Vehicle control method and vehicle

Abstract

The present disclosure relates to the technical field of vehicle intelligent driving, and provides a vehicle control method and a vehicle. In response to monitoring that a vehicle brakes, vehicle acceleration data and wheel speed are acquired, a target disturbance intensity is determined according to the vehicle acceleration data and the wheel speed; it is determined that the target disturbance intensity meets a preset adjustment condition, and vehicle kinetic energy recovery data are acquired; a target recovery intensity is determined according to the target disturbance intensity, the kinetic energy recovery data and the vehicle acceleration data, and the vehicle is controlled to perform energy recovery according to the target recovery intensity. In the present disclosure, when the recovery intensity of energy recovery is determined, the target disturbance intensity of the vehicle at present, the actual running information of the vehicle and the kinetic energy recovery data when the vehicle recovers power are comprehensively considered, and the recovery intensity is more accurate. At the same time, the instability caused by kinetic energy recovery is dynamically inhibited, the imbalance of wheel speed and the decline of vehicle stability caused by recovery control rigidity are avoided, and the consistency of wheel speed and the safety of the vehicle are ensured.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent driving technology for vehicles, and in particular to a vehicle control method and a vehicle. Background Technology

[0002] With the rapid development of vehicle technology, vehicles have become an important means of transportation in people's daily lives. Kinetic energy recovery control systems convert some of the kinetic energy generated during braking or coasting into electrical energy and store it back in the battery, thereby improving energy efficiency and extending driving range.

[0003] Existing vehicle kinetic energy recovery control systems mainly adjust power recovery based on a few characteristic signals such as driver braking requests and vehicle speed. Their control logic is relatively simple and lacks the ability to effectively perceive and dynamically identify complex road disturbance environments. As a result, the energy recovery strategy cannot dynamically adapt to the actual road conditions during the adjustment process, which undermines the longitudinal stability of the vehicle. Summary of the Invention

[0004] In view of this, the purpose of this disclosure is to propose a vehicle control method and a vehicle to solve the problem that current vehicle kinetic energy recovery control systems cannot dynamically adapt to the actual road conditions during the adjustment process, resulting in a mismatch between the recovered torque output and the actual adhesion conditions.

[0005] To achieve the above objectives, a first aspect of this disclosure provides a vehicle control method, the method comprising:

[0006] In response to the detection of vehicle braking, vehicle acceleration data and wheel speed are acquired, and the target disturbance intensity is determined based on the vehicle acceleration data and wheel speed. Based on the target disturbance intensity, determine whether the preset adjustment conditions are met, and obtain vehicle kinetic energy recovery data; The target recovery intensity is determined based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data, and the vehicle is controlled to recover energy based on the target recovery intensity.

[0007] Based on the same inventive concept, a second aspect of this disclosure provides a vehicle control device, comprising: The data acquisition module is configured to acquire vehicle acceleration data and wheel speed in response to detecting vehicle braking, and determine the target disturbance intensity based on the vehicle acceleration data and wheel speed. The judgment module is configured to determine whether the preset adjustment conditions are met based on the target disturbance intensity, and to obtain vehicle kinetic energy recovery data; The vehicle control module is configured to determine the target recovery intensity based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data, and control the vehicle to perform energy recovery based on the target recovery intensity.

[0008] Based on the same inventive concept, a third aspect of this disclosure proposes an electronic device including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the vehicle control method as described above when executing the computer program.

[0009] Based on the same inventive concept, a fourth aspect of this disclosure provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to perform the vehicle control method as described above.

[0010] Based on the same inventive concept, the fifth aspect of this disclosure provides a vehicle including the vehicle control device described in the second aspect, the electronic device described in the third aspect, or the storage medium described in the fourth aspect.

[0011] As can be seen from the above, this disclosure proposes a vehicle control method and a vehicle. When vehicle braking is detected, vehicle acceleration data and wheel speed are acquired. A target disturbance intensity is determined based on the vehicle acceleration data and wheel speed, where the target disturbance intensity represents the impact of the road surface on the vehicle's energy recovery. If the target disturbance intensity is determined to meet preset adjustment conditions, it indicates that the vehicle is frequently subjected to high-frequency local impact interference. To ensure driving stability during the energy recovery process, the energy recovery intensity needs to be adjusted. Vehicle kinetic energy recovery data is acquired, and the target recovery intensity is determined based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data. In determining the energy recovery intensity, the current target disturbance intensity, actual vehicle operating information, and kinetic energy recovery data during kinetic energy recovery are comprehensively considered, resulting in a more accurate recovery intensity. Simultaneously, the instability caused by kinetic energy recovery is dynamically suppressed, avoiding wheel speed imbalance and decreased vehicle stability due to rigid recovery control, ensuring consistent wheel speed and vehicle safety. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in this disclosure or related technologies, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of a vehicle control method according to an embodiment of the present disclosure; Figure 2 This is a structural block diagram of a vehicle control device according to an embodiment of the present disclosure; Figure 3This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0015] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0016] The following are definitions of terms used in this disclosure: Cross-Spectral Mixer (CSM) is a lightweight network architecture specifically designed for extracting spectral features from mixed vibration signals. Its core idea is to efficiently extract local and global spectral patterns from different frequency bands of vibration signals through a multi-branch parallel structure. While maintaining a low parameter count, CSM can effectively model multi-scale, cross-band dependencies in the vibration spectrum, significantly improving feature discriminative and generalization performance in vibration analysis tasks such as fault diagnosis and condition monitoring.

[0017] MTRN (Multi-Tensor Recovery Net) is a deep learning model for multimodal image restoration. Its core idea is to improve overall reconstruction quality by collaboratively restoring multiple related but damaged image modalities (such as RGB images, depth maps, and near-infrared images). This model typically employs a multi-branch encoder-decoder architecture, extracting features from each modality separately. It achieves intermodal information complementarity and guidance through cross-modal attention mechanisms, feature fusion modules, or shared latent representations. This allows it to leverage the structural correlations between multiple modalities to achieve more robust and refined reconstruction results than single-modal restoration in tasks such as denoising, deblurring, super-resolution, or missing region repair.

[0018] FMN (Feedback Modulation Network) is a neural network architecture that introduces a feedback modulation mechanism to enhance the model's adaptability to context or historical states. This network typically includes a forward propagation path and a feedback path, where feedback signals originate from deep or output layers and are used to dynamically modulate feature representations or parameters in shallow or intermediate layers. This modulation can be feature-level weighting, channel attention adjustments, or slight variations in convolutional kernel parameters, allowing the network to flexibly adjust its internal representations based on the correlation between the current input and historical information, thereby improving performance in tasks such as time-series data, video analysis, and incremental learning. The core idea of ​​FMN is to mimic the feedback mechanism of biological neural systems to achieve a more flexible and context-aware computational process.

[0019] DDPG: DDPG (Deep Deterministic Policy Gradient) is a deep reinforcement learning algorithm for continuous action spaces. It adopts an Actor-Critic architecture, in which the Actor network directly outputs deterministic continuous actions (such as specific retrieval intensity coefficients), the Critic network evaluates the value (Q-value) of the action, and the training process is stabilized by empirical replay with the target network.

[0020] SAC: SAC (Soft Actor-Critic) is a continuous action control algorithm based on the maximum entropy reinforcement learning framework. While optimizing the cumulative reward, it maximizes the entropy of the policy (i.e., the randomness of the action), thereby encouraging the agent to explore different behaviors more fully during training and avoiding premature convergence to a suboptimal policy.

[0021] GRU (Gated Recurrent Unit) is a variant of recurrent neural network (RNN) designed to address the vanishing or exploding gradient problems that traditional RNNs often encounter when processing long sequences of data. By introducing two gating mechanisms—an update gate and a reset gate—it effectively controls the fusion of forgotten historical information with the current input, thereby enabling it to capture long-term dependencies in time-series data.

[0022] With the rapid development of vehicle technology, vehicles have become an important means of transportation in people's daily lives. The kinetic energy recovery control system converts some of the kinetic energy during vehicle braking or coasting into electrical energy and stores it back in the battery, thereby improving energy utilization efficiency and extending driving range. At the same time, it achieves a smooth braking feel by coordinating the motor's counter-torque, helping to maintain vehicle driving stability.

[0023] Existing kinetic energy recovery control systems mainly adjust power recovery based on a few characteristic signals such as driver braking requests, vehicle speed, and motor operating status. Their control logic is relatively simple and lacks the ability to effectively perceive and dynamically identify complex road disturbance environments.

[0024] Especially in harsh conditions such as unstructured gravel roads, where numerous discrete obstacles exist on the road surface, vehicles are continuously subjected to high-frequency, unstable local impact disturbances during driving. Since the system relies solely on conventional sensor information, it struggles to accurately distinguish between steady-state driving conditions and transient disturbance scenarios. This results in the energy recovery strategy failing to dynamically adapt to the actual road conditions during adjustment, easily leading to a mismatch between the recovered torque output and the actual adhesion conditions. This causes inconsistent wheel speeds between the drive and driven wheels, imbalances in traction forces across the wheels, and ultimately compromises the vehicle's longitudinal stability. In severe cases, it can even trigger the electronic stability system, resulting in unexpected yaw intervention and affecting ride comfort and safety.

[0025] Traditional control strategies are mostly based on fixed thresholds or simple filtering, which have limited ability to identify the spectral characteristics of road disturbances and lack the ability to analyze and classify disturbance components in different frequency bands. Furthermore, due to the lack of a dynamic consistency prediction mechanism for wheel speed, the system cannot predict and suppress potential slippage trends in the feedforward stage. This results in a highly rigid response from the recovery system, with a small dynamic adjustment margin, making it prone to control lag or overshoot under rapidly changing disturbance excitation, posing certain safety hazards.

[0026] Therefore, there is an urgent need to develop intelligent energy recovery control methods with capabilities for road environment perception, online disturbance spectrum identification, and wheel speed cooperative prediction, in order to adapt to the stability and energy efficiency balance requirements under complex and variable driving environments. To address the above issues, this embodiment proposes a vehicle control method, such as... Figure 1 As shown, the method includes: Step 101: In response to detecting vehicle braking, acquire vehicle acceleration data and wheel speed, and determine the target disturbance intensity based on the vehicle acceleration data and wheel speed.

[0027] In practice, the vehicle's driving status is monitored. If braking is detected, it indicates that the vehicle is decelerating and braking, and energy recovery can be performed. The energy recovery is a technology that converts the kinetic energy of the vehicle during deceleration into electrical energy and stores it for reuse.

[0028] In this embodiment, the determination of whether the vehicle is braking can be made by the vehicle's operating speed, wheel speed, brake pedal opening, or accelerator pedal opening. For example, if the brake pedal opening increases, it indicates that the vehicle is braking. If the vehicle speed decreases, it also indicates that the vehicle is braking.

[0029] The target disturbance intensity is determined based on the vehicle acceleration data and the wheel speed, whereby the target disturbance intensity represents the degree of influence of the road surface on the vehicle's energy recovery. In this embodiment, the target disturbance intensity is in vector form, and the peak response point in the target disturbance intensity represents the location of the high-frequency coupled disturbance caused by gravel impact on the local structure of the vehicle. This can be used to finely constrain the energy recovery release path and amplitude, improving the system stability control capability when driving on gravel roads.

[0030] Step 102: Determine whether the preset adjustment conditions are met based on the target disturbance intensity, and obtain vehicle kinetic energy recovery data.

[0031] In practice, the system determines whether preset adjustment conditions are met based on the target disturbance intensity. These preset adjustment conditions are pre-defined conditions that require adjustment of the vehicle's energy recovery intensity. If the target disturbance intensity is determined to meet the preset adjustment conditions, vehicle kinetic energy recovery data is acquired. This data represents the real-time state of the vehicle kinetic energy recovery control system, specifically including target recovery power, motor feedback request torque, etc.

[0032] Step 103: Determine the target recovery intensity based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data; control the vehicle to perform energy recovery based on the target recovery intensity.

[0033] In specific implementation, after determining the target disturbance intensity and obtaining kinetic energy recovery data, the target recovery intensity is determined based on the target disturbance intensity, the kinetic energy recovery data and the vehicle acceleration data, wherein the vehicle acceleration data includes vehicle longitudinal acceleration, vehicle lateral acceleration and vehicle pitch acceleration.

[0034] After determining the target recovery intensity, the vehicle is controlled to recover energy based on the target recovery intensity. The target recovery intensity represents the level of energy recovery, meaning that based on the target recovery intensity, the motor is automatically controlled to convert into a generator, converting some kinetic energy into electrical energy and storing it back in the battery.

[0035] The above scheme, when vehicle braking is detected, acquires vehicle acceleration data and wheel speed. Based on this data, a target disturbance intensity is determined, whereby the target disturbance intensity represents the impact of the road surface on the vehicle's energy recovery. If the target disturbance intensity meets preset adjustment conditions, it indicates that the vehicle is frequently subjected to high-frequency local impact interference. To ensure driving stability during energy recovery, the energy recovery intensity needs to be adjusted. Vehicle kinetic energy recovery data is acquired, and the target recovery intensity is determined based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data. This determination of the energy recovery intensity comprehensively considers the current target disturbance intensity, actual vehicle operating information, and kinetic energy recovery data during kinetic energy recovery, resulting in a more accurate recovery intensity. Simultaneously, it dynamically suppresses instability caused by kinetic energy recovery, avoiding wheel speed imbalance and decreased vehicle stability due to rigid recovery control, ensuring consistent wheel speed and vehicle safety.

[0036] In some embodiments, when determining the target disturbance intensity based on vehicle acceleration data and wheel speed, the target disturbance spectrum is first determined based on the vehicle acceleration data, and then the target disturbance intensity is determined only if the target disturbance spectrum is greater than a preset spectrum threshold, in order to reduce subsequent useless recovery intensity adjustments. That is, determining the target disturbance intensity based on the vehicle acceleration data and wheel speed in step 101 specifically includes: Step 1011: Based on the vehicle acceleration data, determine the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum; Step 1012: In response to the target disturbance spectrum being greater than a preset spectrum threshold, the target disturbance intensity is determined based on the target disturbance spectrum and the wheel speed.

[0037] In practice, based on the vehicle acceleration data, a target disturbance spectrum and a corresponding first confidence value are determined for the road segment where the vehicle is located. The target disturbance spectrum is used to determine whether the road segment where the vehicle is located affects stable vehicle operation; that is, whether the road segment contains a type of disturbance that could affect stable vehicle operation. The first confidence value indicates the accuracy of the target disturbance spectrum determined based on the vehicle acceleration data.

[0038] A preset spectrum threshold is obtained, and the target disturbance spectrum is compared with the preset spectrum threshold. If it is determined that the target disturbance spectrum is greater than the preset spectrum threshold, it indicates that there is a disturbance in the road segment where the vehicle is located. Then, the intensity of the target disturbance is determined based on the target disturbance spectrum and the wheel speed.

[0039] Specifically, the vehicle acceleration data includes longitudinal acceleration, lateral acceleration, and pitch acceleration. Step 1011, which involves determining the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum based on the vehicle acceleration data, specifically includes: Step 10111: Perform Fourier transform on the longitudinal acceleration, the lateral acceleration and the pitch acceleration to obtain the target acceleration spectrum; Step 10112: Input the target acceleration spectrum into the pre-trained disturbance recognition model, and process it through the disturbance recognition model to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum.

[0040] In practice, the longitudinal acceleration, lateral acceleration, and pitch acceleration are subjected to Fourier transforms to obtain the target acceleration spectrum. The longitudinal acceleration is the acceleration in the vehicle's forward or backward direction. The lateral acceleration is the acceleration perpendicular to the longitudinal direction, in the left-right direction. The pitch acceleration is the angular acceleration about the vehicle's lateral axis, causing the vehicle's nose to nod.

[0041] Specifically, the longitudinal acceleration, lateral acceleration, and pitch acceleration are subjected to short-time Fourier transforms to obtain the complex spectrum of the triaxial vibration signal after resampling, i.e., the target acceleration spectrum. The purpose of the short-time Fourier transform is to segment the non-stationary signal along the time axis, perform a Fourier transform on each segment, and thus obtain a spectrum showing how the signal frequency components change over time.

[0042] After determining the target acceleration spectrum, the target acceleration spectrum is input into a pre-trained disturbance recognition model. Through processing by the disturbance recognition model, the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum are obtained.

[0043] Specifically, the training process of the disturbance recognition model includes: Step a, obtain the first training dataset and the initial perturbation recognition model, wherein the first training dataset includes historical acceleration spectrum, historical perturbation spectrum and historical confidence value.

[0044] Step b: Input the training data in the first training dataset into the initial perturbation recognition model for training, determine that the first preset training termination condition is met, and obtain the perturbation recognition model.

[0045] In specific implementation, a first training dataset and an initial perturbation recognition model are obtained, wherein the first training dataset includes historical acceleration spectra, historical perturbation spectra, and historical confidence values. Training data from the first training dataset is input into the initial perturbation recognition model for training. Once a first preset training termination condition is met, the perturbation recognition model is obtained.

[0046] The first preset training termination condition includes at least one of the following: determining that all data in the first training dataset has been input into the initial perturbation recognition model for training, determining that the loss function of the initial perturbation recognition model has converged to a first convergence threshold, or determining that the initial perturbation recognition model has been iteratively trained to a first preset number of iterations.

[0047] For example, the first preset training termination condition is to determine that all data in the first training dataset has been input into the initial perturbation recognition model for training: The first training dataset contains fifty sets of data, each set including historical acceleration spectrum, historical perturbation spectrum, and historical confidence value. The first preset training termination condition is determined when all data in the first training dataset has been input into the initial perturbation recognition model for training. That is, when all fifty sets of data have been input into the initial perturbation recognition model, there is no training data in the first training dataset that has not yet been input into the initial perturbation recognition model, and the initial perturbation recognition model is determined to be trained and the perturbation recognition model is obtained.

[0048] In another example, the first preset training termination condition is to determine that the loss function of the initial perturbation recognition model converges to a first convergence threshold: The training data from the first training dataset is input into the initial perturbation recognition model for training, and the training results are output. A loss function is determined based on the training results and the actual perturbation spectrum. The loss function may include at least one of the following: mean squared error loss function, cross-entropy loss function, logarithmic loss function, exponential loss function, squared loss function, or absolute value loss function, etc. When the loss function converges to a first convergence threshold, a first preset training termination condition is satisfied, and the perturbation recognition model is obtained.

[0049] In another example, the first preset training termination condition is to determine the initial perturbation recognition model to be iterated and trained to the first preset number of iterations.

[0050] The training data in the first training dataset is input into the initial perturbation recognition model for iterative training. The number of iterations is recorded. When the number of iterations is equal to the first preset number of iterations, the first preset training termination condition is met, and the perturbation recognition model is obtained.

[0051] Specifically, the disturbance identification model includes a spectrum mapping module, an attention module, and a hybrid module. Step 10112, which involves inputting the target acceleration spectrum into the pre-trained disturbance identification model and processing it to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum, specifically includes: Step A: Input the target acceleration spectrum into the pre-trained disturbance recognition model, and use the spectrum mapping module to encode the target acceleration spectrum to obtain the target spectrum vector; Step B: Identify the target frequency band activation spectrum based on the attention module; Step C: Obtain the attention weights corresponding to the attention module, and use the hybrid module to fuse the target frequency band activation spectrum and the attention weights to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum.

[0052] In specific implementation, the target acceleration spectrum is input into a pre-trained disturbance recognition model, and the target acceleration spectrum is encoded using a spectrum mapping module. The target acceleration spectrum is a three-axis vibration spectrum. Specifically, the spectrum mapping unit encodes the three-axis vibration spectrum tensor into intermediate representation vectors to maintain channel independence and avoid the loss of high-frequency disturbance features caused by early channel fusion.

[0053] The target frequency band activation spectrum is obtained by identifying the target spectral vector based on the attention module. That is, a self-attention mechanism across frequency band dimensions is used to identify the frequency range with the most interference effect in different time domain segments. In this embodiment, the focus is particularly on the region of concentrated vibration energy between 60Hz and 180Hz caused by gravel.

[0054] The attention weights corresponding to the attention module are obtained, and the target frequency band activation spectrum and the attention weights are fused using the hybrid module to obtain the fused spectrum embedding vector and the first confidence value. The spectrum embedding vector is the target disturbance spectrum of the road segment where the vehicle is located.

[0055] In this embodiment, the disturbance identification model is a neural network structure model. Preferably, the disturbance identification model is a hybrid vibration spectrum extraction network structure model, wherein the hybrid vibration spectrum extraction network structure is a CSM structure. The target acceleration spectrum is input into the disturbance identification model, and multi-scale frequency bands are extracted within each time window to form a complex spectral tensor of uniform size.

[0056] Specifically, the CSM network consists of a multi-channel spectrum mapping module, an attention module, and a hybrid module. First, the spectrum mapping module encodes the three-axis vibration spectrum tensor into intermediate representation vectors to maintain channel independence and avoid the loss of high-frequency gravel perturbation features caused by early channel fusion. Then, the attention module is introduced, designed as a cross-frequency band self-attention mechanism to identify the frequency ranges with the most perturbation effects in different time domains. Finally, the hybrid module takes the channel attention weights and the frequency band activation spectrum as input to construct a fused spectrum embedding vector, forming the target perturbation spectrum containing joint spatial and frequency domain features as the CSM output.

[0057] In this embodiment, the disturbance recognition model uses labeled road surface type data as supervision signals during the training phase. The gravel road section labels are generated based on the annotations of the visual road surface recognition system on the actual test vehicle. The training objective is to maximize the separability of the spectral embedding space between different road surfaces, while minimizing the feature fluctuations within samples under the same working condition. Furthermore, a large amount of real-world test data from vehicle manufacturers is incorporated into the training data construction, covering typical complex working conditions such as dry gravel, wet gravel, and mixed gravel sections with shallow potholes, ensuring that the model has the ability to generalize and recognize real disturbance patterns.

[0058] In some embodiments, when determining the target disturbance intensity, a disturbance tensor determination model can be used to analyze the target disturbance spectrum and wheel speed. That is, step 1012, which involves determining the target disturbance intensity based on the target disturbance spectrum and the wheel speed, specifically includes: Step 10121: Obtain the pre-trained perturbation tensor determination model; Step 10122: Input the target disturbance spectrum and the wheel speed into the disturbance tensor determination model, and process them through the disturbance tensor determination model to obtain the target disturbance intensity.

[0059] In practice, a pre-trained perturbation tensor determination model is obtained, and the target perturbation spectrum and the wheel speed are input into the perturbation tensor determination model. The target perturbation intensity is obtained by processing the perturbation tensor determination model.

[0060] Specifically, the target disturbance spectrum is used to determine whether the road segment where the vehicle is located affects the vehicle's stable operation; that is, whether the road segment contains a type of disturbance that could affect the vehicle's stable operation. In other words, the target disturbance spectrum reflects the frequency distribution of vibrations generated by the road surface on the vehicle. Different frequency components correspond to the response characteristics of different structural components. Wheel speed directly affects the time scale of the disturbance and the conversion from spatial frequency to time frequency; the same road surface unevenness will produce vibration signals of different frequencies at different vehicle speeds.

[0061] The perturbation tensor determination model first uses wheel speed to convert the perturbation spectrum in the time domain into a road surface roughness spectrum in the spatial domain, thereby removing the scaling effect of vehicle speed on frequency and obtaining the intrinsic features of road surface perturbation independent of vehicle speed. The perturbation tensor determination model extracts key features from the normalized spectrum, such as the location of energy-concentrated frequency bands, the presence of abnormal high-frequency peaks (corresponding to gravel impact), and whether the spectral shape matches known types of perturbation sections (e.g., gravel roads, pothole roads). Finally, the perturbation tensor determination model outputs the target perturbation intensity, which is in vector form. Each element corresponds to the perturbation response amplitude of different local structures of the vehicle (e.g., suspension, chassis, wheel hubs, bearings, etc.), and the dimension of the peak response point represents the part most severely affected by high-frequency coupling perturbation from gravel.

[0062] Because the same road surface unevenness will produce completely different time frequency distributions at different vehicle speeds, the disturbance tensor determination model comprehensively analyzes the wheel speed and the target disturbance spectrum, and corrects the time spectrum to the spatial domain by combining wheel speed, decoupling the essential disturbance characteristics of the road surface, thereby accurately judging the disturbance intensity of the road segment, and avoiding the problem that the spectrum alone cannot distinguish whether the high-frequency disturbance is caused by the impact of gravel on the road surface itself or by the frequency shift caused by the increase in vehicle speed.

[0063] In this embodiment, the perturbation tensor determination model is a neural network structure model. Preferably, the perturbation tensor determination model is a tensor modeling and reconstruction model (MTRN), which is used to identify the implicit perturbation patterns generated by the gravel road surface on the vehicle structure and establish a quantitative correlation with the wheel speed consistency loss, so as to realize the pre-intervention of the unstable behavior of the recycling system.

[0064] Specifically, the target disturbance spectrum and wheel speed determined by the disturbance identification model are input into the disturbance tensor determination model. By constructing a tensor representation method, the frequency domain features of the vibration signal and the dynamic change structure of the wheel speed in the time domain are integrated to form a multidimensional observation tensor for disturbance modeling. The multidimensional observation tensor is the target disturbance intensity.

[0065] In this embodiment, the MTRN internal structure adopts a multi-branch tensor residual encoder framework. The core mechanism involves encoding heterogeneous information from different sources into a low-rank tensor in the same perturbation mode space, and then recovering the perturbation amplitude in missing or weak response channels through a residual reconstruction structure. The encoder part uses a tensor-level weight sharing mechanism to ensure consistent perturbation mapping capability under different operating conditions, while avoiding misleading overall perturbation identification due to fluctuations in data from a single channel. The input tensor structure contains multiple channel axial dimensions within each window period, specifically including spectral channel dimensions, time sliding window dimensions, vehicle front and rear wheel structure dimensions, and joint vibration and wheel speed observation dimensions.

[0066] In this embodiment, during the training phase of the perturbation tensor determination model, wheel speed difference samples constructed from real road test data are used as supervision signals. Particular attention is paid to wheel speed abrupt changes caused by unilateral slippage in gravel road sections. A mapping label between perturbation intensity and wheel speed collapse is constructed, and minimizing the reconstruction error between the perturbation tensor and the actual wheel speed difference tensor is used as the optimization objective. Simultaneously, an incomplete data augmentation strategy is introduced, such as randomly masking some channel signals or simulating signal loss, to enhance the model's robustness to data gaps under real, complex acquisition conditions.

[0067] In some embodiments, to improve the accuracy of adjusting the recovery intensity, when determining whether the target disturbance intensity meets the preset adjustment conditions, the preset adjustment conditions need to consider not only the target disturbance intensity but also the first confidence value corresponding to the target disturbance spectrum. That is, step 102, determining whether the preset adjustment conditions are met based on the target disturbance intensity, specifically includes: Step 1021: Obtain the preset intensity threshold and the preset confidence threshold; Step 1022: In response to the first confidence value being greater than the preset confidence threshold and the target disturbance intensity being greater than the preset intensity threshold, it is determined that the preset adjustment conditions are met.

[0068] In practice, a preset intensity threshold and a preset confidence threshold are obtained, the first confidence value is compared with the preset confidence threshold, and the target disturbance intensity is compared with the preset intensity threshold.

[0069] If it is determined that the first confidence value is greater than the preset confidence threshold and the target disturbance intensity is greater than the preset intensity threshold, it indicates that the vehicle is indeed on a disturbed road surface that affects the stable driving of the vehicle. Therefore, it is determined that the preset adjustment conditions are met, and the energy recovery intensity is adjusted accordingly.

[0070] If the first confidence value is determined to be less than or equal to the preset confidence threshold, and / or the target disturbance intensity is less than or equal to the preset intensity threshold, it indicates that the vehicle is in a low disturbance or over-disclosure state, and the preset adjustment conditions are not met.

[0071] In some embodiments, the kinetic energy recovery data includes the target recovery power and the motor feedback requested torque, and the vehicle acceleration data includes longitudinal acceleration. When determining the target recovery intensity, it can be determined using a pre-trained recovery intensity adjustment model. Specifically, step 103, which involves determining the target recovery intensity based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data, includes: Step 1031: Obtain the pre-trained recovery intensity adjustment model and brake pedal opening; Step 1032: Input the target recovery power, the motor feedback request torque, the brake pedal opening and the longitudinal acceleration into the recovery intensity adjustment model, and process them through the recovery intensity adjustment model to obtain the target recovery intensity.

[0072] In practice, a pre-trained recovery intensity adjustment model and brake pedal opening are acquired. The kinetic energy recovery data includes the target recovery power and the motor feedback requested torque, and the vehicle acceleration data includes longitudinal acceleration. The target recovery power, the motor feedback requested torque, the brake pedal opening, and the longitudinal acceleration are then input into the recovery intensity adjustment model. The model processes these data to obtain the target recovery intensity.

[0073] In this embodiment, the target regenerative power directly reflects the desired amount of electrical power to be recovered and is the core reference for the regeneration intensity. The motor feedback request torque converts the target regenerative power into the actual negative torque generated by the motor, directly determining the magnitude of the regenerative braking force. The brake pedal opening reflects the driver's expectation of total deceleration and also implies their tolerance for the drag sensation caused by regeneration. The longitudinal acceleration is the actual observed vehicle deceleration, used for closed-loop correction to ensure that the current regeneration intensity matches the road conditions and vehicle status.

[0074] The regenerative braking intensity adjustment model first determines the initial regenerative braking intensity level based on the target regenerative power and the requested torque from the motor. Then, it adjusts the upper and lower limits of the regenerative braking intensity according to the brake pedal opening. For example, at a small opening, the upper limit is limited to avoid abruptness, while at a large opening, a higher intensity is allowed to meet deceleration requirements. Finally, longitudinal acceleration plays a crucial corrective role as a feedback quantity. If the actual deceleration is lower than the driver's expectation (calculated from the pedal opening), the regenerative braking intensity adjustment model can appropriately increase the regenerative braking intensity. Conversely, if the deceleration is too large or signs of instability appear, the regenerative braking intensity adjustment model quickly reduces the regenerative braking intensity to ensure driving safety.

[0075] The energy recovery intensity adjustment model learns the complex relationship between target energy recovery power, motor feedback requested torque, brake pedal opening and longitudinal acceleration, and optimal energy recovery intensity through training. This allows it to balance energy recovery efficiency, driver braking intention and vehicle longitudinal stability under different driving conditions, ultimately outputting a safe and comfortable target energy recovery intensity.

[0076] In this embodiment, the target recovery power, the motor feedback request torque, the brake pedal opening, and the longitudinal acceleration are first synchronized in time and normalized in amplitude before being input into the recovery intensity adjustment model, so as to ensure the continuity and interpretability of the feedback adjustment process.

[0077] Specifically, the training process of the recycling intensity adjustment model includes: Step 10a: Obtain the second training dataset and the initial recovery intensity adjustment model, wherein the second training dataset includes historical recovery power, historical motor feedback request torque, historical brake pedal opening, historical longitudinal acceleration, and historical recovery intensity.

[0078] Step 10b: Input the training data from the second training dataset into the initial retrieval intensity adjustment model for training, determine that the second preset training termination condition is met, and obtain the retrieval intensity adjustment model.

[0079] In specific implementation, a second training dataset and an initial recovery intensity adjustment model are obtained. The second training dataset includes historical recovery power, historical motor feedback request torque, historical brake pedal opening, historical longitudinal acceleration, and historical recovery intensity. The training data in the second training dataset is input into the initial recovery intensity adjustment model for training. Once a second preset training termination condition is met, the recovery intensity adjustment model is obtained.

[0080] The second preset training termination condition includes at least one of the following: determining that all data in the second training dataset has been input into the initial retrieval intensity adjustment model for training, determining that the loss function of the initial retrieval intensity adjustment model has converged to the second convergence threshold, or determining that the initial retrieval intensity adjustment model has been iterated for training to the second preset number of iterations.

[0081] For example, the second preset training termination condition is to determine that all data in the second training dataset has been input into the initial retrieval intensity adjustment model for training: The second training dataset contains fifty sets of data, each set including historical regenerative braking power, historical motor feedback request torque, historical brake pedal opening, historical longitudinal acceleration, and historical regenerative braking intensity. The second preset training termination condition is determined when all data in the second training dataset has been input into the initial regenerative braking intensity adjustment model. Specifically, when all fifty sets of data have been input into the initial regenerative braking intensity adjustment model, there is no longer any training data in the second training dataset that has not yet been input into the initial regenerative braking intensity adjustment model. At this point, the initial regenerative braking intensity adjustment model training is considered complete, and the regenerative braking intensity adjustment model is obtained.

[0082] In another example, the second preset training termination condition is to determine that the loss function of the model, which adjusts the initial retrieval intensity, converges to a second convergence threshold: The training data from the second training dataset is input into the initial retrieval intensity adjustment model for training, and the training results are output. A loss function is determined based on the training results and the actual retrieval intensity. The loss function may be of at least one of the following types: mean squared error loss function, cross-entropy loss function, logarithmic loss function, exponential loss function, squared loss function, or absolute value loss function, etc. When the loss function converges to a second convergence threshold, it is determined that the second preset training termination condition is met, and the retrieval intensity adjustment model is obtained.

[0083] In another example, the second preset training termination condition is to determine the initial retrieval intensity and adjust the model iterative training to the second preset number of iterations.

[0084] The training data in the second training dataset is input into the initial recycling intensity adjustment model for iterative training. The number of iterations is recorded. When the number of iterations is equal to the second preset number of iterations, the second preset training termination condition is met, and the recycling intensity adjustment model is obtained.

[0085] In this embodiment, the energy recovery intensity adjustment model is a neural network structure model. Preferably, the energy recovery intensity adjustment model is a feedback modulation network structure model. Exemplarily, the energy recovery intensity adjustment model is DDPG, SAC, GRU, etc. The energy recovery intensity adjustment model is used to continuously, controllably, and dynamically adjust the kinetic energy recovery output with stability constraints after the gravel road disturbance is explicitly modeled, fundamentally avoiding the problem of inconsistent wheel speeds induced by excessive energy recovery intervention.

[0086] Specifically, the regeneration intensity adjustment model employs a hierarchical feedback modulation architecture. The bottom layer is a disturbance sensing encoding unit, used to map road disturbance characteristics (such as wheel speed fluctuations caused by gravel) to a unified modulation factor. This modulation factor characterizes the maximum regeneration intervention amplitude that the system can safely withstand under the current road conditions, avoiding exacerbating wheel instability due to excessive regeneration torque. The middle layer is a regeneration response adjustment unit, which establishes an implicit mapping between regeneration output changes and wheel speed response based on the deviation relationship between historical regeneration commands and actual wheel speed feedback. Its core function is to determine whether the current regeneration intervention is amplifying the disturbance effect caused by low-adhesion road surfaces such as gravel, thereby deciding whether to suppress or slow down the rate of change of regeneration intensity. The upper layer is a stability constraint output unit, which calculates the final regeneration suppression signal based on the modulation factor generated by the bottom layer and the regeneration response state evaluated by the middle layer. This signal continuously reduces or mitigates the target current or feedback torque request of the kinetic energy recovery system, rather than using a traditional step-off shutdown strategy, thereby achieving flexible adjustment of the regeneration intensity and ensuring the vehicle's ride comfort and safety on unstable road surfaces.

[0087] In this embodiment, the training dataset for the recovery intensity adjustment model includes historical recovery power, historical motor feedback request torque, historical brake pedal opening, historical longitudinal acceleration, and historical recovery intensity. Historical recovery power, historical motor feedback request torque, historical brake pedal opening, and historical longitudinal acceleration are used as target input features, and the corresponding historical recovery intensity is used as the target output label. The recovery intensity adjustment model learns the mapping from target input features to target output labels. The loss function typically uses mean squared error (MSE), and an additional wheel speed difference penalty term can be added to guide the recovery intensity adjustment model to output a smoother recovery intensity under perturbation conditions. During training, the recovery intensity adjustment model uses mini-batch stochastic gradient descent, updates network parameters through backpropagation, and monitors the correlation between recovery intensity prediction error and actual wheel speed stability on the validation set to prevent overfitting.

[0088] In this embodiment, the training of the energy recovery intensity adjustment model uses real-vehicle road test data from automakers as the core sample source, focusing on selecting boundary samples in gravel road sections where wheel speeds exhibit transient separation but have not yet triggered intervention from traditional stability control systems (such as ABS / ESC). These samples enhance the model's ability to identify early signs of instability. The training objective focuses on minimizing the magnitude and duration of wheel speed differences between the left and right wheels or between the drive and driven wheels during energy recovery intervention, without significantly reducing energy recovery efficiency. This ensures the energy recovery system exhibits flexible and gradual output characteristics under gravel road conditions, rather than abrupt intervention.

[0089] In some embodiments, if it is determined that the preset adjustment conditions are not met based on the target disturbance intensity, it means that there is no need to adjust the energy recovery intensity using vehicle acceleration data, wheel speed, and vehicle kinetic energy recovery data. In this case, the target recovery intensity can be determined based on the brake pedal opening and vehicle speed. That is, after determining the target disturbance intensity based on the vehicle acceleration data and wheel speed in step 101, the method further includes: Step 10A: Determine whether the preset adjustment conditions are not met based on the target disturbance intensity, and obtain the brake pedal opening, vehicle speed and vehicle battery status. Step 10B: Determine the target recovery intensity based on the brake pedal opening, the vehicle speed, and the vehicle battery status, and control the vehicle to perform energy recovery based on the target recovery intensity.

[0090] In specific implementation, if it is determined that the preset adjustment conditions are not met based on the target disturbance intensity, that is, the first confidence value is less than or equal to the preset confidence threshold, and / or the target disturbance intensity is less than or equal to the preset intensity threshold, then there is no need to adjust the energy recovery intensity through vehicle acceleration data, wheel speed and vehicle kinetic energy recovery data.

[0091] The system acquires brake pedal opening, vehicle speed, and vehicle battery status, including battery charge, battery temperature, and maximum rechargeable power. Based on the brake pedal opening, vehicle speed, and vehicle battery status, a target recovery intensity is determined, and the vehicle is controlled to perform energy recovery based on the target recovery intensity.

[0092] Specifically, the target recovery intensity is determined based on the brake pedal opening, the vehicle speed, and the vehicle battery status. The process of determining the target recovery intensity includes: The system searches a database based on vehicle speed and brake pedal opening to determine the target recovery torque corresponding to the vehicle speed and brake pedal opening. The database stores the correspondence between vehicle speed, brake pedal opening, and target recovery torque.

[0093] The maximum permissible regenerative torque is determined based on the vehicle speed and the battery's maximum rechargeable power. Specifically, the wheel angular velocity is determined based on the vehicle speed. The transmission efficiency is obtained by multiplying the battery's maximum rechargeable power by the transmission efficiency, and then the product value is compared with the wheel angular velocity to obtain the maximum permissible regenerative torque.

[0094] The target recovery torque is compared with the maximum permissible recovery torque. If the target recovery torque is less than or equal to the maximum permissible recovery torque, the battery capacity is sufficient to fully meet the driver's intention to brake smoothly. The target recovery intensity is then determined as the target recovery torque.

[0095] If the target recovery torque is greater than the maximum permissible recovery torque, and the deceleration required by the driver (such as emergency braking) exceeds the battery's receiving capacity, the recovery intensity will be limited to the battery's safe upper limit. That is, the target recovery intensity at this time is the maximum permissible recovery torque.

[0096] In this embodiment, by deploying an online data acquisition and annotation component on the mass-produced vehicle platform, signals such as acceleration, wheel speed, motor braking torque, recovery intensity adjustment command, and yaw rate during vehicle operation are collected in real time. Based on the internal stability estimation module, wheel speed collapse events, recovery instability intervention trigger points, and related disturbance characteristics and environmental conditions are automatically identified.

[0097] In this embodiment, all key operational data is cached in a local cache management system and uploaded to the vehicle manufacturer's cloud-based parameter tuning platform when the vehicle is connected to the network. The cloud system uses the latest uploaded vehicle data to automatically construct training subsets based on operating condition type and stability performance. These subsets are used for retraining the frequency band weights of the CSM spectrum selection mechanism, updating the tensor of the MTRN perturbation peak mapping space, and reconfiguring the policy bandwidth of the FMN modulation rules. The parameter tuning cycle for each model is adaptively set according to the sample trigger frequency; vehicles in areas with frequent gravel road conditions will receive more frequent model fine-tuning updates.

[0098] In this embodiment, a dynamic performance feedback mechanism is introduced to score the performance of the recovery intervention behavior and vehicle response results caused by each model output. This score is determined by the wheel speed consistency change rate, the magnitude of the disturbance residual value change, and the degree of energy recovery loss, and serves as the data selection basis for subsequent model fine-tuning. The vehicle control development platform deployed by the automaker supports prioritizing the injection of high-scoring operational data into the training process, ensuring that the model continuously learns from high-value operational feedback from real-world scenarios. Regarding strategy safety, the system sets up a dual-channel operation logic: one for the currently stable deployed model and the other for the fine-tuned experimental model. Replacement deployment is only performed when the latter continuously runs and meets the stability threshold, preventing model deviation from causing safety risks. Through this closed-loop reinforcement mechanism, the vehicle possesses long-term strategy evolution capabilities during use, significantly improving the intelligent stability control level in disturbed environments such as gravel roads, while maintaining optimal kinetic energy recovery efficiency under actual operating conditions.

[0099] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.

[0100] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0101] Based on the same inventive concept, corresponding to any of the above-described embodiments, this disclosure also provides a vehicle control device.

[0102] refer to Figure 2 , Figure 2 The vehicle control device, as described in this embodiment, includes: The data acquisition module 201 is configured to acquire vehicle acceleration data and wheel speed in response to detecting vehicle braking, and determine the target disturbance intensity based on the vehicle acceleration data and the wheel speed. The judgment module 202 is configured to determine whether the preset adjustment conditions are met based on the target disturbance intensity and to obtain vehicle kinetic energy recovery data; The vehicle control module 203 is configured to determine the target recovery intensity based on the target disturbance intensity, the kinetic energy recovery data and the vehicle acceleration data, and control the vehicle to perform energy recovery based on the target recovery intensity.

[0103] In some embodiments, the data acquisition module 201 is specifically configured as follows: Based on the vehicle acceleration data, determine the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum; In response to the target disturbance spectrum being greater than a preset spectrum threshold, the target disturbance intensity is determined based on the target disturbance spectrum and the wheel speed.

[0104] In some embodiments, the vehicle acceleration data includes longitudinal acceleration, lateral acceleration, and pitch acceleration, and the data acquisition module 201 is specifically configured to: Perform Fourier transforms on the longitudinal acceleration, the lateral acceleration, and the pitch acceleration to obtain the target acceleration spectrum; The target acceleration spectrum is input into a pre-trained disturbance recognition model. After processing by the disturbance recognition model, the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum are obtained.

[0105] In some embodiments, the data acquisition module 201 is specifically configured as follows: Obtain a first training dataset and an initial perturbation recognition model, wherein the first training dataset includes historical acceleration spectrum, historical perturbation spectrum and historical confidence values; The training data in the first training dataset is input into the initial perturbation recognition model for training. Once the first preset training termination condition is met, the perturbation recognition model is obtained.

[0106] In some embodiments, the disturbance identification model includes a spectrum mapping module, an attention module, and a hybrid module, and the data acquisition module 201 is specifically configured as follows: The target acceleration spectrum is input into a pre-trained disturbance recognition model, and the target acceleration spectrum is encoded using a spectrum mapping module to obtain a target spectrum vector. The target frequency band activation spectrum is obtained by identifying the target spectral vector based on the attention module; The attention weights corresponding to the attention modules are obtained, and the target frequency band activation spectrum and the attention weights are fused using the hybrid module to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum.

[0107] In some embodiments, the data acquisition module 201 is specifically configured as follows: Obtain the pre-trained perturbation tensor to determine the model; The target disturbance spectrum and the wheel speed are input into the disturbance tensor determination model, and the target disturbance intensity is obtained by processing the disturbance tensor determination model.

[0108] In some embodiments, the determination module 202 is specifically configured as follows: Obtain the preset intensity threshold and the preset confidence threshold; In response to the first confidence value being greater than the preset confidence threshold and the target disturbance intensity being greater than the preset intensity threshold, it is determined that the preset adjustment conditions are met.

[0109] In some embodiments, the kinetic energy recovery data includes the target recovery power and the motor feedback requested torque, the vehicle acceleration data includes longitudinal acceleration, and the vehicle control module 203 is specifically configured to: Obtain the pre-trained recovery intensity adjustment model and brake pedal opening; The target recovery power, the motor feedback request torque, the brake pedal opening, and the longitudinal acceleration are input into the recovery intensity adjustment model, and the target recovery intensity is obtained by processing the recovery intensity adjustment model.

[0110] In some embodiments, the apparatus further includes an intensity determination module, which is specifically configured to: Based on the target disturbance intensity, it is determined that the preset adjustment conditions are not met, and the brake pedal opening, vehicle speed, and vehicle battery status are obtained. The target recovery intensity is determined based on the brake pedal opening, the vehicle speed, and the vehicle battery status, and the vehicle is controlled to recover energy based on the target recovery intensity.

[0111] For ease of description, the above apparatus is described in terms of its functions, divided into various modules. Of course, in implementing this disclosure, the functions of each module can be implemented in one or more software and / or hardware.

[0112] The apparatus of the above embodiments is used to implement the corresponding vehicle control method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0113] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the vehicle control method described in any of the above embodiments.

[0114] Figure 3 This embodiment illustrates a more specific hardware structure of an electronic device. The device may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0115] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0116] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0117] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0118] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0119] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0120] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0121] The electronic devices described above are used to implement the corresponding vehicle control methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0122] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this disclosure also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to execute the vehicle control method as described in any of the above embodiments.

[0123] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0124] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the vehicle control method as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0125] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a vehicle, including the vehicle control device, the electronic device, and the computer-readable storage medium in the above embodiments, wherein the vehicle device implements the vehicle control method described in any of the above embodiments.

[0126] The vehicles described in the above embodiments are used to implement the vehicle control method described in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0127] It is understood that before using the technical solutions of the various embodiments in this disclosure, users will be informed of the type, scope of use, and usage scenarios of the personal information involved in an appropriate manner, and user authorization will be obtained.

[0128] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media performing the operations of this disclosed technical solution.

[0129] As an optional but not limited implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0130] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0131] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this disclosure as described above, which are not provided in detail for the sake of brevity.

[0132] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this disclosure, the provided drawings may or may not show well-known power / ground connections to integrated circuit (IC) chips and other components. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this disclosure, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this disclosure will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that the embodiments of this disclosure can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0133] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0134] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A vehicle control method, characterized in that, include: In response to the detection of vehicle braking, vehicle acceleration data and wheel speed are acquired, and the target disturbance intensity is determined based on the vehicle acceleration data and wheel speed. Based on the target disturbance intensity, determine whether the preset adjustment conditions are met, and obtain vehicle kinetic energy recovery data; The target recovery intensity is determined based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data, and the vehicle is controlled to recover energy based on the target recovery intensity.

2. The method according to claim 1, characterized in that, Determining the target disturbance intensity based on the vehicle acceleration data and the wheel speed includes: Based on the vehicle acceleration data, determine the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum; In response to the target disturbance spectrum being greater than a preset spectrum threshold, the target disturbance intensity is determined based on the target disturbance spectrum and the wheel speed.

3. The method according to claim 2, characterized in that, The vehicle acceleration data includes longitudinal acceleration, lateral acceleration, and pitch acceleration. The step of determining the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum based on the vehicle acceleration data includes: Perform Fourier transforms on the longitudinal acceleration, the lateral acceleration, and the pitch acceleration to obtain the target acceleration spectrum; The target acceleration spectrum is input into a pre-trained disturbance recognition model. After processing by the disturbance recognition model, the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum are obtained.

4. The method according to claim 3, characterized in that, The training process of the disturbance recognition model includes: Obtain a first training dataset and an initial perturbation recognition model, wherein the first training dataset includes historical acceleration spectrum, historical perturbation spectrum and historical confidence values; The training data in the first training dataset is input into the initial perturbation recognition model for training. Once the first preset training termination condition is met, the perturbation recognition model is obtained.

5. The method according to claim 3, characterized in that, The perturbation identification model includes a spectrum mapping module, an attention module, and a hybrid module. The step of inputting the target acceleration spectrum into a pre-trained disturbance recognition model, and processing it through the disturbance recognition model to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum, includes: The target acceleration spectrum is input into a pre-trained disturbance recognition model, and the target acceleration spectrum is encoded using a spectrum mapping module to obtain a target spectrum vector. The target frequency band activation spectrum is obtained by identifying the target spectral vector based on the attention module; The attention weights corresponding to the attention modules are obtained, and the target frequency band activation spectrum and the attention weights are fused using the hybrid module to obtain the target disturbance spectrum of the road segment where the vehicle is located and the first confidence value corresponding to the target disturbance spectrum.

6. The method according to claim 2, characterized in that, Determining the target disturbance intensity based on the target disturbance spectrum and the wheel speed includes: Obtain the pre-trained perturbation tensor to determine the model; The target disturbance spectrum and the wheel speed are input into the disturbance tensor determination model, and the target disturbance intensity is obtained by processing the disturbance tensor determination model.

7. The method according to claim 2, characterized in that, The step of determining whether the preset adjustment conditions are met based on the target disturbance intensity includes: Obtain the preset intensity threshold and the preset confidence threshold; In response to the first confidence value being greater than the preset confidence threshold and the target disturbance intensity being greater than the preset intensity threshold, it is determined that the preset adjustment conditions are met.

8. The method according to claim 1, characterized in that, The kinetic energy recovery data includes the target recovery power and the requested torque from the motor feedback; the vehicle acceleration data includes longitudinal acceleration. Determining the target recovery intensity based on the target disturbance intensity, the kinetic energy recovery data, and the vehicle acceleration data includes: Obtain the pre-trained recovery intensity adjustment model and brake pedal opening; The target recovery power, the motor feedback request torque, the brake pedal opening, and the longitudinal acceleration are input into the recovery intensity adjustment model, and the target recovery intensity is obtained by processing the recovery intensity adjustment model.

9. The method according to claim 1, characterized in that, After determining the target disturbance intensity based on the vehicle acceleration data and the wheel speed, the method further includes: Based on the target disturbance intensity, it is determined that the preset adjustment conditions are not met, and the brake pedal opening, vehicle speed, and vehicle battery status are obtained. The target recovery intensity is determined based on the brake pedal opening, the vehicle speed, and the vehicle battery status, and the vehicle is controlled to recover energy based on the target recovery intensity.

10. A vehicle, characterized in that, include: Memory, used to store executable programs; processor; When the executable program is executed by the processor, the method as described in any one of claims 1-9 is implemented.

Images