Training method of vehicle suspension control model, vehicle suspension control method and device

Abstract

This application provides a training method for a vehicle suspension control model, a vehicle suspension control method, and an apparatus, relating to the field of artificial intelligence, and solves the problem of poor suspension control performance in related technologies. The vehicle suspension control method includes: acquiring vehicle operating parameters in response to a vehicle suspension adjustment trigger event; determining vehicle suspension control parameters based on the vehicle operating parameters and through a vehicle suspension control model; and controlling the vehicle suspension to perform adjustment operations based on the vehicle suspension control parameters. By inputting the vehicle operating parameters into a vehicle suspension control model trained with vehicle stability-related reward values ​​and training vehicle operating parameters, this approach overcomes the limitations of single-target control in related technologies, transforming the model's output control parameters into suspension adjustment actions. This allows for dynamic adaptation to changing operating conditions, improving the suspension's adaptability to complex road conditions and thus enhancing suspension control performance.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to training methods for vehicle suspension control models, vehicle suspension control methods, and devices. Background Technology

[0002] The vehicle suspension system, as a key component connecting the vehicle body and wheels, buffers vibrations and shocks caused by uneven road surfaces, suppresses changes in vehicle attitude (such as roll and pitch), and ensures good wheel-to-ground grip, balancing the three core requirements of ride comfort, handling stability, and suspension durability. The performance of the vehicle suspension system depends on the suspension control technology employed.

[0003] However, in related technologies, suspension control algorithms rely excessively on model accuracy and empirical rules, which can easily lead to control decision deviations due to model parameter mismatch and incomplete coverage of operating conditions. This not only results in low control accuracy but also fails to unleash the dynamic adaptation potential of active or semi-active suspensions, severely restricting the overall performance improvement of the suspension system. Summary of the Invention

[0004] To overcome the technical problem of poor suspension control performance in related technologies, this application provides a training method for a vehicle suspension control model, a vehicle suspension control method, and a device.

[0005] According to a first aspect of the embodiments of this application, a method for training a vehicle suspension control model is provided. The method includes: determining a reward value based on the operating parameters of the training vehicle and driving environment data; training the suspension control model with the reward value as the training constraint direction based on the operating parameters of the training vehicle, wherein the reward value is used for model parameters and suspension action parameters; and using the suspension control model to control the vehicle suspension to perform adjustment operations.

[0006] In some possible implementations, the method further includes: determining the operating parameters of the training vehicle using a vehicle suspension simulation model based on a preset control current, a first vehicle speed, and the vehicle road surface generation results.

[0007] In some possible implementations, the suspension action parameters include the vehicle suspension control current, and also include: determining the updated training vehicle operating parameters based on the vehicle suspension simulation model using the vehicle suspension control current, the second vehicle speed, and the vehicle road surface generation results.

[0008] In some possible implementations, the method further includes: randomly inserting disturbance excitations based on road surface roughness to generate vehicle road surface generation results; the disturbance excitations include at least one of bump excitations and hump excitations; the vehicle road surface generation results are used to characterize the road surface height sequence of the wheels on the target vehicle road surface; the greater the road surface roughness, the greater the difference between the road surface height sequences of each wheel.

[0009] In some possible implementations, the operating parameters of the training vehicle include at least one of vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed.

[0010] In some possible implementations, driving environment data includes steering wheel angle, vehicle speed, and road surface roughness. A reward value is determined based on the operating parameters of the training vehicle and the driving environment data, including: determining the reward value based on the sum of the products of ride comfort weight and vehicle body vibration parameters, handling stability weight and handling stability parameters, and durability weight and suspension load parameters; wherein, the higher the vehicle speed, the greater the ride comfort weight; the larger the steering wheel angle, the greater the handling stability weight; and the greater the road surface roughness, the greater the durability weight.

[0011] In some possible implementations, vehicle body vibration parameters include vertical acceleration; operational stability parameters include at least one of roll acceleration and pitch acceleration; and suspension load parameters include wheel suspension damper travel.

[0012] In some possible implementations, determining model parameters and suspension action parameters includes: determining the target model parameters of the suspension control model corresponding to the target vehicle road surface when the reward value of the target vehicle road surface meets the reward condition.

[0013] In some possible implementations, the reward conditions are met, including: the ratio of the target reward value to the reward threshold is greater than a constant threshold; the reward threshold is set based on the baseline performance of the target vehicle in the road scenario.

[0014] In some possible implementations, the method further includes: initializing the vehicle suspension parameters of the vehicle suspension simulation model, the vehicle suspension parameters including at least one of vehicle body mass, moment of inertia, suspension stiffness, distance from the front and rear axles to the vehicle center of gravity, and maximum and minimum damping coefficients.

[0015] According to a second aspect of the embodiments of this application, a vehicle suspension control method is provided. The vehicle suspension control method includes: in response to a vehicle suspension adjustment trigger event, acquiring vehicle operating parameters; the vehicle operating parameters being used to characterize the suspension motion state; determining vehicle suspension control parameters based on the vehicle operating parameters and a vehicle suspension control model; the vehicle suspension control model being trained according to the training method of the vehicle suspension control model provided in the first aspect of the embodiments of this application and any one thereof; and controlling the vehicle suspension to perform an adjustment operation based on the vehicle suspension control parameters.

[0016] In some possible implementations, vehicle operating parameters include at least one of the following: vertical velocity of the suspension, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, and roll acceleration.

[0017] In some possible implementations, the vehicle suspension adjustment trigger event includes at least one of the following: actively calling the vehicle suspension control model at a preset period; detecting that the vehicle is traveling to a preset road condition scenario; the preset road condition scenario includes at least one of random road surface, obstacle road surface, continuous sloping road surface, and discrete pothole road surface; receiving a driver operation command; the operation command includes at least one of acceleration command, braking command, steering command, and suspension mode switching command; obtaining road condition forecast information within a preset distance in front of the vehicle; the road condition forecast information includes at least one of road slope change, road surface material change, and construction section warning.

[0018] In some possible implementations, the vehicle suspension control parameters include vehicle suspension control current; and the vehicle suspension is controlled to perform adjustment operations based on the vehicle suspension control parameters, including: controlling the vehicle damping force based on the vehicle suspension control current to dynamically suppress vertical vibration, pitch and roll motion of the vehicle body.

[0019] In some possible implementations, the method further includes: acquiring the adjusted vehicle operating parameters while controlling the vehicle suspension to perform an adjustment operation; obtaining a suspension control evaluation result based on a preset parameter threshold and the adjusted vehicle operating parameters; and updating the suspension control model based on the suspension control evaluation result.

[0020] According to a third aspect of the embodiments of this application, a training device for a vehicle suspension control model is provided. The training device for the vehicle suspension control model includes: a reward calculation module and a model training module; the reward calculation module is configured to determine a reward value based on the operating parameters of the training vehicle and driving environment data; the model training module is configured to train the suspension control model with the reward value as the training constraint direction based on the operating parameters of the training vehicle, and determine model parameters and suspension action parameters; the suspension control model is used to control the vehicle suspension to perform adjustment operations.

[0021] According to a fourth aspect of the embodiments of this application, a vehicle suspension control device is provided, comprising: a sensor group and an on-board control unit; the sensor group is used to acquire vehicle operating parameters in response to a vehicle suspension adjustment trigger event; the vehicle operating parameters are used to characterize the suspension motion state; the on-board control unit is used to determine vehicle suspension control parameters based on the vehicle operating parameters and a vehicle suspension control model; the vehicle suspension control model is trained according to the training method of the vehicle suspension control model provided in the first aspect of the embodiments of this application and any one thereof; and the vehicle suspension is controlled to perform adjustment operations according to the vehicle suspension control parameters.

[0022] According to a fifth aspect of the present application, an electronic device is provided, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method of either the first aspect or the second aspect.

[0023] According to a sixth aspect of the embodiments of this application, a non-transitory computer-readable storage medium is provided storing computer instructions for causing a computer to perform the method of either the first aspect or the second aspect.

[0024] According to a seventh aspect of the embodiments of this application, a computer program product is provided, including a computer program and computer instructions for causing a computer to perform the method of either the first aspect or the second aspect.

[0025] The technical solutions provided by the embodiments of this application can include the following beneficial effects: By triggering suspension adjustment events, changes in operating conditions and driving needs can be accurately captured, and vehicle operating parameters can be collected synchronously to provide data support for subsequent control that adapts to the operating conditions, avoiding control deviations caused by perception lag and blind parameter collection in related suspension control technologies. By inputting vehicle operating parameters into a vehicle suspension control model trained with vehicle stability-related reward values ​​and training vehicle operating parameters, the limitations of single-target control in related technologies can be overcome. The control parameters output by the model can be transformed into suspension adjustment actions, which can dynamically adapt to operating conditions. Compared with fixed damping or segmented control in related technologies, this significantly improves the suspension's adaptability to complex road conditions, thereby improving suspension control performance.

[0026] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0027] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0028] Figure 1 A schematic flowchart illustrating a vehicle suspension control method provided in an embodiment of this application;

[0029] Figure 2 A flowchart illustrating a training method for a vehicle suspension control model provided in an embodiment of this application; Figure 3 A schematic diagram of the architecture of a training method for a vehicle suspension control model provided in an embodiment of this application; Figure 4 A flowchart illustrating a training method for a vehicle suspension control model provided in an embodiment of this application; Figure 5 A flowchart illustrating a training method for a vehicle suspension control model provided in an embodiment of this application; Figure 6 A flowchart illustrating a training method for a vehicle suspension control model provided in an embodiment of this application; Figure 7 A schematic diagram of the structure of a training device for a vehicle suspension control model provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of a vehicle suspension control device provided in an embodiment of this application; Figure 9 A schematic diagram of the structure of a vehicle provided in an embodiment of this application; Figure 10 This is a schematic diagram of the structure of an electronic device according to some embodiments of this application. Detailed Implementation

[0030] Some embodiments of this application will be described in detail herein, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. Various changes, modifications, and equivalents of the methods and apparatus described herein will become apparent upon understanding this application. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but can be changed as will become apparent upon understanding this application, except for operations that must be performed in a specific order. Furthermore, for clarity and brevity, descriptions of features known in the art may be omitted.

[0031] The embodiments described in the following examples of this application do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0032] The vehicle suspension system, as a key load-bearing and vibration-damping component, directly determines the vehicle's ride comfort, handling stability, and suspension structural durability. Ride comfort relates to passenger comfort and is achieved by buffering vibrations and impacts caused by uneven road surfaces. Handling stability ensures controllable vehicle trajectory by suppressing changes in body roll and pitch. Suspension durability depends on the alternating loads borne by the suspension components, and excessive vibration leading to component fatigue damage must be avoided.

[0033] Vehicle suspension systems come in various forms. For example, they can be passive, active, or semi-active suspensions. Passive suspensions control vibrations using springs with fixed stiffness and dampers with fixed damping. However, passive suspensions struggle to balance ride comfort and handling under complex conditions. For instance, they cannot dynamically adjust spring stiffness and damper damping based on real-time road conditions (such as potholes and speed bumps) and driving states (such as high-speed cruising and emergency steering). To overcome the performance limitations of passive suspensions, active and semi-active suspensions have emerged and are gradually becoming the trend. Active suspension control methods involve calculating and outputting the active control force of the actuators in real time to counteract road vibration excitation. Semi-active suspension control methods involve adjusting damping in real time by deciding on key parameters such as the target damping value and adjustment rate of adjustable damping elements. Both active and semi-active suspensions have the potential to dynamically adapt to different operating conditions.

[0034] However, the performance of semi-active and active suspensions is highly dependent on the control algorithm, making it difficult to fully utilize their adjustable capabilities. The control algorithms for active or semi-active suspensions include model-based control algorithms (such as proportional-integral-derivative PID control and model predictive control) and rule-based control algorithms (such as fuzzy control and expert systems). Model-based control requires an accurate vehicle dynamics model, but in actual vehicle operation, factors such as load changes, temperature drift, and random road disturbances can lead to model parameter mismatch, thus reducing control accuracy. While rule-based control is more robust, the rule base relies on engineer experience and cannot cover all complex operating conditions. Furthermore, it cannot learn and optimize autonomously, resulting in a significant decrease in control performance when facing new, unresearched operating conditions. In addition, these algorithms generally suffer from the problem of "difficulty in dynamically coordinating multiple conflicting objectives." For example, when a vehicle simultaneously faces "road bumps" and "emergency steering," the algorithms cannot quickly balance the dynamic switching between "smoothness priority" and "maneuverability priority," easily leading to lagging or biased control decisions and failing to achieve optimal performance balance across all operating conditions.

[0035] To address the aforementioned issues, this application provides a vehicle suspension control method that dynamically adjusts the suspension control strategy based on the real-time status and road conditions of the vehicle during operation, effectively improving vehicle performance. This method is applicable to various complex road conditions. Complex road conditions include, but are not limited to, random road surfaces of different grades (e.g., ISO standards A to F), man-made obstacles (e.g., speed bumps), and continuous or discrete slopes and potholes. This application controls the damping force by adjusting the control current of each shock absorber in the suspension in real time, thereby dynamically suppressing the vehicle's vertical vibration, pitch, and roll movements, comprehensively improving ride comfort, handling stability, and driving safety.

[0036] The vehicle can take many forms. For example, it can be a passenger car, a sports utility vehicle (SUV), or a commercial vehicle equipped with a semi-active or active suspension system. This application does not limit the specific type of vehicle.

[0037] Figure 1 This is a schematic flowchart illustrating a vehicle suspension control method provided in an embodiment of this application. Figure 1 As shown, in some embodiments, the vehicle suspension control method includes the following steps: S11, in response to the vehicle suspension adjustment trigger event, obtains the vehicle operating parameters.

[0038] To achieve timely and proactive suspension control, and to ensure that the suspension system can respond in advance or in real time to various operating condition changes that affect driving performance, in some embodiments, the vehicle suspension adjustment trigger event includes at least one of the following: The vehicle suspension control model is actively invoked at preset intervals. For example, the preset interval can be adaptively adjusted according to the vehicle's speed and the complexity of road conditions. For instance, the preset interval can be set to 10ms-50ms when driving at high speeds and 50ms-100ms when driving at low speeds.

[0039] The system detects that the vehicle is traveling in a preset road condition scenario. These preset road condition scenarios include at least one of the following: random road surface, obstacle road surface, continuous sloping road surface, and discrete pothole road surface. Random road surfaces can be classified into different roughness levels from A to F according to ISO standards. Obstacle road surfaces include speed bumps, protruding stones, sunken manhole covers, and other man-made or naturally formed obstacles. Continuous sloping road surfaces include continuous uphill slopes, continuous downhill slopes, and undulating slopes. Discrete pothole road surfaces include one or more discretely distributed pothole areas. For example, a road condition detection module (such as a camera, LiDAR, or road surface perception sensor) identifies the road condition scenario the vehicle is in in real time and compares this scenario with the conditions corresponding to various preset road condition scenarios. If the road condition scenario matches a preset scenario, suspension adjustments are triggered to adapt to the road conditions.

[0040] The driver's operation command has been received. The operation command includes at least one of the following: acceleration command, braking command, steering command, and suspension mode switching command.

[0041] The system acquires road condition forecast information within a preset distance ahead of the vehicle. This forecast information includes at least one of the following: changes in road gradient, changes in road surface material, and warnings of road construction. For example, this forecast information can be acquired through an in-vehicle navigation system, a vehicle-to-everything (V2X) system, or a forward-facing sensing device. For example, the preset distance ahead of the vehicle can be in the range of 50m to 200m. It is understood that the preset distance ahead of the vehicle can be adjusted according to vehicle speed, and this application does not limit this adjustment. For instance, the higher the vehicle speed, the larger the preset distance, to ensure sufficient adjustment time.

[0042] In some embodiments, vehicle operating parameters include at least one of the following: vertical velocity, pitch angular velocity, roll angular velocity, vertical acceleration, pitch acceleration, and roll acceleration of the suspension. In some embodiments, vehicle operating parameters can be acquired in real time by onboard sensors (such as acceleration sensors, angular velocity sensors, displacement sensors, etc.), with the acquisition frequency matching the aforementioned preset period to ensure the real-time performance and validity of the parameters.

[0043] S12, Based on the vehicle's operating parameters, determine the vehicle suspension control parameters using the vehicle suspension control model.

[0044] The vehicle suspension control model is trained based on vehicle stability-related reward values ​​and training vehicle operating parameters. "Vehicle stability-related" includes, but is not limited to, at least one of "ride comfort, handling stability, and suspension durability." In some embodiments, the vehicle suspension control model can be trained using, for example... Figure 2 The training method shown is used to obtain the vehicle suspension control model.

[0045] In some embodiments, the reward value is determined based on the training vehicle operating parameters and driving environment data output by the vehicle suspension simulation model, to determine the training constraint direction of the suspension control model. These three factors together constitute the optimization objective of vehicle suspension control, preventing optimization of a single objective from leading to the deterioration of other performance aspects.

[0046] In some embodiments, the operating parameters of the training vehicle include at least one of vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed.

[0047] In some embodiments, driving environment data includes at least one of steering wheel angle, vehicle speed, and road surface roughness.

[0048] Based on this, the vehicle suspension control parameters are determined by inputting the vehicle's operating parameters into the trained vehicle suspension control model.

[0049] In some embodiments, vehicle suspension control parameters include vehicle suspension control current. The vehicle suspension control current can be four-wheel suspension control current. It is understood that the number of vehicle suspension control currents is related to the number of wheels and suspensions provided in the vehicle, and this application does not limit this number.

[0050] S13, based on the vehicle suspension control parameters, controls the vehicle suspension to perform adjustment operations.

[0051] In some embodiments, the vehicle damping force is controlled according to the vehicle suspension control current to dynamically suppress the vertical vibration, pitch and roll motion of the vehicle body, thereby realizing the adjustment operation of the vehicle suspension.

[0052] By executing steps S11 to S13, in response to the suspension adjustment trigger event, the system accurately captures changes in operating conditions and driving needs, and synchronously collects vehicle operating parameters. This provides data support for subsequent control to adapt to the operating conditions, avoiding control deviations caused by perception lag and blind parameter collection in related suspension control technologies. Inputting vehicle operating parameters into a vehicle suspension control model trained with vehicle stability-related reward values ​​and training vehicle operating parameters overcomes the limitations of single-target control in related technologies. Transforming the model's output control parameters into suspension adjustment actions allows for dynamic adaptation to operating conditions. Compared to fixed damping or segmented control in related technologies, this significantly improves the suspension's adaptability to complex road conditions, thereby enhancing suspension control performance.

[0053] To further optimize and improve vehicle suspension control, in some embodiments, the vehicle suspension control method further includes the following step: determining the suspension control evaluation result based on the collected real-time vehicle operating parameters under the vehicle suspension control parameters.

[0054] In some embodiments, when the vehicle suspension is adjusted, the adjusted vehicle operating parameters are acquired in real time. These parameters include at least one of the following: vertical velocity, vertical acceleration, pitch velocity, pitch acceleration, roll velocity, and roll acceleration. A suspension control evaluation result is obtained based on preset parameter thresholds and the adjusted vehicle operating parameters. The suspension control model is then updated based on the evaluation result.

[0055] In some embodiments, updating the suspension control model includes updating the model parameters of the suspension control model, including the learning rate. The suspension control model with adjusted model parameters is trained based on the adjusted vehicle operating parameters.

[0056] Among them, the suspension control evaluation results are obtained based on preset parameter thresholds and adjusted vehicle operating parameters, and there are multiple implementation forms.

[0057] For example, a quantitative comparison can be used. The deviation rate of each vehicle's operating parameters from the corresponding preset threshold is calculated. If the deviation rate of all parameters is ≤10%, and no parameter exceeds the threshold within 50 consecutive data acquisition cycles (500ms), the evaluation result is considered satisfactory. If the deviation rate of any parameter is >20%, or the duration of exceeding the threshold exceeds 10 data acquisition cycles (100ms), the evaluation result is considered unsatisfactory.

[0058] For example, a qualitative comparison using graphs can also be used. By comparing the time-series variation curves of each vehicle's operating parameters with the standard interval curves corresponding to preset thresholds, the control effect can be qualitatively judged. For instance, if the vertical acceleration time-series curve frequently exceeds the threshold curve and the fluctuation frequency is >5 times / second, then the qualitative assessment result is determined to not meet the assessment conditions.

[0059] It is understood that the above evaluation conditions can be flexibly set based on the actual needs of the scenario, and are not fixed. For example, for urban paved road scenarios, more stringent parameter thresholds (such as vertical acceleration ±0.3g) can be set to pursue higher comfort; for off-road scenarios, the thresholds can be appropriately relaxed (such as vertical acceleration ±0.8g) to adapt to complex road conditions, and this application does not limit this.

[0060] For example, if the suspension control evaluation result does not meet the evaluation conditions, the convergence curves of the reinforcement learning model (e.g., the actor loss convergence curve and the critic loss convergence curve) are obtained, and the model parameters of the reinforcement learning model are adjusted according to the convergence curves. It is understood that the evaluation conditions are set based on scenario requirements, and this application does not limit them.

[0061] For example, if the fluctuation range of the actor loss curve is greater than 0.1 and cannot reach a stable state, it indicates that the learning rate is too high, which leads to unstable model training. The learning rate should be reduced to 50% of the original value (e.g., from 0.001 to 0.0005), while keeping the discount factor, iteration step size and other parameters unchanged to suppress curve oscillation.

[0062] For example, if the critical loss curve converges too slowly (more than 10,000 iterations are required to reach the loss threshold of 0.01), it indicates that the learning rate is too low. Increasing the learning rate to 1.2 times the original value (e.g., from 0.001 to 0.0012) will accelerate the convergence efficiency of the loss curve.

[0063] For example, if both curves tend to stabilize but the loss value is higher than the preset threshold (actor loss > 0.02, critic loss > 0.01), it indicates that the model is trapped in a local optimum. The learning rate is fine-tuned to 80% of the original value, and the iteration step size of the actor network is adjusted to 80 to break the local optimum and further reduce the loss.

[0064] For example, evaluation indicators can be set, such as driving smoothness can be evaluated by the root mean square value of the vehicle's vertical acceleration, handling stability can be evaluated by the maximum value of the roll rate, and suspension durability can be evaluated by the cumulative value of the shock absorber travel. By substituting the real-time collected vehicle operating parameters into the evaluation indicators, the control effect evaluation results under different road conditions and different driving conditions can be obtained.

[0065] In this way, the target model is experimentally verified in a real road environment, the control effect of the method under different road conditions is evaluated, and the method is optimized and improved based on the verification results.

[0066] The aforementioned vehicle suspension control methods rely on the accuracy and adaptability of the vehicle suspension control model. The training quality of the model directly determines whether multi-objective optimization of ride comfort, handling stability, and suspension durability can be achieved under complex road conditions. To ensure that the vehicle suspension control model can adapt to various complex road conditions (including random road surfaces of different levels, special obstacle roads, continuous or discrete slopes, and potholes), and output accurate control parameters, this application further provides a training method for the vehicle suspension control model. This training method, through scientific sample construction, reward function design, and iterative training process, enables the trained model to possess excellent dynamic response capabilities and operating condition adaptability, and can be directly applied to the aforementioned vehicle suspension control methods.

[0067] Figure 2 This is a flowchart illustrating a training method for a vehicle suspension control model provided in an embodiment of this application. Figure 2 As shown, in some embodiments, the training method for the vehicle suspension control model includes the following steps: S101 determines the reward value based on the operating parameters of the training vehicle and driving environment data.

[0068] In some embodiments, the operating parameters of the training vehicle include at least one of vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed.

[0069] In some embodiments, the operating parameters of the training vehicle are obtained from the output of the vehicle suspension simulation model.

[0070] The vehicle suspension simulation model can be a multi-degree-of-freedom (DOF) suspension simulation model. In some embodiments, the multi-DOF suspension simulation model can be a seven-DOF suspension simulation model. The seven degrees of freedom include three motion degrees of freedom of the vehicle body (vertical translational degree of freedom, pitch rotational degree of freedom, and roll rotational degree of freedom) and one vertical translational degree of freedom for each of the four wheels. In one implementation, a seven-DOF suspension simulation model of the whole vehicle is built based on the Simulink module of the simulation software. For example, the integration and construction of the dynamic equation solving module, signal input module, signal output module, and data calculation module are completed through module construction, parameter configuration, and signal connection to construct the vehicle suspension simulation model. The vehicle suspension parameters of the vehicle suspension simulation model include at least one of the following: vehicle mass, moment of inertia, suspension stiffness, distance from the front and rear axles to the vehicle's center of gravity, and maximum and minimum damping coefficients. In one implementation, the training method of the vehicle suspension control model includes the following steps: initializing the vehicle suspension parameters of the vehicle suspension simulation model.

[0071] The input signals of the vehicle suspension simulation model include at least one of the following: the vehicle road surface generation result (e.g., four-wheel road surface input, i.e., the road surface height sequence corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel), vehicle speed (vehicle speed signal), and four-wheel suspension control current (control current signals of the left front wheel suspension, right front wheel suspension, left rear wheel suspension, and right rear wheel suspension respectively).

[0072] The vehicle road surface generation result includes a road surface height sequence of the wheels. In some embodiments, the greater the road surface roughness, the greater the difference between the road surface height sequences of each wheel. For example, the road surface height sequence can be generated using a randomization algorithm based on a preset road surface amplitude range. The road surface amplitude range can be determined according to the road surface roughness. The vehicle's driving path is divided into continuous spatial segmented regions, corresponding to multiple sampling points, ensuring that the sequence sampling range completely covers the target driving path and ensuring the spatial matching of the generated result with the vehicle's driving scenario. The road surface height sequence can include four sub-sequences, each corresponding to the road surface elevation change trajectory at different wheel positions of the vehicle, such as the right front wheel, left front wheel, and right rear wheel. For example, the sampling points of this road surface height sequence correspond to continuous spatial segmented regions of the road surface under the vehicle's driving path, and the sampling point numbers can cover 0 to 5000. The elevation value fluctuation range of the sequence is -0.6 to 0.4. In the initial segmented region of the driving path (corresponding to sampling points 0 to 1000), the road surface elevation exhibits significant fluctuations, with a maximum fluctuation amplitude of approximately 1.0 unit, corresponding to relatively drastic vertical changes in the corresponding road segment. In the subsequent segmented areas (corresponding to 1000 to 5000 sampling points), the road surface elevation generally fluctuated within a small range of -0.2 to 0.2, indicating that the road surface of the corresponding sections was relatively flat with only minor undulations. At the same time, the changing trends of the four subsequences were highly consistent, indicating that the elevation distribution of the generated road surface in the lateral regions corresponding to the four wheels of the vehicle has strong spatial continuity.

[0073] Understandably, when the road surface roughness is low (i.e., the road surface is relatively smooth), the random deviation of each wheel subsequence is controlled to be within a very small range, so that the changing trends of the four subsequences are highly consistent, reflecting the elevation spatial continuity of the smooth road surface in the lateral region. If the road surface roughness is high, the random deviation between the subsequences is increased, so that the difference between the subsequences expands with the increase of roughness, restoring the lateral non-uniformity of the rough road surface.

[0074] To enhance the generalization ability of the model, in some embodiments, the training method of the vehicle suspension control model further includes: randomly inserting disturbance excitations based on road surface roughness to generate vehicle road surface generation results. The disturbance excitations include at least one of bump excitations and hump excitations. This yields a road surface height sequence for each wheel.

[0075] In some embodiments, the output signal of the vehicle suspension simulation model includes vehicle motion-related parameters and suspension component-related parameters. The vehicle motion parameters include vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, and roll acceleration. The suspension component parameters include at least one of the travel of the shock absorbers for the left front wheel, right front wheel, left rear wheel, and right rear wheel suspensions, and the corresponding shock absorber speeds.

[0076] In some embodiments, the training vehicle operating parameters output by the vehicle suspension simulation model include at least one of vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed.

[0077] Driving environment data can be vehicle operating parameters used for real-time vehicle training. In some embodiments, the vehicle operating parameters used for real-time vehicle training include at least one of vehicle speed, steering angle, pitch angle, and road surface roughness. This allows for the determination of dynamically changing weights to adapt to more driving scenarios. See subsequent embodiments.

[0078] In some embodiments, driving environment data includes steering wheel angle and vehicle speed. Based on the training vehicle operating parameters output by the vehicle suspension simulation model and the driving environment data, a reward value is determined, including: determining the reward value based on the sum of the products of ride comfort weight and body vibration parameters, handling stability weight and handling stability parameters, and durability weight and suspension load parameters. Specifically, the higher the vehicle speed, the greater the ride comfort weight; the larger the steering wheel angle, the greater the handling stability weight; and the greater the road surface roughness, the greater the durability weight.

[0079] In some embodiments, vehicle body vibration parameters include vertical acceleration. Operational stability parameters include at least one of roll acceleration and pitch acceleration. Suspension load parameters include wheel suspension damper travel.

[0080] For example, the formula for calculating the reward value includes: reward = -(w_heave heave_acc + w_roll roll_acc + w_stroke stroke); Where reward represents the reward value; w_heave represents the ride comfort weight; w_roll represents the handling stability weight; w_stroke represents the durability weight; heave_acc represents the vertical acceleration of the vehicle body; roll_acc represents the roll angle acceleration of the vehicle body; and stroke represents the suspension travel.

[0081] For example, the formula for calculating the smoothness weight includes: w_heave = 1.0 + min(0.8, speed / 120); Here, speed represents vehicle speed.

[0082] For example, the formula for calculating the manipulation stability weights includes: w_roll = 1.0 + min(1.0, abs(steer) / 0.3); Here, steer represents the steering wheel angle.

[0083] For example, the formula for calculating the durability weight includes: w_stroke = 1.0 + (1 if road_roughness>0.6, else 0.2); Here, road_roughness represents the road surface roughness.

[0084] When the road surface roughness is greater than 0.6, w_stroke equals 2. When the road surface roughness is less than or equal to 0.6, w_stroke equals 1.2. For example, if the road surface roughness is 0.8, w_stroke equals 2. If the road surface roughness is 0.3, w_stroke equals 1.2.

[0085] In this way, by adaptively and dynamically adjusting the weights of multiple targets such as vehicle acceleration, roll angle acceleration, and suspension travel in the reward function, a dynamic balance between ride comfort, handling, and durability is achieved.

[0086] S102, based on the working parameters of the training vehicle, train the suspension control model with the reward value as the training constraint direction, and determine the model parameters and suspension action parameters.

[0087] The suspension control model is used to control the vehicle suspension to perform adjustment operations.

[0088] In some embodiments, the training vehicle operating parameters and reward values ​​output by the vehicle suspension simulation model are input into the suspension control model to determine the model parameters and suspension action parameters of the suspension control model.

[0089] In some embodiments, the suspension control model can be a neural network model. For example, it can be a reinforcement learning network model. In one implementation, the suspension control model is constructed based on the Proximal Policy Optimization (PPO) algorithm. The PPO algorithm includes an Actor network and a Critic network. The Actor network generates a vehicle suspension control strategy and corresponding suspension action parameters based on the input training vehicle operating parameters and reward values. The Critic network evaluates the execution effect of the vehicle suspension control strategy and suspension action parameters generated by the Actor network, outputting a value evaluation result as a feedback signal to guide the iterative optimization of the Actor network's model parameters. For example, the neural network structure of this model can include an input layer, a hidden layer, and an output layer. The input layer is configured with eight neurons, corresponding to the training vehicle operating parameters for vehicle suspension control, which may include key features such as vehicle vertical acceleration, suspension travel, wheel vertical load, vehicle speed, and road surface grade coefficient. The hidden layer is set to two layers. The first hidden layer is configured with 64 neurons, and the activation function is a Modified Linear Unit (ReLU) to achieve nonlinear feature mapping. The second hidden layer has 32 neurons and also uses the ReLU activation function. It improves the effectiveness of feature representation through feature dimensionality reduction and fusion, while avoiding overfitting and computational latency caused by excessively deep network layers. The output layers are designed according to the network's functional differences. The Actor network output layer has two neurons with the Tanh activation function. The output parameters are mapped to the [-1,1] interval and converted into the two core action parameters: suspension damping adjustment coefficient and stiffness adjustment coefficient. The Critic network output layer has only one neuron and no additional activation function. It provides value evaluation feedback for the Actor network's parameter iteration through a reward value. This network structure effectively controls model complexity while ensuring feature extraction capabilities, adapting to the computational resources of the in-vehicle terminal and meeting the real-time control requirements of the suspension system.

[0090] like Figure 3 As shown, in some embodiments, the suspension control model can be integrated into a reinforcement learning agent. The vehicle suspension simulation model is thus... Figure 3 The "Environment" module is shown. The agent receives vehicle operating parameters and reward values ​​from the environment for training, and outputs actions (e.g., suspension action parameters) to act on the environment. In this way, the suspension control model is trained through reinforcement learning.

[0091] Based on this, the reward value is determined according to the training vehicle operating parameters and driving environment data output by the vehicle suspension simulation model. The objective data output by the simulation is used as the core basis to generate reward feedback, so that the reward value can accurately reflect the actual operating state and performance of the suspension system. This provides a real and reliable feedback basis for the training of the suspension control model, ensures the accuracy and pertinence of the feedback information during the model training process, and provides scientific support for the optimization and adjustment of model parameters.

[0092] In some embodiments, the iterative update mechanism employs suspension action parameters to trigger continuous updates of the training vehicle's operating parameters and reward values. The updated parameters are then re-input into the suspension control model to optimize the model parameters and output new suspension action parameters, forming a closed-loop iterative training logic. This process enables the suspension control model to continuously obtain dynamic feedback based on its output control results, driving the model parameters to continuously correct themselves in a direction that adapts to the dynamic adjustment needs of the suspension system. This gradually optimizes the model's control logic for the suspension system, making the model's output suspension action parameters increasingly aligned with the actual adjustment needs of the suspension system, significantly improving the control accuracy and response adaptability of the suspension control model.

[0093] In some embodiments, based on the vehicle suspension simulation model obtained in steps S101 and S102, by applying such... Figure 1 The suspension control method shown can call up optimized parameters in real time to adjust suspension performance and improve the vehicle's core driving indicators.

[0094] The following is an example illustration.

[0095] like Figure 4 As shown, in some embodiments, the training method for the vehicle suspension control model further includes the following steps: S201, based on the preset control current, the first vehicle speed, and the vehicle road surface generation results, determines the working parameters of the training vehicle through the vehicle suspension simulation model.

[0096] The preset control current is the baseline control parameter for the initial output of the model or before iteration.

[0097] The first vehicle speed is the preset baseline driving speed (such as 30km / h, 60km / h, covering common operating conditions).

[0098] The results of vehicle road surface generation can be found in the aforementioned embodiments, and will not be repeated here.

[0099] The above three parameters are input into the vehicle suspension simulation model. The model simulates the actual suspension working state and outputs training vehicle working parameters such as vertical acceleration, shock absorber travel, and pitch rate, which are used as initial training data. For ease of description, these are referred to as initial training vehicle working parameters.

[0100] Thus, after executing step S201, steps S101 and S102 are executed to determine the model parameters and suspension action parameters corresponding to the initial training vehicle operating parameters. For ease of description, these are referred to as the first model parameters and the first suspension action parameters.

[0101] Among them, the suspension action parameter is the vehicle suspension control current. That is, the first suspension action parameter is the first vehicle suspension control current.

[0102] S202, based on the vehicle suspension control current, the second vehicle speed, and the vehicle road surface generation results, the updated operating parameters of the training vehicle are determined through the vehicle suspension simulation model.

[0103] Input the suspension action parameters (i.e., the first vehicle suspension control current), the second vehicle speed, and the vehicle road surface generation results corresponding to the initial training vehicle operating parameters in step S201 into the vehicle suspension simulation model to determine the updated training vehicle operating parameters.

[0104] The vehicle suspension control current is a new value optimized from the model output parameters in S201, rather than the initial preset value. The second vehicle speed can be the same as the first vehicle speed (for fixed-condition verification) or adjusted (to adapt to multi-speed scenarios, such as adjusting from 60km / h to 80km / h). The vehicle road surface generation results can use the S201 scenario (for iteration under the same condition) or switch to a new scenario (to expand model adaptability). For example, a new road surface height sequence can be randomly generated.

[0105] The updated training vehicle operating parameters are output using the same simulation model and compared with the initial parameters of S201. The reward value is recalculated in conjunction with driving environment data. The updated state parameters and reward value are then fed back to the model to complete one round of parameter optimization.

[0106] like Figure 5 As shown, exemplarily, the training method for this vehicle suspension control model includes the following steps: S301, input the first vehicle suspension control current, the second vehicle speed and the vehicle road surface generation results into the vehicle suspension simulation model, and determine the first training vehicle working parameters output by the vehicle suspension simulation model.

[0107] The first vehicle suspension control current includes the suspension control current corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel at time t.

[0108] The second vehicle speed is the real-time vehicle speed (i.e., the vehicle speed at time t).

[0109] The vehicle road surface generation result includes the road surface height sequence of the wheels corresponding to the target vehicle's road surface. The generation of the road surface height sequence can be found in the description in step S101, and will not be repeated here.

[0110] In some embodiments, the operating parameters of the first training vehicle include at least one of the following at time t: vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed. See the description in step S101 for further details.

[0111] S302, determine the first reward value based on the operating parameters of the first training vehicle and the driving environment data.

[0112] Please refer to the description in step S101, which will not be repeated here.

[0113] S303, input the first training vehicle operating parameters and the first reward value into the suspension control model, update the model parameters of the suspension control model and the first suspension action parameters, so as to obtain the updated model parameters of the suspension control model and the second suspension action parameters.

[0114] The second suspension action parameters include the second vehicle suspension control current.

[0115] The second vehicle suspension control current includes the suspension control current corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel at time t+1.

[0116] S304: Input the second vehicle suspension control current, the third vehicle speed, and the vehicle road surface generation results into the vehicle suspension simulation model to determine the second training vehicle operating parameters output by the vehicle suspension simulation model.

[0117] The third speed is the real-time speed (i.e., the speed at time t+1). It is understandable that the value of the third speed can be the same as or different from the value of the second speed.

[0118] In some embodiments, the operating parameters of the second training vehicle include at least one of the following at time t+1: vertical velocity, pitch velocity, roll velocity, vertical acceleration, pitch acceleration, roll acceleration, wheel suspension damper travel, and wheel suspension damper speed. See the description in step S301 for further details.

[0119] S305, determine the second reward value based on the operating parameters of the second training vehicle and driving environment data.

[0120] S306, input the second training vehicle operating parameters and the second reward value into the suspension control model, update the model parameters of the suspension control model and the second suspension action parameters, so as to obtain the updated model parameters of the suspension control model and the third suspension action parameters.

[0121] The third suspension action parameters include the third vehicle suspension control current.

[0122] The third vehicle suspension control current includes the suspension control current corresponding to the left front wheel, right front wheel, left rear wheel, and right rear wheel at time t+2.

[0123] It is understood that steps S301 to S306 only provide two illustrative iterations and do not constitute a limitation on the number of iterations. In practical applications, the number of iterations can be dynamically adjusted based on preset iteration thresholds (such as 1000 times, 5000 times, etc.), model parameter convergence accuracy (such as the change in model parameters in consecutive iterations being less than 1e-5), or reward value convergence status, etc. As long as the performance optimization goal of the suspension control model can be achieved, it can be included in the protection scope of this application.

[0124] In some embodiments, if the target reward value meets the reward condition, the iteration stops, and the target model parameters of the suspension control model corresponding to the target vehicle road surface are determined. That is, when the above reward condition is met, it is determined that the control performance of the iteratively updated suspension control model in the target vehicle road surface scenario is significantly improved compared to the baseline method. At this point, the iteration stops, and the model parameters of the current suspension control model are determined as the target model parameters corresponding to the target vehicle road surface. If the condition is not met, the iterative update step continues until the reward value meets the above ratio requirement.

[0125] like Figure 6 As shown, in some embodiments, the following steps are included: S401, determine the target reward value for the target vehicle on the road surface.

[0126] The target vehicle's road surface can be any of the following: newly paved road (road surface grade A), smooth road surface (road surface grade B), slightly undulating road surface (road surface grade C), slightly uneven road surface (road surface grade D), rough road surface (road surface grade E), and road surface with potholes (road surface grade F). It is understood that road surface grades (A to F) are classified according to their roughness, with A being extremely smooth and F being extremely rough.

[0127] For example, when a road surface of a certain grade (such as a Class C slightly undulating road surface) is selected as the target vehicle road surface, a road surface scenario is constructed, and corresponding driving environment data such as vehicle speed are input to simulate the suspension working state of the vehicle when driving on this road surface. The output is the training vehicle working parameters such as vertical acceleration, shock absorber travel, and pitch rate. Then, based on the preset reward value calculation rules, the above training vehicle working parameters are weighted (the weights correspond to the three dimensions of ride comfort, handling stability, and suspension durability) to finally obtain the target reward value corresponding to the target vehicle road surface.

[0128] S402, if the target reward value meets the reward conditions, determine the target model parameters of the suspension control model corresponding to the road surface of the target vehicle.

[0129] In some embodiments, satisfying the reward condition includes: the ratio of the target reward value to the reward threshold is greater than a constant threshold. The reward threshold is set based on the baseline performance of the target vehicle in the road scenario.

[0130] For example, the target reward value is denoted as reward1. Under the same road surface scenario and the same initial simulation conditions, a baseline control method (e.g., the canopy control algorithm) is invoked to obtain the reward value corresponding to the canopy control algorithm, denoted as reward2 (i.e., the "reward threshold" described in this application). The reward condition is satisfied, i.e., the formula is satisfied: ; The constant threshold 'a' is a value greater than 1. For example, 'a' can take the values ​​1.1, 1.2, or 1.3.

[0131] It is understood that the value of the constant threshold 'a' can be set based on the performance design requirements of the target vehicle, industry standards, and practical engineering experience, and this application does not impose any restrictions on this.

[0132] It is understood that the use of reward2, the reward value of the ceiling control algorithm, as the benchmark in this embodiment is merely illustrative. In other feasible embodiments, the benchmark method may also use conventional suspension control strategies in the industry, such as floor control algorithm, PID control algorithm, and fuzzy control algorithm. This application does not limit the specific methods used.

[0133] To enable reinforcement learning models to adapt to different road conditions and improve the control adaptability and accuracy of active or semi-active suspensions, in some embodiments, a phased progressive training is used to achieve accurate matching between model parameters and road surface grades. The specific training process is described in detail below.

[0134] In some embodiments, by following a progressive training logic of "from simple scenarios to complex scenarios," the complexity of the task is gradually increased, helping the reinforcement learning agent efficiently learn the core suspension control strategy. The progression from simple to complex scenarios can be done by gradually increasing the difficulty according to road surface level gradients. Simple scenarios refer to low-complexity road surface conditions, corresponding to the very smooth road surface of Level A mentioned earlier. In this scenario, suspension vibration interference is small and the control task is easy, allowing the reinforcement learning agent to quickly master the basic suspension control logic and complete the initial strategy learning. Complex scenarios refer to high-complexity road surface conditions, corresponding to the very rough road surface of Level F mentioned earlier. In this scenario, suspension vibration interference is strong and the demand for control parameter adjustment is high, requiring the agent to further optimize and adapt its capabilities based on the basic strategy.

[0135] For example, during training, the training stages are traversed according to the road surface level gradient preset by the course learning mechanism (gradually transitioning from level A to level F). This achieves progressive training, "building a foundation in simple scenarios and improving capabilities in complex scenarios," ensuring that the reinforcement learning agent smoothly learns the core suspension control strategies under all road surface conditions. This avoids the problems of low exploration efficiency and slow convergence caused by directly entering high-complexity scenarios. In other words, after completing the current stage of training and passing the stage evaluation, the next stage of training begins. Each training stage can be referred to in the aforementioned embodiments.

[0136] Based on this, a road environment adapted to the complexity of the current training stage is generated according to the road surface level corresponding to that stage. The road surface level increases sequentially according to the complexity gradient preset by the course learning mechanism. The reinforcement learning agent interacts with the suspension simulation environment, synchronously collecting the suspension training vehicle operating parameters at time t, the control action parameters output by the reinforcement learning agent, the reward value, and the suspension training vehicle operating parameters at time t+1 (collectively referred to as the trajectory dataset). The reward value is calculated using a preset reward function that adopts an adaptive dynamic adjustment of reward weights. Based on the collected trajectory dataset, the PPO algorithm iteratively updates the model parameters of the reinforcement learning agent. The PPO algorithm effectively suppresses parameter oscillations during training and improves the stability of reinforcement learning model training by limiting the difference between the new and old strategies. After each training stage is completed, the performance of the reinforcement learning agent trained in the current stage is evaluated. If the evaluation result meets the standard, the training stage is considered successful, the reinforcement learning model parameters corresponding to the road surface level of the current stage are recorded and saved, and then the next training stage begins. If the evaluation fails to meet the requirements, the current stage's road surface generation, trajectory collection, and PPO algorithm iteration steps are repeated until the stage evaluation requirements are met. Through the above-mentioned phased training method, reinforcement learning model parameters adapted to the corresponding road surface conditions are trained for different road surface levels preset for each stage. This achieves accurate matching between model parameters and road surface levels, ensuring that the reinforcement learning agent can output appropriate suspension control strategies under different road surface scenarios.

[0137] After all preset training phases are completed, the model parameters obtained from each phase are integrated and optimized, and finally, a reinforcement learning model with the ability to adapt to all road conditions is saved. This reinforcement learning model can call parameters corresponding to the road surface level to execute suspension control, providing model support for the actual control of subsequent active or semi-active suspensions.

[0138] Thus, this embodiment, by introducing a course-based learning mechanism, replaces the direct training of high-complexity scenarios with a progressively increasing complexity training mode. This not only helps the reinforcement learning agent efficiently master the core suspension control strategy but also effectively solves the technical problems of low exploration efficiency and slow model convergence speed in direct training schemes of related technologies, ensuring the training efficiency and control accuracy of the reinforcement learning model. Through the above reinforcement learning technology, leveraging its core advantages of being model-free and autonomously learning and optimizing, the agent gradually learns the optimal decision-making strategy through real-time interaction with the environment, without the need for pre-setting precise models or empirical rule bases, and can dynamically adjust the decision logic according to environmental changes. Applying reinforcement learning technology to active or semi-active suspension control can effectively overcome the dependence of related control algorithms on model accuracy and empirical rules, while possessing the potential to dynamically coordinate multi-objective conflicts. This can fully unleash the dynamic adjustability of active or semi-active suspensions, providing a new technical path for achieving a dynamic balance of the three core requirements of suspension systems under all operating conditions.

[0139] In summary, this application constructs a vehicle suspension dynamics simulation model to reproduce the suspension structure and mechanical characteristics, serving as the basis for subsequent simulation verification. The suspension model and control model are co-simulated to simulate various road conditions, and iterative calculations output suspension operating parameter schemes adapted to the scenarios. The control model is trained and optimized based on simulation data until indicators such as control accuracy and response speed meet preset standards. The qualified model is deployed to a real vehicle, and optimized parameters are called in real time to adjust suspension performance and improve the vehicle's core driving indicators.

[0140] This application also provides a training device for a vehicle suspension control model, used to execute the training method for the aforementioned vehicle suspension control model.

[0141] Figure 7 This is a schematic diagram of the structure of a training device for a vehicle suspension control model provided in an embodiment of this application. Figure 7 As shown, in some embodiments, the training device 500 for the vehicle suspension control model includes a reward calculation module 501 and a model training module 502.

[0142] The reward calculation module 501 is configured to determine the reward value based on the operating parameters of the training vehicle and driving environment data. For example, the reward value is determined based on the operating parameters of the training vehicle and driving environment data output from the vehicle suspension simulation model.

[0143] The model training module 502 is configured to train the suspension control model based on the operating parameters of the training vehicle, with the reward value as the training constraint direction, and to determine the model parameters and suspension action parameters. The suspension control model is used to control the vehicle suspension to perform adjustment operations. For example, the operating parameters of the training vehicle and the reward value output from the vehicle suspension simulation model are input into the suspension control model to determine the model parameters and first suspension action parameters of the suspension control model. Furthermore, the model parameters of the suspension control model are iteratively updated based on the first suspension action parameters.

[0144] In some embodiments, the training device 500 for the vehicle suspension control model further includes a suspension simulation module; the suspension simulation module is used to determine the operating parameters of the training vehicle through the vehicle suspension simulation model based on the preset control current, the first vehicle speed and the vehicle road surface generation results.

[0145] In some embodiments, the suspension simulation module is also used to determine updated training vehicle operating parameters based on the vehicle suspension control current, the second vehicle speed, and the vehicle road surface generation results, using the vehicle suspension simulation model.

[0146] In some embodiments, the training device 500 for the vehicle suspension control model further includes: a road surface generation module; the road surface generation module is used to randomly insert disturbance excitations according to the road surface roughness to generate vehicle road surface generation results; the disturbance excitations include at least one of bump excitations and hump excitations; the vehicle road surface generation results are used to characterize the road surface height sequence of the wheels; the greater the road surface roughness, the greater the difference between the road surface height sequences of each wheel.

[0147] In some embodiments, the reward calculation module 501 is further configured to determine a reward value based on the sum of the products of ride comfort weight and vehicle vibration parameters, operation stability weight and operation stability parameters, and durability weight and suspension load parameters; wherein, the higher the vehicle speed, the greater the ride comfort weight; the greater the steering wheel angle, the greater the operation stability weight; and the greater the road surface roughness, the greater the durability weight.

[0148] In some embodiments, the model training module 502 is further configured to determine the target reward value of the target vehicle road surface; and, if the target reward value meets the reward conditions, determine the target model parameters of the suspension control model corresponding to the target vehicle road surface.

[0149] In some embodiments, the suspension simulation module is also used to initialize the vehicle suspension parameters of the vehicle suspension simulation model.

[0150] In some embodiments, the training device 500 for the vehicle suspension control model further includes a parameter storage module for storing model parameters during the iterative update process and storing target model parameters when the reward value meets the conditions.

[0151] This application also provides a vehicle suspension control device for performing the above-described vehicle suspension control method.

[0152] Figure 8 This is a schematic diagram of a vehicle suspension control device provided in an embodiment of this application. Figure 8 As shown, in some embodiments, the vehicle suspension control device 600 includes a sensor group 601 and an on-board control unit 602.

[0153] Sensor group 601 is used to acquire vehicle operating parameters in response to vehicle suspension adjustment trigger events. Sensor group 601 can periodically collect driving environment data in real time. For example, sensor group 601 includes, but is not limited to, lidar and road perception sensors.

[0154] The on-board control unit 602 is used to determine vehicle suspension control parameters based on vehicle operating parameters and a vehicle suspension control model. For example, the on-board control unit 602 internally stores and runs the vehicle suspension control model. The on-board control unit 602 receives driving environment data from the sensor group 601, inputs the driving environment data into the vehicle suspension control model, and determines the vehicle suspension control strategy, including the vehicle suspension control parameters.

[0155] The vehicle suspension control model can be trained based on the vehicle suspension control method provided in the embodiments of this application; the vehicle suspension is controlled to perform adjustment operations according to the vehicle suspension control parameters.

[0156] In some embodiments, the vehicle control unit 602 is further configured to determine the suspension control evaluation result based on the collected real-time vehicle operating parameters under the vehicle suspension control parameters.

[0157] In some embodiments, the vehicle control unit 602 is also used to monitor vehicle suspension adjustment trigger events.

[0158] In some embodiments, the vehicle control unit 602 is further configured to control the vehicle damping force according to the vehicle suspension control current, so as to dynamically suppress the vertical vibration, pitch and roll motion of the vehicle body.

[0159] This application also provides a vehicle for executing the training method of the above-described vehicle suspension control model or the above-described suspension control method.

[0160] Figure 9 This is a schematic diagram of the structure of a vehicle provided in an embodiment of this application. Figure 9 As shown, in some embodiments, vehicle 700 can be a hybrid vehicle, a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. Vehicle 700 can be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0161] Reference Figure 9 The vehicle 700 may include various subsystems, such as an infotainment system 710, a perception system 720, a decision control system 730, a drive system 740, and a computing platform 750. The vehicle 700 may also include more or fewer subsystems, and each subsystem may include multiple components. Furthermore, each subsystem and each component of the vehicle 700 can be interconnected via wired or wireless means.

[0162] In some embodiments, the infotainment system 710 may include a communication system, an entertainment system, and a navigation system, etc.

[0163] The perception system 720 may include several sensors for sensing information about the environment surrounding the vehicle 700. For example, the perception system 720 may include a global positioning system (which may be GPS, BeiDou, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0164] The decision control system 730 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.

[0165] The drive system 740 may include components that provide powered motion to the vehicle 700. In one embodiment, the drive system 740 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of internal combustion engines, electric motors, and compressed air engines. The engine is capable of converting energy provided by the energy source into mechanical energy.

[0166] Some or all of the functions of vehicle 700 are controlled by computing platform 750. Computing platform 750 may include at least one processor 751 and memory 752, and processor 751 may execute instructions 753 stored in memory 752.

[0167] Processor 751 can be any conventional processor, such as a commercially available CPU. Processors may also include graphics processing units (GPUs), field-programmable gate arrays (FPGAs), systems-on-chips (SoCs), application-specific integrated circuits (ASICs), or combinations thereof.

[0168] The memory 752 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0169] In addition to instruction 753, memory 752 can also store data, such as road maps, route information, vehicle position, direction, speed, and other data. The data stored in memory 752 can be used by computing platform 750.

[0170] In this embodiment of the application, the processor 751 can execute instruction 753 to complete all or part of the steps of the above-described vehicle suspension control model training method or the above-described suspension control method.

[0171] In some embodiments, the vehicle includes a body and wheels, a suspension system, a vehicle suspension control device and a power control system provided in the above embodiments.

[0172] The suspension system includes multiple active vehicle suspension actuators and wheel suspension components, the wheel suspension components including shock absorbers, springs, and stiffness or damping adjustment mechanisms.

[0173] The on-board control unit of the vehicle suspension control device is connected to the suspension system, and the sensor group of the vehicle suspension control device is installed on the main body of the vehicle body and at the wheel positions.

[0174] The powertrain control system connects to the vehicle control unit to synchronize driving environment data such as vehicle speed and steering wheel angle to the vehicle control unit.

[0175] This embodiment also provides an electronic device for executing the training method of the vehicle suspension control model or the suspension control method described above.

[0176] Figure 10 This is a schematic diagram illustrating the structure of an electronic device according to some embodiments of this application. For example... Figure 10 As shown, the electronic device 800 includes at least one processor 820; and a memory 804 communicatively connected to the at least one processor 820; wherein the memory 804 stores instructions executable by the at least one processor 820, which, when executed by the at least one processor 820, enable the at least one processor to train a vehicle suspension control model.

[0177] Electronic device 800 can be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc. Furthermore, electronic device 800 can also be configured in vehicle form to meet the application requirements of vehicle suspension control.

[0178] In some embodiments, when the electronic device 800 is operating in vehicle mode, it can collect vehicle operating parameters in real time through sensors, and dynamically adjust the vehicle suspension control parameters based on these parameters to optimize the suspension control effect, thereby improving the ride comfort and handling stability of the vehicle during driving.

[0179] In other embodiments, the electronic device 800 can serve as a training device (e.g., a computer) for the vehicle suspension control model. It generates training vehicle operating parameters through simulation technology and uses these parameters to simulate the suspension's performance under different operating conditions in advance, providing data support and simulation basis for the training optimization of the vehicle suspension control model and the design iteration of the suspension structure.

[0180] Electronic device 800 may include one or more of the following components: processing component 802, memory 804, power component 806, multimedia component 808, audio component 810, input / output (I / O) interface 812, sensor component 814, and communication component 816.

[0181] Processing component 802 typically controls the overall operation of electronic device 800, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the methods described above. Furthermore, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

[0182] Memory 804 is configured to store various types of data to support the operation of electronic device 800. Examples of such data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0183] Power component 806 provides power to various components of electronic device 800. Power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800.

[0184] Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the electronic device 800 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0185] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when electronic device 800 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

[0186] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0187] Sensor assembly 814 includes one or more sensors for providing state assessments of various aspects of electronic device 800. For example, sensor assembly 814 can detect the on / off state of electronic device 800, the relative positioning of components such as the display and keypad of electronic device 800, changes in position of electronic device 800 or a component of electronic device 800, the presence or absence of user contact with electronic device 800, orientation or acceleration / deceleration of electronic device 800, and temperature changes of electronic device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 814 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0188] Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices. Electronic device 800 can access wireless networks based on communication standards, such as WiFi, 3G, 4G, 5G, other communication standards, or combinations thereof. In some embodiments of this application, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In some embodiments of this application, communication component 816 further includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0189] In some embodiments of this application, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0190] In some embodiments of this application, a non-transitory computer-readable storage medium storing computer instructions is also provided, such as a memory 804 including instructions that can be executed by a processor 820 of an electronic device 800 to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0191] In some embodiments of this application, a computer program product is also provided, including a computer program and computer instructions for causing a computer to perform the method described in any one of the first aspects.

[0192] In the above detailed description, reference has been made to the accompanying drawings, which illustrate specific aspects of how this application can be practiced. In this regard, terms indicating direction or positional relationship, such as “center,” “longitudinal,” “lateral,” “length,” “width,” “thickness,” “upper,” “lower,” “front,” “rear,” “left,” “right,” “vertical,” “horizontal,” “top,” “bottom,” “inner,” “outer,” “clockwise,” “counterclockwise,” “axial,” “radial,” and “circumferential,” can be used with reference to the orientation of the described figures. Since components of the described device can be positioned in multiple different orientations, directional terms are used for illustrative purposes and are not restrictive. It should be understood that other aspects can be utilized and structural or logical changes can be made without departing from the concept of this application. Therefore, the following detailed description should not be considered limiting.

[0193] It should be understood that, unless otherwise specifically indicated, features of various embodiments of this application described herein can be combined with each other. As used herein, the term “and / or” includes any one of the relevant listed items and any combination of any two or more; similarly, “at least one of…” includes any one of the relevant listed items and any combination of any two or more.

[0194] It should be understood that, unless otherwise expressly specified and limited, the terms "joining," "attaching," "installing," "connecting," "linking," and "fixing," as used in the embodiments of this application, should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms herein based on the specific circumstances.

[0195] Furthermore, the term "above" as used herein with respect to components, elements, or material layers formed or located "above" a surface may be used to indicate that the component, element, or material layer is "indirectly" positioned (e.g., placed, formed, deposited, etc.) on the surface such that one or more additional components, elements, or layers are arranged between the surface and the component, element, or material layer. However, the term "above" as used with respect to components, elements, or material layers formed or located "above" a surface may also optionally have a specific meaning: that the component, element, or material layer is "directly" positioned (e.g., placed, formed, deposited, etc.) on the surface, for example, in direct contact with the surface.

[0196] Although terms such as “first,” “second,” and “third” may be used herein to describe various components, parts, regions, layers, or sections, these components, parts, regions, layers, or sections are not limited to these terms. Rather, these terms are used only to distinguish one component, part, region, layer, or section from another. Therefore, without departing from the teachings of the examples described herein, the first component, part, region, layer, or section mentioned in the examples may also be referred to as the second component, part, region, layer, or section. Furthermore, the terms “first” and “second” are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as “first” or “second” may explicitly or implicitly include at least one of that feature. In the description herein, “a plurality” means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0197] It should be understood that spatial relative terms, such as “above,” “upper,” “below,” and “lower,” are used herein to describe the relationship between one element and another shown in the figures. In addition to the orientation depicted in the figures, these spatial relative terms are also intended to encompass different orientations of the device in use or operation. For example, if the device in the figures is flipped, an element described as “above” or “upper” relative to another element would be “below” or “lower” relative to that other element. Thus, depending on the spatial orientation of the device, the term “above” encompasses both above and below orientations. Devices may have other orientations (e.g., rotated 90 degrees or in other orientations), and the spatial relative terms used herein should be interpreted accordingly.

[0198] Furthermore, the term “exemplary” is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as “exemplary” is not necessarily to be construed as advantageous compared to other aspects or designs. Rather, the use of the term “exemplary” is intended to present the concept in a concrete manner. As used herein, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from the context, “X applies A or B” is intended to mean any of the natural inclusive arrangements. That is, “X applies A or B” satisfies any of the foregoing instances if X applies A; X applies B; or both X applies A and B. Additionally, unless otherwise specified or clear from the context to refer to the singular form, the articles “a” and “an” as used in this application and the appended claims are generally understood to mean “one or more.”

[0199] Similarly, although this application has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding the specification and drawings. This application includes all such modifications and variations and is limited only by the scope of the claims. In particular, with respect to the various functions performed by the components described above (e.g., elements, resources, etc.), unless otherwise indicated, the terminology used to describe such components is intended to correspond to any component (functionally equivalent) that performs the specific function of the described component, even if structurally not equivalent to the disclosed structure. Furthermore, although a particular feature of this application may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations, as may be desired and advantageous to any given or particular application. Moreover, with regard to the terms “comprising,” “owning,” “having,” “having,” or variations thereof as used in the Detailed Description or the claims, such terms are intended to be inclusive in a manner similar to the term “including.”

[0200] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0201] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A training method for a vehicle suspension control model, characterized in that, include: The reward value is determined based on the operating parameters of the training vehicle and the driving environment data. Based on the operating parameters of the training vehicle, the suspension control model is trained with the reward value as the training constraint direction to determine the model parameters and suspension action parameters. The suspension control model is used to control the vehicle suspension to perform adjustment operations.

2. The training method for the vehicle suspension control model according to claim 1, characterized in that, Also includes: Based on the preset control current, the first vehicle speed, and the generated road surface, the operating parameters of the training vehicle are determined through a vehicle suspension simulation model.

3. The training method for the vehicle suspension control model according to claim 2, characterized in that, The suspension action parameters include the vehicle suspension control current, and determining the operating parameters of the training vehicle includes: Based on the vehicle suspension control current, the second vehicle speed, and the vehicle road surface generation results, the updated operating parameters of the training vehicle are determined through the vehicle suspension simulation model.

4. The training method for the vehicle suspension control model according to claim 2 or 3, characterized in that, Also includes: Based on the road surface roughness, a disturbance excitation is inserted to generate the vehicle road surface generation result; The disturbance excitation includes at least one of turbulence excitation and hump excitation; The vehicle road surface generation results are used to characterize the road surface height sequence of the wheels on the target vehicle road surface; the greater the road surface roughness, the greater the difference between the road surface height sequences of each wheel.

5. The training method for the vehicle suspension control model according to claim 4, characterized in that, The driving environment data includes steering wheel angle, vehicle speed, and road surface roughness. The determination of the reward value based on the training vehicle's operating parameters and the driving environment data includes: The reward value is determined by summing the products of ride comfort weight and vehicle vibration parameters, handling stability weight and handling stability parameters, and durability weight and suspension load parameters. The higher the vehicle speed, the greater the weight of ride smoothness; the greater the steering wheel angle, the greater the weight of handling stability; and the greater the road surface roughness, the greater the weight of durability.

6. The training method for the vehicle suspension control model according to claim 5, characterized in that, The vehicle body vibration parameters include vertical acceleration; the operational stability parameters include at least one of roll angle acceleration and pitch angle acceleration; and the suspension load parameters include the travel of the wheel suspension shock absorbers.

7. The training method for the vehicle suspension control model according to claim 6, characterized in that, The determination of model parameters and suspension action parameters includes: If the reward value of the target vehicle road surface meets the reward conditions, the target model parameters of the suspension control model corresponding to the target vehicle road surface are determined.

8. The training method for the vehicle suspension control model according to claim 7, characterized in that, The conditions for fulfilling the reward include: The ratio of the reward value of the target vehicle road surface to the reward threshold is greater than a constant threshold; the reward threshold is set based on the baseline performance of the target vehicle road surface scenario.

9. The training method for the vehicle suspension control model according to claim 3, characterized in that, The method further includes: Initialize the vehicle suspension parameters of the vehicle suspension simulation model. The vehicle suspension parameters include at least one of the following: vehicle mass, moment of inertia, suspension stiffness, distance from the front and rear axles to the vehicle's center of gravity, and maximum and minimum damping coefficients.

10. A vehicle suspension control method, characterized in that, include: In response to a vehicle suspension adjustment trigger event, vehicle operating parameters are acquired; these vehicle operating parameters are used to characterize the suspension motion state. Based on the vehicle operating parameters, the vehicle suspension control parameters are determined using a vehicle suspension control model; the vehicle suspension control model is trained using the training method for the vehicle suspension control model according to any one of claims 1-9. Based on the vehicle suspension control parameters, the vehicle suspension is controlled to perform adjustment operations.

11. The vehicle suspension control method according to claim 10, characterized in that, The vehicle suspension adjustment trigger event includes at least one of the following: The vehicle suspension control model is actively invoked according to a preset cycle; The vehicle was detected to be traveling in a preset road condition scenario; the preset road condition scenario includes at least one of random road surface, obstacle road surface, continuous sloping road surface, and discrete pothole road surface. The driver's operation command is received; the operation command includes at least one of the following: acceleration command, braking command, steering command, and suspension mode switching command. Obtain road condition forecast information within a preset distance ahead of the vehicle; the road condition forecast information includes at least one of the following: changes in road slope, changes in road surface material, and construction section warnings.

12. The vehicle suspension control method according to claim 10 or 11, characterized in that, The vehicle suspension control parameters include the vehicle suspension control current; the step of controlling the vehicle suspension to perform adjustment operations based on the vehicle suspension control parameters includes: Based on the vehicle suspension control current, the damping force of each shock absorber in the vehicle suspension is controlled to dynamically suppress the vertical vibration, pitch and roll motion of the vehicle body.

13. The vehicle suspension control method according to claim 10 or 11, characterized in that, Also includes: While controlling the vehicle suspension to perform adjustment operations, the adjusted vehicle operating parameters are obtained; Based on the preset parameter thresholds and the adjusted vehicle operating parameters, the suspension control evaluation results are obtained; The suspension control model is updated based on the suspension control evaluation results.

14. A training device for a vehicle suspension control model, characterized in that, include: The reward calculation module is configured to determine the reward value based on the operating parameters of the training vehicle and driving environment data; The model training module is configured to train the suspension control model according to the operating parameters of the training vehicle and with the reward value as the training constraint direction, and to determine the model parameters and suspension action parameters. The suspension control model is used to control the vehicle suspension to perform adjustment operations.

15. A vehicle suspension control device, characterized in that, include: A sensor array is used to acquire vehicle operating parameters in response to a vehicle suspension adjustment trigger event; the vehicle operating parameters are used to characterize the suspension motion state. The vehicle control unit is used to determine vehicle suspension control parameters based on the vehicle operating parameters and through a vehicle suspension control model; the vehicle suspension control model is trained using the training method for the vehicle suspension control model according to any one of claims 1-9; and controls the vehicle suspension to perform adjustment operations based on the vehicle suspension control parameters.

16. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the method of any one of claims 1-9 or any one of claims 10-13.

17. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-9 or any one of claims 10-13.

18. A program product, characterized in that, Includes computer instructions for causing a computer to perform the method of any one of claims 1-9 or any one of claims 10-13.

Images