Active suspension control method and device based on driving style recognition and road preview

Abstract

The invention relates to an active suspension control method and device based on driving style recognition and pavement preview. The method comprises the steps that vehicle state parameters and driving environment data in the vehicle driving process are collected in real time; determining a driving style based on the vehicle state parameters, and determining pavement preview information of the vehicle reaching different pavement areas based on the driving environment data; according to the driving style, the vehicle state parameters and the road surface preview information, determining a target regulation and control strategy for an active suspension of the vehicle; and according to the target regulation and control strategy, the active suspension executing mechanism is controlled to act. By adopting the method, the adaptive capacity of the vehicle suspension system can be improved.

Description

Technical Field

[0001] This application relates to the field of intelligent transportation technology, and in particular to an active suspension control method and device based on driving style recognition and road surface prediction. Background Technology

[0002] With the rapid development of the automotive industry, the vehicle suspension system, as a key component connecting the vehicle body and wheels, directly affects the vehicle's ride comfort, handling stability, and passability.

[0003] However, traditional suspension systems struggle to achieve a good balance between ride comfort, handling stability, and off-road capability in complex and varied usage scenarios. This results in insufficient adaptability to different usage needs and scenarios, failing to meet the comprehensive performance requirements of modern vehicles for suspension systems. Therefore, there is an urgent need to propose a new technical solution to address these issues and improve the adaptability of vehicle suspension systems. Summary of the Invention

[0004] Therefore, it is necessary to provide an active suspension control method and device based on driving style recognition and road surface prediction that can improve the adaptability of vehicle suspension systems and address the aforementioned technical problems.

[0005] In a first aspect, this application provides an active suspension control method based on driving style recognition and road surface prediction, the method comprising:

[0006] Real-time collection of vehicle status parameters and driving environment data during vehicle operation;

[0007] The driving style is determined based on vehicle state parameters, and the road surface prediction information for the vehicle to reach different road surface areas is determined based on driving environment data.

[0008] Based on driving style, vehicle status parameters, and road surface preview information, determine the target control strategy for the vehicle's active suspension.

[0009] Based on the target control strategy, control the movement of the active suspension actuator.

[0010] Secondly, this application also provides an active suspension control device based on driving style recognition and road surface prediction, the device comprising:

[0011] The data acquisition module is used to collect vehicle status parameters and driving environment data in real time during the vehicle's driving process.

[0012] The information determination module is used to determine the driving style based on vehicle status parameters and to determine the road surface prediction information for the vehicle to reach different road surface areas based on driving environment data.

[0013] The strategy determination module is used to determine the target control strategy for the vehicle's active suspension based on driving style, vehicle state parameters and road surface preview information.

[0014] The control module is used to control the actions of the active suspension actuators according to the target control strategy.

[0015] Thirdly, this application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0016] Real-time collection of vehicle status parameters and driving environment data during vehicle operation;

[0017] The driving style is determined based on vehicle state parameters, and the road surface prediction information for the vehicle to reach different road surface areas is determined based on driving environment data.

[0018] Based on driving style, vehicle status parameters, and road surface preview information, determine the target control strategy for the vehicle's active suspension.

[0019] Based on the target control strategy, control the movement of the active suspension actuator.

[0020] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the following steps:

[0021] Real-time collection of vehicle status parameters and driving environment data during vehicle operation;

[0022] The driving style is determined based on vehicle state parameters, and the road surface prediction information for the vehicle to reach different road surface areas is determined based on driving environment data.

[0023] Based on driving style, vehicle status parameters, and road surface preview information, determine the target control strategy for the vehicle's active suspension.

[0024] Based on the target control strategy, control the movement of the active suspension actuator.

[0025] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0026] Real-time collection of vehicle status parameters and driving environment data during vehicle operation;

[0027] The driving style is determined based on vehicle state parameters, and the road surface prediction information for the vehicle to reach different road surface areas is determined based on driving environment data.

[0028] Based on driving style, vehicle status parameters, and road surface preview information, determine the target control strategy for the vehicle's active suspension.

[0029] Based on the target control strategy, control the movement of the active suspension actuator.

[0030] The aforementioned active suspension control method and device based on driving style recognition and road surface prediction provides comprehensive and real-time perception data for the suspension system by frequently collecting vehicle state parameters and fusing driving environment data from multiple sensors. Based on this data, it determines the driving style that aligns with the driver's preferences and provides forward-looking road surface prediction information, clarifying the direction and lead time for suspension adaptation. Subsequently, it integrates driving style, vehicle state parameters, and road surface prediction information to formulate a target control strategy, forming a suspension parameter scheme that accurately matches multi-dimensional needs. Finally, according to the target control strategy, it controls the action of the active suspension actuator. This method forms a closed loop from perception and decision-making to execution, achieving flexible adaptation to different driving preferences, vehicle dynamics, and road conditions, significantly improving the adaptability of the vehicle suspension system. Attached Figure Description

[0031] Figure 1 This is a diagram of the internal structure of an electronic device in one embodiment;

[0032] Figure 2 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in one embodiment.

[0033] Figure 3 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0034] Figure 4 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0035] Figure 5 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0036] Figure 6 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0037] Figure 7 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0038] Figure 8 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0039] Figure 9 This is a flowchart illustrating an active suspension control method based on driving style recognition and road surface prediction in another embodiment;

[0040] Figure 10 This is a structural block diagram of an active suspension control device based on driving style recognition and road surface prediction in one embodiment. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0042] With the rapid development of the automotive industry, the vehicle suspension system, as a key component connecting the vehicle body and wheels, directly affects the vehicle's ride comfort, handling stability, and off-road capability. Traditional passive suspension systems, due to their inherent structural characteristics, struggle to achieve a good balance among these three aspects, exhibiting significant limitations, especially under complex and varied road conditions and different driving styles.

[0043] To address this issue, active suspension systems have emerged. Active suspension systems improve a vehicle's overall performance to some extent by adjusting suspension parameters (such as stiffness, damping, and height) in real time. Existing technology includes top-level mode control—which involves assigning algorithms based on operating conditions and driving style determination, and dynamically adjusting parameters—as well as low-level algorithm control for specific execution of control algorithms. This allows for the dynamic selection and adjustment of low-level control methods and parameters, thereby optimizing the performance of electronically controlled semi-active suspension systems.

[0044] In recent years, with the development of artificial intelligence and sensor technology, driving style recognition technology has gradually become an important component of vehicle control systems. Existing technology two constructs a driving condition information database by acquiring brake, accelerator pedal opening, steering wheel angle, and vehicle speed information from the CAN bus. It then identifies the current driving condition by extracting pedal opening and steering wheel angle information and comparing them with preset thresholds. Simultaneously, it performs rolling segmentation of historical vehicle speed trajectories based on time windows to extract statistical and temporal features to identify driving styles. Finally, it uses the driving condition and driving style as inputs to achieve active suspension control.

[0045] On the other hand, significant progress has also been made in the application of anti-sighting control technology in active suspension systems. Existing technology three uses a binocular vision recognition algorithm to identify and classify random road surface information during vehicle operation. The anti-sighting control algorithm then inputs this road surface classification information into the vehicle-seat active suspension system model as an excitation signal, thereby improving the magnitude of the vehicle's vertical acceleration and effectively enhancing the ride comfort and handling stability of autonomous vehicles.

[0046] To further expand the range of pre-aiming control, existing technology 4 uses a combined positioning system based on low-orbit satellite positioning, high-precision maps, and inertial navigation systems to locate the vehicle and identify road conditions. At the same time, it establishes signal interaction between vehicles to realize the interactive transmission of information about road conditions, suspension parameters, and vehicle positioning between the vehicle and surrounding vehicles. This allows for the formulation of beyond-line-of-sight pre-aiming control strategies, effectively solving problems such as limited perception range and insufficient computing power of onboard electronic control units during pre-aiming control.

[0047] In addition, the existing technology 5 sets the basic suspension parameters of the whole vehicle based on the driving mode, and then determines the suspension system correction parameters based on the driving style analysis of the driver's intention, thereby controlling the suspension system and effectively improving the adaptability of suspension control and the responsiveness of the suspension system.

[0048] Existing technologies still have the following shortcomings: First, most existing active suspension control systems are still based on fixed control logic. Although they consider driving style or road condition information, they lack a comprehensive control strategy that organically combines the two, making it impossible to fully adapt to different driving styles and changing road conditions. Second, existing driving style recognition technologies are mostly based on single or limited feature parameters, and their recognition accuracy and real-time performance need improvement, making it difficult to capture subtle changes in driver behavior. Third, anticipation control technology still has limitations in road information acquisition and processing, and the efficiency of multi-sensor fusion and information processing needs further improvement. Finally, existing technologies lack multi-objective optimization considerations in suspension parameter adjustment strategies, making it difficult to achieve an optimal balance among multiple performance indicators such as ride comfort, handling stability, and traction.

[0049] Therefore, there is a need for an active suspension control method that can integrate driving style recognition and road surface preview information. By recognizing driving style and perceiving road conditions ahead in real time, the suspension parameters can be dynamically adjusted to achieve multi-objective optimization of vehicle performance, thereby meeting the personalized needs of different drivers under various road conditions.

[0050] To address the technical challenges of traditional vehicle suspension systems in balancing ride comfort, handling stability, and passability, and the inability of existing active suspensions based on fixed control logic to adapt to different driving styles and road conditions, this application provides an active suspension control method based on driving style recognition and road surface prediction to achieve personalized driving experiences, improved vehicle performance, and intelligent energy-saving optimization.

[0051] In one embodiment, an electronic device is provided, the internal structure of which can be shown as follows: Figure 1As shown, this electronic device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores relevant data in the active suspension control process based on driving style recognition and road surface prediction. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an active suspension control method based on driving style recognition and road surface prediction.

[0052] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0053] In one embodiment, such as Figure 2 As shown, an active suspension control method based on driving style recognition and road surface prediction is provided, which is then applied to... Figure 1 Taking an electronic device as an example, the explanation includes the following steps:

[0054] Step 201: Collect vehicle status parameters and driving environment data in real time during the vehicle's driving process.

[0055] Among them, vehicle status parameters are key physical quantities and operational data that reflect the real-time operating status of a vehicle during driving, and serve as the basis for identifying driving style and judging the dynamic operating conditions of the vehicle. These mainly include accelerator pedal travel, brake pedal travel, steering wheel angle, vehicle longitudinal acceleration, and vehicle lateral acceleration, which are collected in real time through the vehicle's internal sensor network.

[0056] Driving environment data refers to raw data related to the surrounding driving scene collected by onboard external sensing devices during vehicle operation. It forms the basis for obtaining road surface preview information. Common data collection devices include visual cameras, LiDAR, and millimeter-wave radar. The data content covers image information, point cloud information, and target detection information of the road surface ahead.

[0057] In this embodiment, the electronic device collects vehicle status parameters and driving environment data in real time during vehicle operation. First, a connection is established with the vehicle's internal sensor network via the vehicle bus, with the data acquisition frequency set to 100Hz. When collecting vehicle status parameters, real-time accelerator pedal travel data is received to record the displacement changes when the driver presses the accelerator; brake pedal travel is simultaneously acquired to capture the magnitude of deceleration; steering wheel angle is also collected to track dynamic changes in the vehicle's steering angle; furthermore, longitudinal and lateral acceleration data are continuously received to reflect the dynamic state of the vehicle during acceleration, deceleration, and turning. In the driving environment data collection phase, the electronic device captures road surface images within a 100-meter range ahead using a front-facing visual camera, generating image data every 30 frames per second; it scans the three-dimensional space within a 200-meter range ahead using lidar to obtain point cloud information on road undulations and obstacles; and it detects the distance and relative speed of targets ahead using millimeter-wave radar, updating the detection data every 20 times per second. During the data acquisition process, the electronic equipment performs time calibration on all data, using its own clock as a reference to unify the data timestamp, and controls the synchronization error within 1 millisecond. At the same time, spatial calibration is performed to map the data collected by different sensors to a unified vehicle coordinate system to ensure data consistency.

[0058] In another embodiment, after initializing the data acquisition module, the electronic device connects to the pedal sensor, steering angle sensor, and acceleration sensor via a dedicated interface. When acquiring vehicle status parameters, the accelerator pedal travel is first filtered to remove instantaneous fluctuations and retain only valid travel changes. During brake pedal travel acquisition, a threshold is set to filter out minor accidental touch signals, recording only the travel corresponding to valid braking operations. When acquiring steering wheel angle, the angle value is read at a frequency of 50Hz, and the rate of change of angle is calculated. Longitudinal and lateral acceleration acquisition utilizes a sliding window algorithm to smooth the data curve and reduce measurement noise. For driving environment data acquisition, the electronic device controls the vision camera to focus on the road surface area, prioritizing the acquisition of key information such as road texture and markings. The lidar uses a layered scanning method to focus on acquiring road elevation data and obstacle outlines. The millimeter-wave radar adjusts its detection angle to cover the area within the vehicle's driving trajectory, outputting target distance and speed information in real time. After acquisition, the electronic device categorizes and stores the data: vehicle status parameters are stored in a real-time database, and driving environment data is stored in a dedicated image and point cloud storage module for easy retrieval during subsequent processing.

[0059] Step 202: Determine the driving style based on vehicle state parameters, and determine the road surface prediction information for the vehicle to reach different road surface areas based on driving environment data.

[0060] Among these, driving style reflects the typified results of the driver's driving preferences and operating habits, and is a key basis for personalized adjustment of suspension performance. For example, driving style can be comfort, sport, or normal.

[0061] Road surface preview information is predictive information related to the road surface ahead, obtained through data processing and analysis based on driving environment data. Its function is to allow the active suspension to adapt to the road conditions ahead in advance. Specifically, it includes road surface feature information and time preview information. Road surface feature information covers road surface type, road surface unevenness, road surface obstacles, and road surface elevation. Time preview information combines the vehicle's current speed and driving direction to predict the estimated time for the vehicle to reach different road surface feature areas ahead.

[0062] In this embodiment, the electronic device determines the driving style based on vehicle state parameters and determines road surface prediction information for different road surface areas based on driving environment data. For example, features are first extracted from the collected vehicle state parameters, a 5-second time window is selected, and the maximum rate of change of the accelerator pedal travel within the window is calculated to reflect the aggressiveness of the driver's acceleration operation; at the same time, the maximum rate of change of the brake pedal travel is calculated to reflect the decisiveness of the deceleration operation; the number of times the absolute value of the steering wheel angle is greater than 10° within the window is counted to obtain the steering wheel angle frequency, which measures the frequency of steering operations; in addition, the average value of longitudinal acceleration and the variance of lateral acceleration are calculated to comprehensively reflect the stability of vehicle driving. The extracted features are input into a preset machine learning model. The model performs weight allocation and classification operations on the features. If the rate of change of the accelerator and brake pedals is less than 0.2% / ms, the steering wheel angle frequency is less than 0.5 times / second, the mean longitudinal acceleration is less than 0.8m / s², and the variance of lateral acceleration is less than 0.3(m / s²)², the driving style is determined to be comfort. If the values ​​of all features are higher than the preset high threshold, the driving style is determined to be sport. All other cases are determined to be normal.

[0063] In another embodiment, the electronic device preprocesses vehicle state parameters, removes outliers, and extracts features in 1-second increments. It calculates the cumulative change in accelerator pedal travel over each time interval to assess the total magnitude of acceleration; calculates the ratio of brake pedal travel to braking time to reflect braking intensity; determines the magnitude of steering operation by the maximum deviation of the steering wheel angle; and analyzes the difference between peak and trough values ​​of longitudinal acceleration and the fluctuation range of lateral acceleration to obtain dynamic characteristics of the vehicle. These features are input into a trained classification model. The model compares the similarity of the features with preset style templates. If the similarity with the comfort template exceeds 80%, it is identified as a comfort driving style; if the similarity with the sport template exceeds 80%, it is identified as a sport driving style; if the similarity with both templates is less than 60%, it is classified as a normal driving style. The model also outputs the confidence level of the style determination to ensure the reliability of the results.

[0064] Step 203: Determine the target control strategy for the vehicle's active suspension based on driving style, vehicle status parameters, and road surface preview information.

[0065] Among them, the target control strategy is a specific parameter scheme for the active suspension system, determined through algorithm optimization by combining driving style, vehicle state parameters, and road surface preview information. It guides the adjustment of suspension performance and serves as the core bridge connecting perceived information and executed actions. The content mainly includes key control parameters of the active suspension, such as suspension stiffness coefficient, suspension damping coefficient, and suspension height adjustment. Parameter values ​​must take into account driving style preferences, real-time road condition adaptability, and vehicle dynamic performance requirements.

[0066] The active suspension actuator is a hardware component used to receive commands from the target control strategy and adjust the parameters of the active suspension. It is the execution vehicle for the active suspension system to perform its functions. It typically includes solenoid valves, adjusting motors, hydraulic pumps, etc., and can adjust the damping, stiffness, and height of the suspension according to control commands. Through coordinated actions, the suspension performance reaches the state set by the target control strategy.

[0067] In this embodiment, the electronic device determines the target control strategy for the vehicle's active suspension based on driving style, vehicle state parameters, and road surface preview information. First, it integrates the driving style, vehicle state parameters, and road surface preview information. If the current driving style is comfort, the vehicle state parameters are: accelerator pedal travel 30%, brake pedal travel 0%, longitudinal acceleration 0.6 m / s², and lateral acceleration 0.3 m / s². The road surface preview information is: 0.6 seconds to a slightly uneven asphalt road ahead, and 1.2 seconds to a pothole area ahead. Next, it calls a pre-stored association mapping table and matches the suspension stiffness reference value of 20 kN / m, damping coefficient of 800 N·s / m, and height adjustment reference amount of 0 mm according to the comfort driving style. Considering the slightly uneven road surface, the stiffness is adjusted down by 10% and the damping down by 15% based on the reference values. For pothole areas, the height is adjusted up by 15 mm to obtain the initial suspension parameters. Then, the initial parameters were optimized using a model predictive control algorithm. The goal was to minimize the vertical acceleration fluctuation of the vehicle, the body roll angle, and the tire ground pressure uniformity. A dynamic model was constructed based on the current vehicle state parameters. The model was then solved by rolling time-domain optimization to finally obtain the target control strategy with a suspension stiffness of 17 kN / m, a damping coefficient of 650 N·s / m, and a height adjustment of 18 mm.

[0068] In another embodiment, the electronic device first prioritizes driving style, vehicle status parameters, and road surface preview information. The driving style determines the overall trend of the suspension parameters, the road surface preview information determines the adjustment range of the parameters, and the vehicle status parameters serve as the basis for dynamic correction. If the current driving style is sporty, the vehicle status parameters show a longitudinal acceleration of 1.5 m / s² and a lateral acceleration of 0.8 m / s², and the road surface preview information indicates a 0.8-second wait for a smooth concrete road and a 1.5-second wait for a curve. The system matches the corresponding stiffness reference value of 28 kN / m, damping coefficient of 1000 N·s / m, and height adjustment reference of -5 mm (lowering the vehicle body to improve handling) from the association mapping table. No stiffness or damping adjustment is needed for smooth concrete roads, but the damping needs to be increased by 10% in curves to suppress body roll. Combining the current high longitudinal and lateral acceleration, the stiffness is further increased by 5% to obtain the initial parameters. Subsequently, by using a model predictive control algorithm, the influence of centrifugal force on the vehicle in curves is considered, and the stiffness and damping distribution are optimized to ensure that the body roll angle is less than 3°. Finally, the target control strategy of suspension stiffness of 30kN / m, damping coefficient of 1100N・s / m, and height adjustment of -5mm is determined.

[0069] Step 204: Control the action of the active suspension actuator according to the target control strategy.

[0070] In this embodiment, the electronic device controls the active suspension actuator according to the target control strategy. First, the suspension parameters in the target control strategy are converted into actuator control signals. If the target stiffness is 17kN / m, the pre-stored "stiffness-motor speed" correspondence is consulted to determine that the motor needs to output a speed of 800rpm. If the target damping is 650N・s / m, the solenoid valve drive current is determined to be 1.2A according to the "damping-solenoid valve current" mapping table. If the target height adjustment is 18mm, the hydraulic pump working pressure is set to 1.8MPa according to the "height-hydraulic pump pressure" lookup table. Next, the electronic device synchronously transmits the control signals to the adjustment motor, solenoid valve, and hydraulic pump via the vehicle bus. The control timing is as follows: the hydraulic pump is started first, with pressurization beginning 0.3 seconds in advance to ensure height adjustment is completed when the vehicle reaches a bumpy area; after a delay of 0.1 seconds, the adjustment motor and solenoid valve are started synchronously to adjust stiffness and damping to adapt to slightly uneven road surfaces. During execution, the actual parameters are collected in real time by the suspension status feedback sensor. If the actual height is found to be only 16mm, which is 2mm away from the target value and exceeds the preset threshold of ±1mm, the electronic equipment will immediately increase the hydraulic pump pressure to 1.9MPa until the actual height reaches 18mm, ensuring that the suspension performance meets the control requirements.

[0071] In another embodiment, after receiving the target control strategy, the electronic device first verifies the rationality of the parameters, confirming that the suspension stiffness of 30kN / m, damping coefficient of 1100N・s / m, and height adjustment of -5mm are within the operating range of the actuator. Then, the parameters are converted into control signals: stiffness of 30kN / m corresponds to a regulating motor speed of 1200rpm, damping of 1100N・s / m corresponds to a solenoid valve current of 1.8A, and height adjustment of -5mm corresponds to a hydraulic pump pressure of 1.2MPa (pressure relief lowers the vehicle body). The electronic device sends signals through a dedicated control interface, first controlling the hydraulic pump to depressurize, slowly lowering the vehicle body height to avoid sudden height changes affecting driving stability; simultaneously, the regulating motor operates at the set speed, gradually increasing the suspension stiffness; and the solenoid valve synchronously adjusts the current to increase damping. During the process, the feedback sensor transmits data in real time. If the actual damping value is detected to be 1050 N·s / m, which deviates from the target value by 50 N·s / m, the electronic device will increase the current of the solenoid valve to 1.9A and continue to correct until the damping reaches 1100 N·s / m, ensuring that the active suspension moves accurately according to the target control strategy.

[0072] In the aforementioned active suspension control method based on driving style recognition and road surface prediction, high-frequency acquisition of vehicle state parameters and multi-sensor fusion of driving environment data provide the suspension system with comprehensive and real-time perception data. Based on this data, a driving style that aligns with the driver's preferences and forward-looking road surface prediction information are determined, clarifying the direction and lead time for suspension adaptation. Subsequently, a target control strategy is formulated by integrating driving style, vehicle state parameters, and road surface prediction information, forming a suspension parameter scheme that accurately matches multi-dimensional needs. Finally, according to the target control strategy, the active suspension actuator is controlled. This method forms a closed loop from perception and decision-making to execution, achieving flexible adaptation to different driving preferences, vehicle dynamics, and road conditions, significantly improving the adaptability of the vehicle suspension system.

[0073] In one embodiment, such as Figure 3 As shown, the aforementioned "determining the target control strategy for the vehicle's active suspension based on driving style, vehicle state parameters, and road surface preview information" includes:

[0074] Step 301: Determine the initial control strategy based on driving style.

[0075] In this embodiment of the application, the electronic device determines the initial control strategy based on the driving style.

[0076] For example, the identified driving style result "Comfort" is first obtained. This result is derived from the analysis of previous vehicle state parameters, specifically including the accelerator pedal change rate of 0.15% / ms and the steering wheel angle frequency of 0.4 times / s. Then, the pre-stored "Driving Style - Initial Suspension Parameters" mapping table is called. The table clearly shows that the core parameter benchmarks corresponding to the Comfort driving style are a suspension stiffness benchmark value of 22kN / m, a damping coefficient benchmark value of 750N・s / m, and a height adjustment benchmark amount of 0mm. The stiffness parameter focuses on reducing vibration transmission, the damping parameter focuses on filtering bumps, and the height parameter maintains a normal vehicle height to balance comfort and passability. Next, the benchmark parameters are checked for reasonableness, confirming that the stiffness of 22kN / m and the damping of 750N・s / m are both within the working range of the active suspension actuator. The actuator's stiffness working range is 15-35kN / m, and its damping working range is 500-1200N・s / m; no parameter exceeds the limit. Finally, the verified parameters are integrated into an initial control strategy, specifically "suspension stiffness 22kN / m, damping coefficient 750N・s / m, height adjustment 0mm", and stored in a temporary parameter library for later use.

[0077] In another embodiment, the electronic device receives a "sporty" result from the driving style recognition module. This result is derived from the driver's continuous aggressive operations, with corresponding operating parameters including accelerator pedal change rate of 0.6% / ms, brake pedal change rate of 0.55% / ms, and lateral acceleration variance of 0.9 (m / s²)². Then, a preset mapping rule is retrieved. For a sporty driving style, prioritizing handling stability, the initial parameter benchmarks are: suspension stiffness benchmark of 30 kN / m, damping coefficient benchmark of 1000 N·s / m, and height adjustment benchmark of -8 mm. The stiffness parameter enhances vehicle support, the damping parameter suppresses body roll and pitch, and the height parameter reduces the impact of centrifugal force during cornering by lowering the vehicle's center of gravity. Then, the parameters are adjusted based on the vehicle's current load. The vehicle load is estimated to be 500kg based on wheel speed sensor data, which is a light load condition. Under light load, the body support requirements are reduced, so the stiffness is reduced by 5% from the baseline value to 28.5kN / m, the damping coefficient remains unchanged at 1000N・s / m, and the height adjustment remains at -8mm, as the lower center of gravity requirement is not affected by the light load. Finally, the initial adjustment strategy is generated as "suspension stiffness 28.5kN / m, damping coefficient 1000N・s / m, height adjustment -8mm", and marked with the "Sporty - Light Load" scenario for subsequent use.

[0078] Step 302: Adjust the initial control strategy based on road surface preview information and vehicle state parameters to obtain the target control strategy for the vehicle's active suspension.

[0079] In this embodiment, the electronic device adjusts the initial control strategy based on road surface preview information and vehicle state parameters to obtain a target control strategy for the vehicle's active suspension.

[0080] Taking an initial control strategy based on a comfort baseline (stiffness 22kN / m, damping 750N・s / m, height 0mm) as an example, the road surface preview information and current vehicle status parameters are first obtained. The road surface preview information indicates that the vehicle will reach a "slightly uneven gravel road area" 1.2 seconds ahead. The standard deviation of the road surface unevenness in this area is 0.12m, and there are no obstacles. The vehicle status parameters are a longitudinal acceleration of 0.4m / s², indicating a constant speed, and a steering wheel angle of 0°, indicating a straight-line driving state. Next, the requirements of the road surface preview information on the suspension are analyzed. The slight unevenness of the gravel road requires enhanced bump filtering ability, thus requiring a reduction in stiffness and damping. The absence of obstacles eliminates the need to adjust the height. The adjustment range was then determined based on vehicle state parameters. Since the vehicle was traveling at a constant speed in a straight line with no steering or acceleration requirements, the stiffness was reduced by 15% from the initial 22 kN / m to 18.7 kN / m, and the damping was reduced by 20% from 750 N·s / m to 600 N·s / m, while the height remained unchanged at 0 mm. The adjustment effect was then verified using a model predictive control algorithm. The adjusted parameters were input into the vehicle dynamics model, and the simulation results showed that the vertical acceleration fluctuation when traversing a gravel road was 0.18 m / s², lower than the comfort threshold of 0.2 m / s², indicating stable vehicle posture. Finally, the target control strategy was determined to be "suspension stiffness 18.7 kN / m, damping coefficient 600 N·s / m, height adjustment 0 mm".

[0081] In one embodiment, such as Figure 4 As shown, the above-mentioned "adjusting the initial control strategy based on road surface preview information and vehicle state parameters to obtain a target control strategy for the vehicle's active suspension" includes:

[0082] Step 401: The road surface prediction information is processed in a hierarchical manner to classify the road surface prediction information of the vehicle approaching the road surface area into multiple prediction levels.

[0083] In this embodiment of the application, the electronic device performs hierarchical processing on the road surface prediction information to classify the road surface prediction information of the vehicle approaching the road surface area into multiple prediction levels.

[0084] In one specific embodiment, the electronic device first acquires processed road surface preview information, including the location of different road surface areas ahead, road surface characteristics (such as smooth roads, slightly uneven roads, heavily uneven roads, and obstacle areas), and the estimated time for the vehicle to reach each area. Then, a classification rule is set: the estimated time for the vehicle to reach the road surface area ahead is the core classification criterion, combined with the complexity of the road surface characteristics for auxiliary adjustment—an estimated time ≤ 0.5 seconds and road surface characteristics of heavily uneven roads or obstacles are classified as emergency preview level, where the suspension system needs to respond quickly to sudden road conditions; an estimated time between 0.5 and 1.5 seconds and road surface characteristics of slightly uneven roads or regular curves are classified as normal preview level, where the suspension system needs to adjust smoothly to adapt to changes in road conditions; an estimated time > 1.5 seconds and road surface characteristics of smooth roads or no special road conditions are classified as long-term preview level, where the suspension system can adjust slowly or maintain the current parameters. Next, the specific road surface advance information is classified into levels: "Pothole / obstacle area" (estimated time 0.3-0.5 seconds, including obstacles) is categorized as emergency advance; "Slightly uneven asphalt road area" (estimated time 0.5-1.5 seconds) is categorized as normal advance; and "Smooth cement road area" is categorized as categorized as expedited advance. (estimated time 2.0-1.5 seconds) is categorized as smooth road. Finally, the classification results are associated with the corresponding road surface characteristics and estimated time, providing a basis for subsequent adjustments to control strategies.

[0085] Step 402: For each pre-aiming level, adjust the initial control strategy according to the vehicle state parameters corresponding to the pre-aiming level to obtain the target control strategy.

[0086] In this embodiment, for each pre-aiming level, the electronic device adjusts the initial control strategy according to the vehicle state parameters corresponding to the pre-aiming level to obtain the target control strategy.

[0087] In one specific embodiment, taking "initial adjustment strategy as comfort benchmark (suspension stiffness 22kN / m, damping coefficient 750N・s / m, height adjustment 0mm)" as an example, adjustments are made for different pre-aiming levels.

[0088] For the Emergency Pre-aiming Level (reaching the pothole / obstacle area 0.3 seconds ahead), the corresponding vehicle status parameters are first obtained: longitudinal acceleration 0.5 m / s² (constant speed driving), steering wheel angle 0° (straight-line driving). These parameters indicate that the vehicle has no steering requirement and should prioritize dealing with pothole impacts. Since the Emergency Pre-aiming Level needs to quickly improve passability, the adjustments are set as follows: height is increased by 20 mm from the initial 0 mm to avoid collisions; stiffness is temporarily increased by 10% to 24.2 kN / m to enhance vehicle body support (to cope with pothole impacts); and damping remains unchanged at 750 N·s / m (while also considering impact filtering).

[0089] For the standard anti-collision level (reaching a slightly uneven area of ​​asphalt road 1.2 seconds ahead), the corresponding vehicle state parameters were obtained: longitudinal acceleration 0.4 m / s² (constant speed driving), steering wheel angle 5° (small steering). These parameters indicate that for small steering, a balance needs to be struck between bump filtering and steering stability. The adjustment range was set as follows: stiffness was reduced by 15% from the initial 22 kN / m to 18.7 kN / m (enhancing bump filtering), damping was reduced by 20% to 600 N·s / m (improving comfort), and height was kept at 0 mm (no height adjustment required). At the same time, due to the small steering, an additional 5% (630 N·s / m) of damping was retained on the steering side to suppress body roll.

[0090] For the long-term advance targeting level (reaching a smooth concrete road area 2.0 seconds ahead), the corresponding vehicle state parameters are obtained: longitudinal acceleration of 0.3 m / s² (constant speed driving) and steering wheel angle of 0° (straight driving). These parameters indicate that the vehicle is driving smoothly and there are no special road conditions required. The adjustment range is set as follows: stiffness, damping, and height are all maintained at the initial control strategy parameters (stiffness 22 kN / m, damping 750 N・s / m, height 0 mm), and the parameters are only adjusted slowly through iteration (fine-tuning stiffness ±1% every 0.5 seconds) to ensure that the parameters are adapted to the smooth road surface.

[0091] Finally, the adjustment results of each aiming level were integrated, and a target control strategy was formed according to the expected time sequence: "0-0.7 seconds (before reaching the conventional aiming level area): stiffness 18.7kN / m (steering side damping 630N·s / m, non-steering side 600N·s / m), height 0mm; 0.7-0.9 seconds (transition from conventional aiming level to emergency aiming level): stiffness 18.7kN / m→24.2kN / m, damping 600N·s / m." / m→750N・s / m, height 0mm→20mm; 0.9-1.0 seconds (emergency pre-aiming level area): stiffness 24.2kN / m, damping 750N・s / m, height 20mm; after 1.0 second (from the end of the emergency pre-aiming level to the long-term pre-aiming level): stiffness 24.2kN / m→22kN / m, damping 750N・s / m→750N・s / m, height 20mm→0mm, then maintain the initial parameters.

[0092] In one embodiment, such as Figure 5 As shown, the above "determining driving style based on vehicle state parameters" includes:

[0093] Step 501: Perform feature extraction processing on the vehicle state parameters to obtain multiple state features.

[0094] In this embodiment of the application, the electronic device performs feature extraction processing on the vehicle state parameters to obtain multiple state features.

[0095] In one specific embodiment, the electronic device first acquires real-time vehicle state parameters, including accelerator pedal travel, brake pedal travel, steering wheel angle, longitudinal acceleration, and lateral acceleration data over a continuous 5-second period. The acquisition frequency is 100Hz, resulting in 500 sets of time-series data. The accelerator pedal travel data is then processed to calculate the change in travel every 100ms. The maximum value within 5 seconds is taken as the "accelerator pedal travel change rate," reflecting the aggressiveness of the driver's acceleration operation. The brake pedal travel data is processed using the same method to calculate the "brake pedal travel change rate," reflecting the decisiveness of the deceleration operation. For the steering wheel angle data, the number of times the absolute value of the angle is ≥10° within 5 seconds is counted, and this number is divided by the time length to obtain the "steering wheel angle frequency," measuring the frequency of steering operations. Simultaneously, the maximum deviation of the angle within 5 seconds is calculated as the "steering wheel angle amplitude," reflecting the force of the steering operation. For longitudinal acceleration data, the average value over 5 seconds is calculated to obtain the "mean longitudinal acceleration," which reflects the overall smoothness of vehicle acceleration or deceleration. For lateral acceleration data, the variance over 5 seconds is calculated to obtain the "variance lateral acceleration," which reflects the fluctuation of the vehicle's lateral posture during cornering. Finally, the extracted "accelerator pedal travel rate of change, brake pedal travel rate of change, steering wheel angle frequency, steering wheel angle amplitude, mean longitudinal acceleration, and variance lateral acceleration" are integrated into six state features to form a feature dataset for subsequent classification.

[0096] Step 502: Input the extracted multiple state features into the preset feature classification model. The feature filtering layer of the feature classification model sorts and filters the multiple state features by importance. The classification layer of the feature classification model is used to classify the filtered features to obtain the driving style.

[0097] In this embodiment, the electronic device inputs multiple extracted state features into a preset feature classification model. The feature filtering layer of the feature classification model sorts and filters the multiple state features by importance. The classification layer of the feature classification model is used to classify the filtered features to obtain the driving style.

[0098] In one specific embodiment, the electronic device first invokes a preset feature classification model, which consists of a feature selection layer and a classification layer. The feature selection layer uses a random forest algorithm, and the classification layer uses a support vector machine algorithm. This model has been trained using 1000 sets of driver operation data. The extracted six state features (accelerator pedal travel rate of change 0.18% / ms, brake pedal travel rate of change 0.15% / ms, steering wheel angle frequency 0.4 times / s, steering wheel angle amplitude 15°, mean longitudinal acceleration 0.6 m / s², and variance of lateral acceleration 0.3 (m / s²)²) are input into the model. First, through a feature selection layer, the random forest algorithm ranks the features by calculating their contribution to the classification results. The results show that "accelerator pedal travel rate of change, steering wheel angle frequency, and mean longitudinal acceleration" have the highest importance scores (all > 0.2), followed by "brake pedal travel rate of change and lateral acceleration variance" (0.1-0.2), and "steering wheel angle amplitude" has the lowest (< 0.1). Based on a preset threshold (importance score ≥ 0.1), the top 5 features are selected, and "steering wheel angle amplitude" is removed to reduce redundant information. Next, the selected 5 features are input into the classification layer. The support vector machine algorithm compares these features with feature templates for various driving styles (comfort, sport, and normal) in the training set to calculate the matching degree: the matching degree between this feature set and the comfort template is 89%, the matching degree with the sport template is 32%, and the matching degree with the normal template is 65%. Based on the principle of highest matching degree, the current driving style is determined to be "comfort". At the same time, the classification confidence level of 89% is output to ensure the reliability of the result. Finally, "comfort" is used as the driving style identification result for subsequent suspension adjustment strategy formulation.

[0099] In one embodiment, such as Figure 6 As shown, the above-mentioned "determining road surface preview information based on driving environment data for vehicles to reach different road surface areas" includes:

[0100] Step 601: Perform fusion processing and error elimination processing on the driving environment data in sequence to obtain the processed data.

[0101] In this embodiment, the electronic device sequentially performs fusion processing and error elimination processing on the driving environment data to obtain processed data.

[0102] In one specific embodiment, the electronic device first collects driving environment data from multiple sensors, including an RGB image of the road surface ahead output by a visual camera, three-dimensional point cloud data generated by a lidar, and target distance and relative speed data acquired by a millimeter-wave radar. First, a fusion process is performed: using the visual camera image as a reference, feature point matching (extracting road surface edges and corner points of marker lines in the image as feature points) is used to map the lidar point cloud data to the image pixel coordinate system, ensuring that the spatial position of the point cloud accurately corresponds to the road surface area in the image; simultaneously, the obstacle distance data detected by the millimeter-wave radar is associated with the three-dimensional coordinates of the obstacles in the lidar point cloud to supplement the obstacle's speed information, forming a preliminary dataset after multi-source data fusion. Next, error elimination processing is performed: the Kalman filter algorithm is used, with the vehicle dynamics model (built based on the current vehicle speed and steering angle) as the state equation. Anomalies in the fused data (such as isolated point clouds caused by rain and snow interference from lidar and static targets misidentified by millimeter-wave radar) are treated as observation noise, and data deviations are corrected through iterative calculation. For areas with uneven illumination in the visual image, the histogram equalization algorithm is used to eliminate brightness fluctuation errors. Finally, the processed data with no obvious noise and spatial and temporal synchronization is obtained, which includes the fused road surface image, the corrected point cloud coordinates, and the dynamic and static status information of obstacles.

[0103] Step 602: Perform image segmentation and feature recognition processing on the processed data in sequence to obtain a road surface feature map of the driving road surface condition.

[0104] In this embodiment, the electronic device sequentially performs image segmentation and feature recognition processing on the processed data to obtain a road feature map of the driving road surface conditions.

[0105] In one specific embodiment, the electronic device first calls a semantic segmentation model to segment the road surface image in the processed data. The model extracts the road surface texture, color, and edge features in the image through an encoder, and then classifies the image pixels into categories such as "asphalt road", "cement road", "gravel road", "dirt road", "obstacles", and "marking lines" through a decoder, generating a pixel-level road surface segmentation map, in which different categories are marked with different colors (dark gray for asphalt road, light gray for cement road, beige for gravel road, and red for obstacles). The following feature recognition processing is then performed: For the segmented road surface areas, the area proportion of each road surface type is calculated (e.g., asphalt road accounts for 90%, gravel road accounts for 10%), and the road surface unevenness parameters are extracted by combining the point cloud data corrected by LiDAR (a 10m×10m sampling area is selected in the segmentation image, and the standard deviation of the point cloud elevation in this area is calculated; a standard deviation of 0.12m corresponds to slight unevenness); For the segmented obstacle areas, the location of the obstacles is determined by the bounding box detection algorithm (with the vehicle's current position as the origin, the longitudinal distance of the obstacle center point is 25m, and the lateral distance is 1.5m) and dimensions (length 1.2m, width 0.8m, height 0.15m). Finally, the segmentation category, unevenness parameters, and obstacle information are superimposed on the original road surface image to generate a road surface feature map containing detailed road condition annotations. The feature map also marks the coordinate range of each road surface area to facilitate the subsequent calculation of vehicle arrival time.

[0106] Step 603: Based on the road surface feature map and vehicle status parameters, determine the road surface pre-aiming information for the vehicle to reach different road surface areas.

[0107] In this embodiment, the electronic device determines the road surface pre-aiming information for the vehicle to reach different road surface areas based on the road surface feature map and vehicle state parameters.

[0108] In one specific embodiment, the electronic device first extracts the coordinate information of key road surface areas from the road surface feature map, including "slightly uneven asphalt road area" (longitudinal distance 15-25m, lateral distance -2 to 2m), "gravel road area" (longitudinal distance 25-35m, lateral distance -2 to 2m), and "obstacle area" (longitudinal distance 25m, lateral distance 1.5m, dimensions 1.2m × 0.8m × 0.15m). Next, it acquires the current vehicle status parameters, including vehicle speed (converted from wheel speed sensor data to 60km / h, i.e., 16.67m / s), driving direction (steering wheel angle 0°, determined to be straight driving), and longitudinal acceleration (0.4m / s², determined to be constant speed driving). The estimated time for the vehicle to reach each road surface area was then calculated: For the "slightly uneven asphalt road area," the initial longitudinal distance of 15m was taken, and the time = 15m ÷ 16.67m / s ≈ 0.9s; for the "gravel road area," the initial longitudinal distance was taken, and the time = 25m ÷ 16.67m / s ≈ 1.5s; for the "obstacle area," the longitudinal distance from the center point was taken, and the time was the same as for the gravel road area, 1.5s. Simultaneously, the lateral overlap between the vehicle's trajectory and each road surface area was confirmed by combining the vehicle's driving direction parameters (when driving in a straight line, the trajectory completely overlaps with the lateral range of each area, and no direction adjustment is needed), and the speed change trend was determined based on the longitudinal acceleration (when driving at a constant speed, no correction to the estimated time is needed). Finally, the categories of each road surface area (slightly uneven asphalt road, gravel road, obstacles), estimated arrival time (0.9s, 1.5s, 1.5s), location coordinates (longitudinal 15-25m, 25-35m, 25m; lateral -2 to 2m, -2 to 2m, 1.5m) and size information (obstacle 1.2m×0.8m×0.15m) are integrated to form complete road surface pre-aiming information for vehicles arriving at different road surface areas, which is stored in the control strategy database for subsequent retrieval.

[0109] In one embodiment, such as Figure 7 As shown, the above-mentioned "performing sequential fusion processing and error elimination processing on driving environment data to obtain processed data" includes:

[0110] Step 701: Perform data fusion processing on the driving environment data to obtain fused driving environment data.

[0111] In this embodiment of the application, the electronic device performs data fusion processing on the driving environment data to obtain fused driving environment data.

[0112] In one specific embodiment, the electronic device first collects raw driving environment data from three types of sensors: a visual camera, a lidar, and a millimeter-wave radar. The visual camera outputs an RGB road surface image (1920×1080 resolution, 30fps) within a 100m range ahead, including road texture, color, and obstacle outline information. The lidar generates three-dimensional point cloud data (128 lines density, 10Hz frame rate) within a 200m range ahead, which can accurately reflect road undulations, obstacle height, and spatial coordinates. The millimeter-wave radar acquires distance and relative speed data of targets within a 80m range ahead (detection angle ±60°, ranging accuracy ±0.1m, 20Hz frame rate), which can effectively identify dynamic obstacles.

[0113] The data fusion process is then initiated: using the visual camera image as a spatial reference, the SIFT feature extraction algorithm is used to select key feature points such as road edges and traffic sign corners from the image (50-80 stable feature points are extracted in total). The coordinates of the LiDAR point cloud data and these feature points are then matched. The spatial transformation matrix is ​​calculated using the least squares method to transform the point cloud data from the LiDAR coordinate system to the image pixel coordinate system, thereby achieving spatial alignment between the point cloud and the image. This allows the road elevation information marked in the point cloud to be accurately superimposed onto the corresponding area of ​​the image.

[0114] Simultaneously, the distance and velocity data of obstacles detected by millimeter-wave radar are correlated with the three-dimensional coordinates of obstacles identified in the lidar point cloud. This is achieved by calculating the positional deviation between the millimeter-wave radar target and the lidar obstacle (with a set deviation threshold of ≤0.3m), matching multi-source information of the same obstacle, and supplementing the obstacle's dynamic velocity parameters (e.g., if the speed of an obstacle 25m ahead relative to the vehicle is 0m / s, it is determined to be a static obstacle). Finally, the aligned image, point cloud data, and correlated obstacle dynamic and static information are integrated to form fused driving environment data. This data simultaneously includes visual texture, spatial three-dimensional morphology, and target dynamic attributes, providing complete information support for subsequent processing.

[0115] Step 702: Use the Kalman filter algorithm or particle filter algorithm to perform error elimination processing on the fused driving environment data to obtain the processed data.

[0116] In this embodiment, the computing device uses a Kalman filter algorithm or a particle filter algorithm to perform error elimination processing on the fused driving environment data to obtain processed data.

[0117] In one specific embodiment, the electronic device first constructs the state equation and observation equation of the Kalman filter model: based on the vehicle dynamics model, the state equation is established by combining the current vehicle state parameters (vehicle speed 60km / h, steering wheel angle 0°), and the road surface elevation and obstacle position in the fused data are used as state variables (the state vector dimension is set to 6, including the elevation of the road surface sampling point and the X / Y / Z coordinates of the obstacle). The theoretical value of the state variable at the next moment is predicted by the vehicle speed (e.g., based on a vehicle speed of 16.67m / s, it is predicted that the relative distance between the vehicle and the obstacle will decrease by 1.67m after 0.1 seconds).

[0118] The fused driving environment data was used as the input of the observation values ​​into the filtering model to analyze the sources of error in the observation values: isolated point clouds caused by rain and snow interference from lidar (elevation deviation from surrounding point clouds > 0.2m), local brightness anomalies in visual images caused by uneven illumination (grayscale value fluctuation > 30), and false targets caused by multipath reflection from millimeter-wave radar (no corresponding lidar point cloud matching) were all identified as observation noise.

[0119] Initiate Kalman filter iterative calculation: First, calculate the residual between the predicted state value and the observed value (e.g., the predicted elevation of a road sampling point is 0.05m, the observed elevation is 0.28m, and the residual is 0.23m). Then, based on the preset process noise covariance matrix (based on sensor accuracy settings, the lidar elevation error variance is set to 0.01(m)²) and the observation noise covariance matrix (the isolated point cloud noise variance is set to 0.04(m)²), calculate the Kalman gain (the gain value for this sampling point is 0.22). Correct the predicted state value through gain weighting to obtain the optimal estimated value after filtering (the corrected elevation of this sampling point is 0.09m).

[0120] To address illumination errors in visual images, a histogram equalization algorithm is further employed after filtering to stretch the image's grayscale range (expanding the original grayscale interval from 0-180 to 0-255) and eliminate local brightness fluctuations. For false targets from millimeter-wave radar, data on targets without corresponding LiDAR point cloud matching after filtering are directly removed. The final processed data after error elimination shows: the standard deviation of road elevation data decreases from 0.18m to 0.08m after fusion; obstacle position error is ≤0.1m; image brightness uniformity is improved by 40%; and data noise is significantly reduced, accurately reflecting the actual road surface conditions.

[0121] In one embodiment, such as Figure 8 As shown, the above-mentioned "controlling the action of the active suspension actuator according to the target control strategy" includes:

[0122] Step 801: Convert the adjusted suspension control parameters in the target control strategy into actuator control signals; the actuator control signals include at least the solenoid valve drive current, the motor speed adjustment, and the hydraulic pump working pressure.

[0123] In this embodiment, the electronic device converts the adjusted suspension control parameters in the target control strategy into actuator control signals.

[0124] In one specific embodiment, the electronic device first retrieves the adjusted suspension control parameters determined in the target control strategy, specifically "suspension stiffness 24kN / m, damping coefficient 820N・s / m, height adjustment amount 15mm", which need to be adapted to the front emergency pre-aiming pothole area. Then, the pre-stored "Suspension Control Parameters - Actuator Signals" mapping table is invoked. This table, calibrated through previous bench tests, establishes the correspondence between each parameter and the control signal: For suspension stiffness, the stiffness value is positively correlated with the speed of the regulating motor, with 24 kN / m corresponding to a regulating motor speed of 950 rpm (tests verify that at this speed, the motor-driven stiffness regulating mechanism can reach the target stiffness within 0.2 seconds); For damping coefficient, the damping value is linearly mapped to the solenoid valve drive current, with 820 N·s / m corresponding to a solenoid valve drive current of 1.4 A (the current controls the opening of the solenoid valve core; at 1.4 A, the valve core opening is 45%, matching the target damping); For height adjustment, a 15 mm upward adjustment requirement corresponds to a hydraulic pump working pressure of 1.9 MPa (the hydraulic pump pressure achieves height adjustment by pushing the piston rod to extend and retract; 1.9 MPa can drive the piston rod to extend 15 mm, and the pressure is stable without overshoot). During the conversion process, the electronic equipment verifies the validity of the signals: confirming that the regulating motor speed of 950 rpm is within its operating range (500-1500 rpm), the solenoid valve current of 1.4A is within the safe range (0.5-2.0A), and the hydraulic pump pressure of 1.9MPa does not exceed the rated pressure (2.5MPa), indicating no risk of parameter exceeding limits. Finally, the verified "solenoid valve drive current 1.4A, regulating motor speed 950 rpm, and hydraulic pump operating pressure 1.9MPa" are integrated into the actuator control signal, encapsulated into a standard control frame via the CAN bus protocol, and await synchronous transmission.

[0125] Step 802: Synchronously adjust the solenoid valve, regulating motor, and hydraulic pump of the active suspension according to the control signal of the actuator.

[0126] In this embodiment, the electronic device performs synchronous adjustment of the solenoid valve, regulating motor, and hydraulic pump of the active suspension according to the control signal of the actuator.

[0127] In one specific embodiment, the electronic device first determines the synchronous adjustment sequence: considering that the response speed of the hydraulic pump driving the height adjustment (0.15 seconds from start to reach the target pressure) is slower than that of the solenoid valve (0.05 seconds) and the adjustment motor (0.08 seconds), the sequence is set as "the hydraulic pump starts 0.07 seconds earlier, and the solenoid valve and adjustment motor start 0.07 seconds later", to ensure that the three complete the adjustment action simultaneously and avoid attitude fluctuations caused by asynchronous suspension parameters. Subsequently, the encapsulated control signals are sent to each actuator via the vehicle's CAN bus according to a set timing sequence: First, a pressure control command of 1.9MPa is sent to the hydraulic pump controller. After the hydraulic pump starts, the pressure value is fed back in real time through the pressure sensor. When the pressure rises to 1.7MPa (approaching 90% of the target value), the controller automatically reduces the pump speed to smoothly increase the pressure and prevent pressure overshoot. 0.07 seconds later, a drive current command of 1.4A is sent synchronously to the solenoid valve (current output is controlled by PWM signal with an accuracy of ±0.02A), and a speed command of 950rpm is sent to the regulating motor (the motor controller uses a vector control algorithm, and the speed fluctuation is controlled within ±10rpm). During the adjustment process, the electronic equipment collects feedback data in real time through suspension status feedback sensors: the stiffness sensor detects the process of the actual stiffness value increasing from the initial 22kN / m to 24kN / m, the damping sensor monitors the dynamic curve of the damping coefficient changing with the solenoid valve current, and the displacement sensor records the displacement amount from 0mm to 15mm. When the feedback data shows "stiffness 24kN / m, damping 820N・s / m, height 15mm" and remains stable for 30ms, the synchronization adjustment is considered complete. If a deviation occurs (such as the height only reaching 14mm), a pressure compensation command of 0.05MPa is added to the hydraulic pump until all parameters match the target values, ultimately achieving precise adjustment of the active suspension according to the target control strategy.

[0128] In one embodiment, such as Figure 9 As shown, the above method also includes:

[0129] Step 901: Collect the actual working parameters of the active suspension in real time and compare them with the adjusted suspension control parameters in the target control strategy.

[0130] In this embodiment of the application, the electronic device collects the actual working parameters of the active suspension in real time and compares them with the adjusted suspension control parameters in the target control strategy.

[0131] In one specific embodiment, the electronic device first activates the suspension state feedback sensor network, which includes a stiffness sensor, a damping sensor, and a displacement sensor. These sensors respectively collect the actual stiffness value, actual damping coefficient, and actual height value of the active suspension. The sampling frequency is set to 200Hz (higher than the actuator adjustment frequency to ensure dynamic tracking). The adjusted suspension control parameters in the target control strategy are "stiffness 24kN / m, damping 820N・s / m, height 15mm", with preset deviation thresholds of "stiffness ±5% (i.e., 22.8-25.2kN / m), damping ±8% (i.e., 754.4-885.6N・s / m), height ±3mm (i.e., 12-18mm)", used to determine whether the parameters meet the standards.

[0132] The sensors collect data in real time and transmit it to the electronic equipment via the CAN bus: the stiffness sensor detects the relationship between the deformation and force of the suspension spring and outputs the actual stiffness value (e.g., continuously collected values ​​of 23.2kN / m, 23.1kN / m, and 23.3kN / m); the damping sensor monitors the piston movement speed and resistance changes of the shock absorber and calculates the actual damping coefficient (e.g., 805N・s / m, 803N・s / m, and 806N・s / m); the displacement sensor measures the piston rod extension and retraction to obtain the actual height value (e.g., 13mm, 12.8mm, and 12.9mm).

[0133] The electronic equipment performs frame-by-frame comparison calculations between the collected actual parameters and target parameters: the deviation between the actual average stiffness of 23.2 kN / m and the target of 24 kN / m is -0.8 kN / m (deviation rate -3.3%, within ±5% threshold); the deviation between the actual average damping of 805 N·s / m and the target of 820 N·s / m is -15 N·s / m (deviation rate -1.8%, within ±8% threshold); the deviation between the actual average height of 12.9 mm and the target of 15 mm is -2.1 mm (deviation rate -14%, exceeding the ±3 mm threshold, i.e., exceeding the lower limit of the 12-18 mm range). After comparison, "height parameter deviation exceeds limit" is marked, triggering the subsequent compensation and correction process, while the deviation data of stiffness and damping parameters are recorded for traceability.

[0134] Step 902: If the deviation between the actual working parameters and the adjusted suspension control parameters exceeds a preset threshold, the actuator control signal is compensated and corrected to match the performance adjustment accuracy of the active suspension with the vehicle's dynamic driving state.

[0135] In this embodiment, when the deviation between the actual operating parameters and the adjusted suspension control parameters exceeds a preset threshold, the electronic device performs compensation and correction processing on the actuator control signal to match the performance adjustment accuracy of the active suspension with the vehicle's dynamic driving state.

[0136] In one specific embodiment, based on the previous comparison results, the actual height value (12.9mm) deviates from the target height value (15mm) by -2.1mm, exceeding the ±3mm threshold. The electronic device then initiates a compensation correction process. First, the cause of the deviation is analyzed: height adjustment is controlled by the hydraulic pump's working pressure. The current hydraulic pump control signal is 1.9MPa. The actual height did not reach the target, suggesting a slight leak in the hydraulic system causing pressure loss, which needs to be compensated for by increasing the pressure.

[0137] The mapping relationship between "height deviation and pressure compensation" was retrieved. This relationship was calibrated through bench testing. For every 1mm increase in deviation, an additional 0.08MPa of pressure is required. The current deviation is -3.2mm (the actual value is lower than the target value of 3.2mm). The required additional pressure compensation is calculated to be 3.2mm × 0.08MPa / mm = 0.256MPa. The corrected hydraulic pump working pressure control signal is the original 1.9MPa + 0.256MPa = 2.156MPa. At the same time, it was confirmed that this pressure does not exceed the rated pressure of the hydraulic pump of 2.5MPa, and there is no safety risk.

[0138] The compensation timing is adjusted synchronously based on the vehicle's dynamic driving status: The current vehicle status parameters are "vehicle speed 55km / h (constant speed), longitudinal acceleration 0.3m / s² (stable)", with no need for rapid acceleration or sharp turns. The compensation process is set as "step-by-step pressure increase" to avoid sudden pressure increases that could cause vehicle body vibration. The first stage is 0-0.05 seconds, increasing the pressure from 1.9MPa to 2.0MPa; the second stage is 0.05-0.1 seconds, increasing it to 2.1MPa; and the third stage is 0.1-0.15 seconds, increasing it to 2.156MPa, achieving pressure compensation in stages.

[0139] The corrected actuator control signals are: "Solenoid valve drive current 1.4A (unchanged, as damping deviation is within limits), regulating motor speed 950rpm (unchanged, as stiffness deviation is within limits), hydraulic pump working pressure 2.156MPa," which are retransmitted to each actuator via the CAN bus. After compensation, actual height values ​​are continuously collected. When the average actual height stabilizes at 14.9mm (deviation -0.1mm, within the ±3mm threshold) and remains stable for 50ms without fluctuation, compensation correction is stopped, confirming that the active suspension height adjustment accuracy meets the standard and its performance matches the vehicle's constant speed driving state.

[0140] In one embodiment, the above method further includes:

[0141] Step 1: Collect vehicle status parameters and driving environment data in real time during the vehicle's driving process.

[0142] Step 2: Perform feature extraction processing on the vehicle state parameters to obtain multiple state features.

[0143] Step 3: Input the extracted multiple state features into the preset feature classification model. The feature filtering layer of the feature classification model sorts and filters the multiple state features by importance. The classification layer of the feature classification model is used to classify the filtered features to obtain the driving style.

[0144] Step 4: Perform data fusion processing on the driving environment data to obtain the fused driving environment data.

[0145] Step 5: Use the Kalman filter algorithm or particle filter algorithm to perform error elimination processing on the fused driving environment data to obtain the processed data.

[0146] Step 6: Perform image segmentation and feature recognition processing on the processed data in sequence to obtain a road surface feature map of the driving road surface condition.

[0147] Step 7: Based on the road surface feature map and vehicle status parameters, determine the road surface pre-aiming information for the vehicle to reach different road surface areas.

[0148] Step 8: Classify the road surface prediction information into multiple prediction levels to determine the road surface prediction information for the area where the vehicle is about to reach.

[0149] Step 9: For each pre-aiming level, adjust the initial control strategy according to the vehicle state parameters corresponding to the pre-aiming level to obtain the target control strategy.

[0150] Step 10: Convert the adjusted suspension control parameters in the target control strategy into actuator control signals; the actuator control signals include at least the solenoid valve drive current, the motor speed adjustment, and the hydraulic pump working pressure.

[0151] Step 11: Synchronously adjust the solenoid valve, regulating motor, and hydraulic pump of the active suspension according to the control signal of the actuator.

[0152] Step 12: Collect the actual working parameters of the active suspension in real time and compare them with the adjusted suspension control parameters in the target control strategy.

[0153] Step 13: If the deviation between the actual working parameters and the adjusted suspension control parameters exceeds a preset threshold, the actuator control signal is compensated and corrected to match the performance adjustment accuracy of the active suspension with the vehicle's dynamic driving state.

[0154] It should be noted that this application does not limit the execution steps of steps 2-3 and steps 4-7. In actual execution, steps 4-7 can be executed first, followed by steps 2-3, or steps 2-3 and steps 4-7 can be executed in parallel.

[0155] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0156] Based on the same inventive concept, this application also provides an active suspension control device based on driving style recognition and road surface prediction for implementing the active suspension control method based on driving style recognition and road surface prediction described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the active suspension control device based on driving style recognition and road surface prediction provided below can be found in the limitations of the active suspension control method based on driving style recognition and road surface prediction described above, and will not be repeated here.

[0157] In one embodiment, such as Figure 10 As shown, an active suspension control device based on driving style recognition and road surface prediction is provided, including: a data acquisition module 1001, an information determination module 1002, a strategy determination module 1003, and a control module 1004, wherein:

[0158] The data acquisition module 1001 is used to collect vehicle status parameters and driving environment data in real time during the vehicle's driving process.

[0159] The information determination module 1002 is used to determine the driving style based on vehicle state parameters and to determine the road surface prediction information for the vehicle to reach different road surface areas based on driving environment data.

[0160] The strategy determination module 1003 is used to determine the target control strategy for the vehicle's active suspension based on driving style, vehicle state parameters and road surface preview information.

[0161] The control module 1004 is used to control the action of the active suspension actuator according to the target control strategy.

[0162] In one embodiment, the strategy determination module 1003 is specifically used to determine an initial control strategy based on driving style;

[0163] The initial control strategy is adjusted based on road surface preview information and vehicle state parameters to obtain a target control strategy for the vehicle's active suspension.

[0164] In one embodiment, the strategy determination module 1003 is specifically used to perform hierarchical processing on the road surface pre-aiming information to divide the road surface pre-aiming information of the vehicle reaching the road surface area ahead into multiple pre-aiming levels.

[0165] For each pre-aiming level, the initial control strategy is adjusted based on the vehicle state parameters corresponding to the pre-aiming level to obtain the target control strategy.

[0166] In one embodiment, the information determination module 1002 is specifically used to perform feature extraction processing on the vehicle state parameters to obtain multiple state features.

[0167] The extracted state features are input into a preset feature classification model. The feature filtering layer of the feature classification model sorts and filters the multiple state features by importance. The classification layer of the feature classification model is used to classify the filtered features to obtain the driving style.

[0168] In one embodiment, the information determination module 1002 is specifically used to perform fusion processing and error elimination processing on the driving environment data in sequence to obtain the processed data.

[0169] The processed data is then subjected to image segmentation and feature recognition processing to obtain a road surface feature map of the driving road surface conditions.

[0170] Based on the road surface feature map and vehicle status parameters, determine the road surface pre-aiming information for the vehicle to reach different road surface areas.

[0171] In one embodiment, the information determination module 1002 is specifically used to perform data fusion processing on the driving environment data to obtain fused driving environment data.

[0172] The Kalman filter algorithm or particle filter algorithm is used to perform error elimination processing on the fused driving environment data to obtain the processed data.

[0173] In one embodiment, the control module 1004 is specifically used to convert the adjusted suspension control parameters in the target control strategy into actuator control signals; the actuator control signals include at least the solenoid valve drive current, the motor speed adjustment, and the hydraulic pump working pressure.

[0174] The solenoid valves, regulating motors, and hydraulic pumps of the active suspension are synchronously adjusted according to the control signals from the actuators.

[0175] In one embodiment, the aforementioned active suspension control device based on driving style recognition and road surface prediction is also used to collect the actual working parameters of the active suspension in real time and compare them with the adjusted suspension control parameters in the target control strategy.

[0176] If the deviation between the actual working parameters and the adjusted suspension control parameters exceeds a preset threshold, the control signal of the actuator is compensated and corrected to match the performance adjustment accuracy of the active suspension with the dynamic driving state of the vehicle.

[0177] The modules in the aforementioned active suspension control device based on driving style recognition and road surface prediction can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the electronic device in hardware form or independently of it, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0178] In one embodiment, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0179] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0180] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0181] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0182] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0183] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0184] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An active suspension control method based on driving style recognition and road surface prediction, characterized in that, The method includes: Real-time collection of vehicle status parameters and driving environment data during vehicle operation; The driving style is determined based on the vehicle state parameters, and the road surface prediction information for the vehicle to reach different road surface areas is determined based on the driving environment data. Based on the driving style, the vehicle state parameters, and the road surface preview information, a target control strategy for the vehicle's active suspension is determined. The active suspension actuator is controlled to move according to the target control strategy.

2. The method according to claim 1, characterized in that, The step of determining a target control strategy for the vehicle's active suspension based on the driving style, vehicle state parameters, and road surface preview information includes: Determine the initial control strategy based on the driving style described; The initial control strategy is adjusted based on the road surface preview information and the vehicle state parameters to obtain a target control strategy for the vehicle's active suspension.

3. The method according to claim 2, characterized in that, The step of adjusting the initial control strategy based on the road surface preview information and the vehicle state parameters to obtain a target control strategy for the vehicle's active suspension includes: The road surface prediction information is processed in a hierarchical manner to classify the road surface prediction information of the vehicle approaching the road surface area into multiple prediction levels; For each of the aforementioned pre-aiming levels, the initial control strategy is adjusted based on the vehicle state parameters corresponding to the pre-aiming level to obtain the target control strategy.

4. The method according to claim 1, characterized in that, The process of determining the driving style based on the vehicle state parameters includes: The vehicle state parameters are subjected to feature extraction processing to obtain multiple state features; The extracted state features are input into a preset feature classification model. The feature filtering layer of the feature classification model sorts and filters the state features by importance. The classification layer of the feature classification model is used to classify the filtered features to obtain the driving style.

5. The method according to claim 1, characterized in that, The step of determining the road surface prediction information for the vehicle to reach different road surface areas based on the driving environment data includes: The driving environment data is sequentially fused and error-eliminating to obtain the processed data; The processed data is then subjected to image segmentation and feature recognition processing in sequence to obtain a road surface feature map of the driving road surface condition; Based on the road surface feature map and the vehicle state parameters, road surface pre-aiming information for the vehicle to reach different road surface areas is determined.

6. The method according to claim 5, characterized in that, The process of fusing and eliminating errors in the driving environment data sequentially yields processed data, including: The driving environment data is fused to obtain fused driving environment data; The fused driving environment data is processed by using a Kalman filter algorithm or a particle filter algorithm to eliminate errors, resulting in the processed data.

7. The method according to claim 1, characterized in that, The step of controlling the action of the active suspension actuator according to the target control strategy includes: The adjusted suspension control parameters in the target control strategy are converted into actuator control signals; the actuator control signals include at least the solenoid valve drive current, the regulating motor speed, and the hydraulic pump working pressure. The solenoid valves, regulating motors, and hydraulic pumps of the active suspension are synchronously adjusted according to the control signals of the actuators.

8. The method according to claim 7, characterized in that, The method further includes: The actual operating parameters of the active suspension are collected in real time and compared with the adjusted suspension control parameters in the target control strategy. If the deviation between the actual operating parameters and the adjusted suspension control parameters exceeds a preset threshold, the actuator control signal is compensated and corrected to match the performance adjustment accuracy of the active suspension with the vehicle's dynamic driving state.

9. An active suspension control device based on driving style recognition and road surface prediction, characterized in that, The device includes: The data acquisition module is used to collect vehicle status parameters and driving environment data in real time during the vehicle's driving process. The information determination module is used to determine the driving style based on the vehicle state parameters, and to determine the road surface prediction information for the vehicle to reach different road surface areas based on the driving environment data. The strategy determination module is used to determine a target control strategy for the active suspension of the vehicle based on the driving style, the vehicle state parameters and the road surface preview information. The control module is used to control the action of the active suspension actuator according to the target control strategy.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images