test: Under standard working conditions, the consistency rate between saturated calculation and traditional calculation results reached 100%; under extreme overload conditions, such as a load of 600kg and a slope of 15%, traditional calculation resulted in 128 overflow values, with a maximum error of 32768; saturated calculation achieved zero overflow, and all outputs remained stable within the preset boundary value range.

[0076] Real-time performance test: The additional time overhead of saturation operation is less than 2 microseconds / time; under worst conditions, the overall calculation delay increases by less than 0.5 milliseconds; fully meets the strict real-time requirements of the control system (system response time is less than 10 milliseconds).

[0077] The following uses stress calculation as an example to illustrate the specific implementation method of protection.

[0078] In the heavy-load climbing scenario, the system needs to perform stress calculation: stress value = current load × slope compensation coefficient × dynamic safety factor. When the load value is 4200N, the slope compensation coefficient is 2.8, and the dynamic safety factor is 1.5, the theoretical calculation result is 4200 × 2.8 × 1.5 = 17640N.

[0079] Specific protection process: The system detects that it is currently in an extreme working condition and automatically activates the saturated arithmetic operation mode; it uses 32-bit signed integers for calculation, with a valid range of -2147483648 to 2147483647; the processor's SSAT instruction ensures that intermediate results and final output values ​​are always kept within the valid range; even if a calculation anomaly occurs, the output value is limited to the preset safe range [-5000, 5000].

[0080] This implementation method uses a saturated arithmetic operation mechanism that combines hardware and software to ensure the numerical integrity of all calculation processes and processing results under extreme operating conditions, providing key technical support for the safe and stable operation of the vehicle's electronic control intelligent adjustment system.

[0081] Step 3: Cross-validate the calculation results after saturated arithmetic operations with the expected range output by the data dynamic feature model. Based on the principle of consistency with physical laws, identify and correct abnormal values ​​that do not conform to physical laws, and obtain the verified calculation results.

[0082] For example, cross-validation employs a Bayesian probability model. The Bayesian probability model uses the previous time-series valid computation results as the prior probability and the current saturated arithmetic operation result as the observation. The posterior probability is calculated using Bayes' theorem. When the posterior probability falls below a preset threshold, it is identified as an outlier and corrected. Furthermore, the correction process of the Bayesian probability model uses a Kalman filter algorithm. The calculation results are optimally estimated using a state-space equation, which includes the process noise covariance matrix derived from the data dynamic feature model.

[0083] Specifically, the first step is the architecture design of the cross-validation system. In this embodiment, the cross-validation system can include three core components: a baseline value generation module, which outputs the expected reasonable range of each parameter based on the dynamic feature model of the data; a consistency verification module, which uses a Bayesian probability model to evaluate the conformity with physical laws; and an anomaly correction module, which implements an adaptive correction algorithm for the identified abnormal values.

[0084] The benchmark generation module calculates the expected reasonable range for each parameter based on the trend data output by the dynamic feature model. For each key parameter... Its expected range is expressed as:

[0085] ;

[0086] in, The minimum value of the parameter. The maximum value of the parameter. It is a moving average calculated based on historical data. The standard deviation within the sliding window. This is the confidence coefficient, which is usually taken as 2.0.

[0087] Next, the system establishes a set of physical constraints based on vehicle dynamics principles, including: linear relationship constraints between suspension stress and vehicle load; dynamic response constraints between vehicle body attitude and steering input; balance constraints on force distribution at each suspension point; and constraints for each calculation result after saturated arithmetic operations. Calculate its physical law compliance degree :

[0088] ;

[0089] in, Let j represent the constraint function of the j-th physical law. The system state vector output by the dynamic feature model. This represents the total number of physical law constraint functions.

[0090] Then, anomaly identification and correction are performed, and the calculation results are... A value is considered abnormal if it does not meet any of the following conditions:

[0091] Condition 1: ;

[0092] Condition 2: ;in, To determine the compliance threshold, a preset judgment threshold (e.g., 0.85) can be set. If the result is below this threshold, it is considered to be inconsistent with physical laws.

[0093] For outlier values, a correction algorithm based on Kalman filtering is used, as follows:

[0094] ;

[0095] in, The corrected output value is the reasonable value that is finally adopted by the system after being processed by the Kalman filter algorithm. This is the theoretical expected value of the parameter at the current moment, which can be calculated based on the system state and physical law model; The difference between observed and predicted values ​​reflects the uncertainty or error in the current measurement or calculation; For adaptive Kalman gain.

[0096] Through real-vehicle testing, the results of this cross-validation mechanism can be summarized in two main aspects: First, the accuracy of anomaly detection: under normal operating conditions, the false alarm rate is less than 0.1%; the detection rate of abnormal commands caused by numerical overflow reaches 99.8%; and the average detection latency is less than 2ms. Second, the evaluation of correction effect: the physical rationality of the corrected abnormal values ​​is improved by 85%; the stability of the system is significantly improved after correction; and the control accuracy is improved by 40% compared with traditional methods.

[0097] This implementation method utilizes a cross-validation mechanism to ensure the rationality of the control system's decision-making under extreme conditions through a multi-level and multi-angle verification strategy. This forms a complete technical closed loop with the preceding and following steps, thereby improving the safety and reliability of the vehicle's electronic control intelligent adjustment system.

[0098] Step four: Establish a dynamic constraint boundary model based on the mechanical performance limits of the vehicle parts, and limit the output amplitude of the verified calculation results to restrict the output commands within the safe operating range of the actuator.

[0099] For example, the dynamic constraint boundary model establishes a safe operating area (SOA) based on the material fatigue limit and structural strength of vehicle components. The SOA is represented by a convex polyhedron in a multi-dimensional space, where each dimension corresponds to a key mechanical parameter. Furthermore, the boundary values ​​of the SOA can be dynamically adjusted by real-time monitoring of component temperature and operating cycle count, while establishing adaptive boundary conditions that consider material performance degradation.

[0100] Specifically, the dynamic constraint boundary model is established based on the material properties and mechanical structure limit parameters of vehicle components. Its main core data includes material fatigue limit data, namely the SN curves of key components such as suspension links and spring assemblies obtained through laboratory testing; structural strength parameters, namely the maximum allowable stress of each connection point obtained based on finite element analysis; actuator mechanical limits, namely physical limitation parameters such as the maximum stroke and maximum output force of electric actuators; and thermal management limits, namely thermal performance parameters such as the temperature rise of motor windings and the heat dissipation capacity of the driver.

[0101] When using the dynamic constraint boundary model, the safe operating region (SOA) is first modeled. This region consists of three parameter spaces: the stress-life two-dimensional plane, the displacement-velocity two-dimensional plane, and the temperature-load two-dimensional plane. The stress-life two-dimensional plane represents the allowable stress range based on the material's SN curve; the displacement-velocity two-dimensional plane represents the range of motion parameters based on the actuator's mechanical characteristics; and the displacement-velocity two-dimensional plane represents the operating condition range based on thermal management capabilities. The boundary function of the SOA can be expressed as:

[0102] ;

[0103] in, This is the system state vector; Let i be the i-th constraint function; Let be an n-dimensional real space, representing the space where the system state vector resides; The dimension of the system state vector represents the number of independent parameters that need to be considered. This represents the total number of constraints.

[0104] After the safe operating area is modeled, real-time boundary calculations are performed. The constraint boundaries can be dynamically calculated based on the current system state. The steps include calculating the material strength reduction factor based on the real-time temperature of the components; calculating the cumulative fatigue damage based on the load history; considering the impact of ambient temperature on actuator performance; and comprehensively calculating the safe operating boundary value at the current moment.

[0105] Finally, hard constraints are applied to the cross-validated calculation results. , can be represented as:

[0106] ;

[0107] in, These are verified calculation results. and This represents the dynamic boundary value at the current moment.

[0108] Through system testing, the dynamic constraint boundary model demonstrates the following safety assurance performance:

[0109] Under extreme operating conditions, the rate of output command over-limit occurrence is reduced by 99.5%; the stress of key components is always controlled below 80% of the material's yield strength; and the actuator's operating point is always maintained within the safe range of mechanical limits.

[0110] Furthermore, the dynamic constraint boundary model exhibits the following performance in terms of dynamic adaptability:

[0111] Boundary values ​​are adjusted in real time according to system status, with a response time of less than 10ms; they can accurately reflect the influence of slowly changing parameters such as temperature and fatigue accumulation; and they automatically adapt to the optimal operating range under different environmental conditions.

[0112] This implementation utilizes a dynamic constraint boundary model to establish an intelligent constraint mechanism based on physical characteristics, ensuring that the control system will not generate commands that exceed the mechanical system's capacity under any operating condition. This forms a complete technical closed loop with the preceding and following steps, providing the ultimate safety guarantee for the vehicle's electronic control intelligent adjustment system.

[0113] Step 5, please refer to Figure 2 As shown, when any of the extreme condition identification, cross-validation, or output amplitude limitation processes detects an unrecoverable abnormal state, the system protection mechanism is automatically triggered to switch control and perform orderly degradation of the system state; if no unrecoverable abnormal state is detected, the instructions are output to the actuator normally.

[0114] For example, the system protection mechanism can be triggered using a tiered strategy: when a Level 1 anomaly is detected, the system control performance is reduced; when a Level 2 anomaly is detected, the system switches to a backup control unit; and when a Level 3 anomaly is detected, an orderly shutdown procedure is executed. Furthermore, the method for executing the orderly shutdown procedure may include completing the entire process—from stopping non-critical loads to gradually reducing the control frequency and then maintaining basic safety monitoring functions—within a preset time frame.

[0115] Specifically, in this embodiment, the system protection mechanism can adopt a distributed fault-tolerant architecture, whose core components include an abnormal state monitoring network, a multi-level triggering decision unit, a control switching executor, and a state degradation sequence generator. The abnormal state monitoring network can deploy health status monitoring points in each processing module, the multi-level triggering decision unit is a comprehensive judgment system based on a voting mechanism, the control switching executor can realize a disturbance-free transfer of control, and the state degradation sequence generator has multiple pre-set degradation operation schemes.

[0116] Regarding the criteria for determining unrecoverable abnormal states, the system can establish a three-level abnormal state determination system:

[0117] For Level 1 anomalies (mild), such as data exceeding limits from a single sensor or an output anomaly from a single computing module, the trigger condition is set to... The handling strategy is generally to make local corrections without triggering system protection.

[0118] For Level 2 anomalies (moderate), such as multiple related parameters being abnormal simultaneously or critical calculation modules being abnormal consecutively, the trigger condition is set as follows: The typical handling strategy is to prepare for early warning and preload backup systems.

[0119] For level 3 anomalies (severe), such as core algorithm failure or executor response anomalies, the trigger condition is set to... If the system consistency check fails, the handling strategy needs to immediately trigger the system protection mechanism.

[0120] in, , , These represent the sets of Level 1, Level 2, and Level 3 abnormal states, respectively. This represents the j-th abnormal event; This represents the anomaly metric function corresponding to the j-th anomalous event, whose input is the system state vector. The output is a non-negative scalar, and the larger the value, the more severe the anomaly. Weights for different abnormal events; This is an indicator function; the function value is 1 when the condition inside the parentheses is true, and 0 otherwise; max indicates taking the maximum value. , The threshold is used to classify different levels of anomaly severity.

[0121] For the execution process of control handover, a three-stage non-disruptive handover strategy can be adopted. In the preparation stage, the backup control system first completes the initialization and synchronizes the operating status, then the main control system prepares to hand over control authority, and finally the system verifies the consistency of the status of the two control units.

[0122] During the switching phase, the switching is first performed when the main control system outputs a stable point, and then the output signal superposition method is used to achieve a smooth transition, and the switching time is controlled within 10ms.

[0123] During the confirmation phase, first verify that the output of the backup control system meets expectations, then shut down the output channel of the main control system, and finally update the system status flags.

[0124] In this embodiment, the system can be preset with four degradation modes for the orderly degradation of system state: performance optimization mode: reducing control accuracy requirements while maintaining basic functions; function simplification mode: disabling advanced adjustment functions while retaining safety assurance; minimum operation mode: maintaining only the most basic control capabilities; and safety maintenance mode: executing a safe shutdown sequence.

[0125] After system testing, the system's protection mechanism achieved the following performance indicators: response time from anomaly detection to protection trigger <5ms; complete control handover time <15ms; and degradation mode switching time <20ms. Based on these indicators, the system's protection mechanism can effectively prevent the spread of abnormal states and system crashes, ensuring that the system can maintain basic safe operation even in the event of a severe fault, and providing a final safety guarantee for the vehicle, thus preventing major accidents.

[0126] This implementation establishes a complete anomaly handling chain through a system protection mechanism, providing an ultimate safety barrier for the vehicle's electronic control intelligent adjustment system, and forming a complete technical closed loop with all the aforementioned technical steps, ensuring that the system can operate safely and orderly under any abnormal circumstances.

[0127] In summary, by using a dynamic feature model to monitor sensor data in real time and identify extreme operating conditions, early warnings are provided for subsequent protection. Upon identification of extreme conditions, a saturated arithmetic operation mechanism is activated to ensure all numerical calculations remain within a valid range, preventing numerical overflow at the source. Next, by cross-validating the calculation results with physical laws, abnormal values ​​are identified and corrected to ensure the rationality of control commands. Then, dynamic constraint boundaries are established based on mechanical performance limits to restrict the amplitude of output commands, ensuring the actuator operates within a safe range. Finally, when any link detects an unrecoverable anomaly, the system protection mechanism is triggered, executing control switching and orderly degradation to form the final safety guarantee. These five steps are interconnected, from early warning to protection, from verification to constraint, and finally to safety guarantee, forming a complete closed-loop protection system. This achieves end-to-end protection from the source of calculation to the execution terminal, solving the technical problem that existing technologies cannot handle numerical overflow caused by the multiplication of multiple extreme parameters. It effectively avoids abnormal control commands due to numerical overflow, reducing the probability of vehicle component damage and traffic accidents.

[0128] By calling the processor's saturation arithmetic instructions through compiler built-in functions, numerical overflow protection is elevated from the software level to the hardware instruction level. By directly calling the processor's saturation arithmetic instructions, nanosecond-level overflow protection response is achieved, ensuring that the system can still meet real-time requirements under the highest load conditions. At the same time, the hardware-level data clamping mechanism avoids the boundary condition omission problem that may occur in traditional software protection.

[0129] The safe operating area is upgraded from a traditional fixed boundary to a multi-dimensional dynamic boundary based on materials science. By dynamically adjusting the safety boundary through real-time monitoring of parameters such as component temperature and load history, the system can fully tap the hardware potential while ensuring safety. This avoids performance waste caused by conservative design and prevents mechanical failures caused by overload operation, achieving the optimal balance between safety and performance indicators.

[0130] Example 2:

[0131] This embodiment 2 provides an electronically controlled intelligent adjustment system for vehicle parts, used to implement the electronically controlled intelligent adjustment method for vehicle parts disclosed in embodiment 1.

[0132] An electronically controlled intelligent adjustment system for vehicle parts includes a dynamic range sensing module, a saturation arithmetic processing module, a cross-validation module, a dynamic constraint module, and a system protection module.

[0133] The dynamic range sensing module is configured to continuously analyze the extreme value change trends of data monitored by multiple sensors and build a dynamic feature model of the data.

[0134] The saturated arithmetic processing module is configured to enable saturated arithmetic operations in the calculation process of all values ​​in the control unit when extreme operating conditions are identified.

[0135] The cross-validation module is configured to cross-validate the results of saturated arithmetic operations with the expected range of the data dynamic feature model.

[0136] The dynamic constraint module is configured to establish a dynamic constraint boundary model based on the mechanical performance limits of vehicle components.

[0137] The system protection module is configured to automatically trigger the system protection mechanism when an unrecoverable abnormal state is detected.

[0138] Since this system uses an electronically controlled intelligent adjustment method for vehicle accessories as described in Example 1, it has the same effect and will not be repeated here.

[0139] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0140] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An electrically controlled intelligent adjustment method for a vehicle accessory, the vehicle including an adjustment system, the adjustment system including a plurality of sensors that monitor a state of operation of the vehicle, an actuator that executes an adjustment instruction, and a control unit that connects the sensors and the actuator, characterized by, The method comprises: A data dynamic characteristic model is constructed to identify whether the vehicle accessory adjustment system is in a limit working condition by continuously analyzing the extreme value change trend of the sensor monitoring data; When the limit working condition is identified, a saturation arithmetic operation mechanism is enabled for the calculation process of all values in the control unit to ensure that all intermediate calculation results and final output values are kept within the effective value range; The calculation results processed by the saturation arithmetic operation are cross-verified with the expected range output by the data dynamic characteristic model, and abnormal values that do not conform to the physical law are identified and corrected based on the consistency principle of physical law to obtain verified calculation results; A dynamic constraint boundary model is established according to the mechanical performance limit of the vehicle accessory to limit the output amplitude of the verified calculation results, and the output command is limited within the safe operation range of the actuator; When any of the limit working condition identification, cross-verification or output amplitude limitation processes detects an unrecoverable abnormal state, the system protection mechanism is automatically triggered to perform control switching and orderly degradation of system state.

2. The electrically controlled intelligent conditioning method for vehicle accessories according to claim 1, characterized in that, The saturation arithmetic operation mechanism is realized by calling the saturation arithmetic instruction of the processor through the built-in function of the compiler. The saturation arithmetic instruction includes the SSAT and USAT instructions of the ARM processor. When the numerical calculation result exceeds the representation range of the target data type, the saturation arithmetic instruction clamps the result to the maximum or minimum value that can be represented by the data type.

3. The method of claim 1, wherein the method is an electronically controlled intelligent conditioning method for a vehicle accessory. The construction of the data dynamic characteristic model comprises: The standard deviation and gradient of the sensor data are calculated within a sliding time window. When the standard deviation exceeds a first threshold value and the gradient exceeds a second threshold value, it is determined that the limit working condition is reached. The data dynamic characteristic model outputs a confidence score representing the danger level of the working condition.

4. The electronically controlled intelligent conditioning method for vehicle accessories according to claim 1, characterized in that, The cross-verification adopts a Bayesian probability model. The Bayesian probability model takes the valid calculation result at the previous moment as the prior probability and takes the current saturation arithmetic operation processing result as the observation value. The posterior probability is calculated by the Bayes formula. When the posterior probability is lower than a preset threshold value, it is determined that the abnormal value is corrected.

5. The method of claim 4, wherein the method further comprises: The correction process of the Bayesian probability model adopts the Kalman filter algorithm to optimally estimate the calculation result through the state space equation. The state space equation contains the process noise covariance matrix derived from the data dynamic characteristic model.

6. The electronically controlled intelligent conditioning method for vehicle accessories according to claim 1, characterized in that, The dynamic constraint boundary model establishes a safe operation area SOA based on the material fatigue limit and structural strength of the vehicle accessory. The safe operation area SOA is represented by a convex polyhedron in a multi-dimensional space, where each dimension corresponds to a key mechanical parameter.

7. The electrically controlled intelligent conditioning method for vehicle accessories according to claim 6, characterized in that, The boundary values of the safe operation area SOA are dynamically adjusted by the real-time monitored accessory temperature and working cycle number, and adaptive boundary conditions considering material performance degradation are established.

8. The method of claim 1, wherein the method is an electronically controlled intelligent conditioning method for a vehicle accessory. The triggering of the system protection mechanism adopts a hierarchical strategy: When a first-level abnormality is detected, the system control performance is reduced; When a second-level abnormality is detected, the backup control unit is switched to; When a third-level abnormality is detected, an orderly shutdown process is performed.

9. The method of claim 8, wherein the method further comprises: The method for performing the orderly shutdown process comprises: The entire process from stopping non-critical loads to gradually reducing the control frequency and then maintaining the basic safety monitoring function is completed within a preset time.

10. An electrically controlled intelligent adjustment system for vehicle accessories, for implementing the electrically controlled intelligent adjustment method for vehicle accessories according to any one of claims 1-9, characterized in that, The system comprises: a dynamic range awareness module for continuously analyzing extreme value change trends of the plurality of sensor monitoring data and constructing a data dynamic characteristic model; a saturation arithmetic processing module for enabling saturation arithmetic operations on all numerical value calculation processes in the control unit when the limit working condition is identified; a cross-validation module for cross-validating the saturation arithmetic operation results with the expected range of the data dynamic characteristic model; a dynamic constraint module for establishing a dynamic constraint boundary model according to the mechanical performance limit of the vehicle accessories; a system protection module for automatically triggering a system protection mechanism when an unrecoverable abnormal state is detected.

Images