Steer-by-wire control method, device and computer device

Abstract

The present disclosure relates to a kind of drive-by-wire steering control method, device and computer device.The problem of response lag of drive-by-wire steering system is solved by involving drive-by-wire steering technology.The method comprises: obtaining the road surface sensing data of the preview area in the driving direction of the vehicle, the road surface sensing data at least includes the road surface image data and / or road surface three-dimensional point cloud data of the preview area;According to the road surface sensing data, extract road surface characteristic parameters, and construct road surface digital elevation preview model according to the road surface characteristic parameters, the road surface digital elevation preview model outputs the road surface excitation input time sequence of the preview area;According to the road surface excitation input time sequence and real-time vehicle driving state, generate feedforward compensation instruction, the feedforward control instruction is used to indicate that steering mechanism inhibits the harmful vibration caused by the preview area.The technical scheme provided by the present disclosure is suitable for intelligent driving, realizes the early active vibration prediction and timely intervention.

Description

Technical Field

[0001] This disclosure relates to steer-by-wire technology, and more particularly to a steer-by-wire control method, apparatus, and computer device. Background Technology

[0002] With the rapid development of automotive electrification and intelligence, steer-by-wire (SBW) systems have become a core component of advanced autonomous driving vehicles. SBW eliminates the mechanical connection between the steering wheel and the steering wheels, using electrical signals to transmit and execute steering commands. It offers advantages such as adjustable steering ratios, flexible vehicle layout, compatibility with advanced autonomous driving control, and good collision safety.

[0003] SBW's vibration control relies heavily on feedback signals such as rack force and steering motor status. The controller only performs vibration damping after the wheels have rolled over speed bumps, potholes, or bumpy roads, resulting in significant lag and causing steering wheel kickback and high-frequency vibration.

[0004] In summary, due to the lag in the response of the steer-by-wire system, the user's driving experience is poor. Summary of the Invention

[0005] To overcome the problems existing in related technologies, this disclosure provides a steer-by-wire control method, apparatus, and computer device. Based on a pre-aiming area, a road excitation input timing sequence is obtained to predict the future road surface, and a feedforward compensation command is generated on this basis to promptly generate a scheme to suppress upcoming harmful vibrations. This solves the problem of response lag in steer-by-wire systems, realizes proactive vibration prediction and timely intervention, and significantly improves driving experience and safety.

[0006] According to a first aspect of the embodiments of this disclosure, a steer-by-wire control method is provided, comprising: Acquire road perception data of a pre-aimed area in the vehicle's driving direction, wherein the road perception data includes at least road image data and / or three-dimensional point cloud data of the pre-aimed area; Based on the road surface perception data, road surface feature parameters are extracted, and a road surface digital elevation prediction model is constructed based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area. Based on the road excitation input timing sequence and the real-time vehicle driving status, a feedforward compensation command is generated. The feedforward control command is used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

[0007] Furthermore, the step of acquiring road perception data of the pre-aimed area in the vehicle's driving direction includes: The pre-aiming area is divided according to the vehicle's driving status in the driving direction; The road surface image data of the pre-aiming area is acquired using an image acquisition device; The three-dimensional point cloud data of the road surface in the pre-aiming area is obtained by lidar; Synchronize the road surface image data with the road surface 3D point cloud data in time and / or in space; The road surface image data and the road surface 3D point cloud data are fused to obtain the road surface perception data.

[0008] Furthermore, the step of extracting road feature parameters based on the road surface perception data and constructing a road surface digital elevation prediction model based on the road surface feature parameters includes: The road surface feature parameters are extracted from the road surface perception data, and the road surface feature parameters include at least any one or more of the following parameters: Road surface geometric features, road surface paving and adhesion material features, and road surface damage features; Establish a vehicle coordinate system with the center point of the front axle of the vehicle as the origin, the direction of travel as the X-axis, the left and right lateral directions of the vehicle as the Y-axis, and the vertical height of the road surface as the Z-axis. The road surface feature parameters are mapped to the vehicle coordinate system, and a road surface digital elevation prediction model covering the prediction area is constructed along the positive X-axis. The road surface digital elevation prediction model outputs the road surface excitation input time series sequence in the prediction time domain.

[0009] Furthermore, the vehicle driving state includes at least one or more of the following parameters: Vehicle motion dynamics parameters, steering system state parameters, driving mode calibration parameters, and intelligent driving state parameters. The step of generating feedforward compensation commands based on the road excitation input time sequence and the real-time vehicle driving state includes: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact; When the vibration frequency meets the preset harmful vibration conditions, the feedforward compensation command is generated. The feedforward compensation command includes the steering system damping coefficient and rack force compensation amount to offset the corresponding axial impact.

[0010] Furthermore, the method also includes: The feedforward control command is executed at the control lead time before the wheel contacts the axial impact.

[0011] Furthermore, the step of executing the feedforward control command at the control lead time before the wheel contacts the axial impact includes: At the predicted moment when the vehicle is in contact with the axial impact, closed-loop compensation is performed on the feedforward control command based on the real-time feedback signal from the rack force sensor.

[0012] Furthermore, the method also includes: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact; When the vibration frequency meets the preset beneficial vibration frequency band, a road feel torque feedforward command is generated. The road feel torque feedforward command includes a steering wheel feedback torque base value and a dynamic compensation torque that match the corresponding axial impact.

[0013] Furthermore, the method also includes: At the predicted moment when the wheel contacts the axial impact, the road feel torque feedforward command is executed.

[0014] According to a second aspect of the embodiments of this disclosure, a steer-by-wire control device is provided, comprising: The data pre-acquisition module is used to acquire road perception data of a pre-aiming area in the direction of vehicle travel. The road perception data includes at least road image data and / or three-dimensional point cloud data of the road surface in the pre-aiming area. The road surface excitation simulation module is used to extract road surface feature parameters based on the road surface sensing data, and to construct a road surface digital elevation prediction model based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area. The vibration feedforward compensation module is used to generate feedforward compensation commands based on the road excitation input timing sequence and the real-time vehicle driving status. The feedforward control commands are used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

[0015] According to a third aspect of the embodiments of this disclosure, a computer apparatus is provided, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to execute the steer-by-wire control method provided in the embodiments of this disclosure.

[0016] The technical solutions provided by the embodiments of this disclosure can include the following beneficial effects: acquiring road surface perception data of the pre-aiming area in the vehicle's driving direction; extracting road surface feature parameters based on the road surface perception data; constructing a road surface digital elevation pre-aiming model based on the road surface feature parameters; outputting a road surface excitation input time sequence of the pre-aiming area from the road surface digital elevation pre-aiming model; and generating a feedforward compensation command based on the road surface excitation input time sequence and the real-time vehicle driving state. The feedforward control command is used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area. Relying on multi-sensor fusion pre-aiming and feedforward control, the problem of lag in the steer-by-wire system is solved, harmful vibrations are attenuated in advance, and the driving experience and safety are improved.

[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0019] Figure 1 This is a flowchart illustrating a steer-by-wire control method according to an exemplary embodiment.

[0020] Figure 2 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0021] Figure 3 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0022] Figure 4 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0023] Figure 5 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0024] Figure 6 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0025] Figure 7 This is a flowchart illustrating yet another steer-by-wire control method according to an exemplary embodiment.

[0026] Figure 8 This is a block diagram illustrating a steer-by-wire control device according to an exemplary embodiment. Detailed Implementation

[0027] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0028] With the rapid development of vehicle electrification and intelligence, steer-by-wire systems have become a core component of advanced intelligent driving vehicles. Steer-by-wire systems eliminate the mechanical connection between the steering wheel and the steering wheels, using electrical signals to transmit and execute steering commands. They offer key advantages such as adjustable steering ratios, flexible vehicle layout, compatibility with advanced autonomous driving control, and good collision safety, making them a core technology direction for future intelligent chassis.

[0029] However, existing mass-produced and publicly available steer-by-wire systems suffer from two major technical challenges that are difficult to reconcile: the steering wheel vibration filtering is both delayed and passive. Vibration suppression schemes in steer-by-wire systems typically rely on feedback control based on steering rack force sensors and the steering motor's operating status. This means that the controller only triggers vibration damping after the road surface excitation has been transmitted to the steering system via the steering tie rod and rack, resulting in an inherent control delay. When the vehicle travels over speed bumps, on roads with continuous potholes, or on bumpy unpaved roads, steering wheel "kickback" and high-frequency vibrations are highly likely to occur, severely impacting driving comfort and, in extreme conditions, interfering with the driver's steering input and posing a safety hazard.

[0030] Furthermore, the simulated road feel is disconnected from the actual road surface, resulting in poor realism. The road feel feedback of steer-by-wire systems is mostly based on passive feedback from steering motor current and rack force, or on preset calibration lookup models based on vehicle speed and steering wheel angle. These systems cannot perceive the actual working conditions of the road ahead in advance, leading to a significant timing difference between the simulated road feel and the excitation of the real road surface, resulting in the common industry problem of "false feel".

[0031] Furthermore, harmful vibrations and beneficial road feel are generated simultaneously. Excessive filtering can lead to a lack of road feel and the driver's inability to perceive the road surface conditions. Insufficient damping can result in poor comfort, making it difficult to balance driving comfort and road feel realism.

[0032] To address the aforementioned issues, embodiments of this disclosure provide a steer-by-wire control method, apparatus, and computer device. Based on a pre-aiming region, a road excitation input timing sequence is acquired to predict future road conditions. Feedforward compensation commands are then generated to promptly create a scheme to suppress impending harmful vibrations. This solves the problem of lag in the steer-by-wire system's response, enabling proactive vibration prediction and timely intervention, significantly improving driving experience and safety. While reducing harmful vibrations, it simultaneously restores realistic road feel, balancing driving comfort and handling quality, making it suitable for advanced intelligent driving and human-machine co-driving scenarios, and possessing strong mass production feasibility.

[0033] An exemplary embodiment of this disclosure provides a steer-by-wire control method, the process of which uses this method to suppress harmful vibrations is as follows: Figure 1 As shown, it includes: Step 101: Obtain road perception data of the pre-aimed area in the vehicle's driving direction.

[0034] In this step, road perception data of the area ahead of the vehicle is acquired, providing a data foundation for subsequent road feature recognition, advance modeling, and the generation of feedforward control commands.

[0035] The road surface perception data includes at least road surface image data and / or road surface 3D point cloud data. That is, the system can acquire road surface image data solely through an image acquisition device, or it can acquire road surface 3D point cloud data solely through a LiDAR, or it can acquire both types of data simultaneously through an image acquisition device and a LiDAR and fuse them to obtain more comprehensive and accurate road surface perception information. The image acquisition device can be a forward-looking binocular camera.

[0036] According to one exemplary implementation, the preview area refers to a road surface perception range dynamically defined by the system based on the current driving state, located ahead of the vehicle's direction of travel. The system collects road surface data within this preview area to perceive road features ahead in advance, reserving sufficient time for subsequent feedforward control. The division of the preview area is dynamically adjusted according to the vehicle's driving state; when the vehicle speed is high, the preview area expands accordingly to ensure the system has sufficient control lead time; when the vehicle speed is low, the preview area shrinks accordingly to reduce unnecessary computational burden and optimize system resource allocation.

[0037] The estimated time to pass through the pre-aiming area based on the current driving status of the vehicle is the pre-aiming time domain.

[0038] Step 102: Extract road feature parameters based on the road surface perception data, and construct a road surface digital elevation prediction model based on the road surface feature parameters.

[0039] The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area.

[0040] In this step, feature extraction and modeling are performed on the acquired road perception data, transforming the raw sensor data into a structured road model to provide accurate input for subsequent control decisions. By constructing a digital elevation prediction model of the road surface, the system can predict the road excitation characteristics that the vehicle is about to pass over, including information such as road unevenness, the geometric dimensions and spatial location of bumps or depressions, road paving type, and adhering materials. This allows the system to grasp the temporal characteristics and intensity distribution of road excitation before the wheels actually contact obstacles, laying the foundation for the precise implementation of feedforward control.

[0041] The road surface digital elevation prediction model stores three-dimensional road surface information, is constructed along the vehicle's travel direction, and covers the entire prediction area. The road surface excitation input time sequence output by the digital elevation prediction model describes the variation of the road excitation the vehicle will experience over time within the prediction time domain, including key parameters such as the excitation's occurrence time, duration, amplitude, and frequency characteristics. These parameters directly serve subsequent vibration filtering and road feel simulation decisions, and are the core data input for the entire feedforward control link.

[0042] Step 103: Based on the road excitation input timing sequence and the real-time vehicle driving status, generate a feedforward compensation command. The feedforward control command is used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

[0043] In this step, control decisions are made based on the excitation timing information output by the road surface digital elevation prediction model and the real-time driving status of the vehicle. Feedforward compensation commands are generated to guide the steering mechanism to execute vibration reduction control before the wheels contact road obstacles. This transforms the results of perception and modeling into specific control actions, achieving a technological leap from delayed passive response to timely proactive prevention.

[0044] The generation process of the feedforward compensation command comprehensively considers the intensity and frequency characteristics of the road excitation, as well as the current driving state of the vehicle. The system first identifies the axial impact and its vibration frequency caused by the road excitation based on the road excitation input time sequence, and then determines whether the vibration frequency falls within a preset harmful vibration frequency band. If the vibration frequency falls within the harmful vibration frequency band, the system calculates the steering system damping coefficient and rack force compensation amount required to offset the impact based on the vehicle dynamics model, generating a feedforward compensation command that includes the target damping coefficient and rack force feedforward compensation amount. In this way, the system can selectively suppress harmful vibrations rather than treating all road excitations indiscriminately, thus achieving an optimal balance between vibration reduction effect and system efficiency.

[0045] An exemplary embodiment of this disclosure also provides a steer-by-wire control method, the process of which involves acquiring road perception data of a preview area in the vehicle's driving direction as follows: Figure 2 As shown, it includes: Step 201: Based on the vehicle's driving status, divide the pre-aiming area in the driving direction.

[0046] In this step, the spatial range of the preview area is dynamically determined based on the vehicle's current driving status, setting a clear perception boundary for subsequent road data collection. The reasonable division of the preview area directly affects the sufficiency of the system's perception of the road surface ahead and the timing accuracy of the feedforward control.

[0047] According to one exemplary embodiment, the aiming area is defined along the vehicle's travel direction, and its leading edge distance (i.e., aiming distance) is dynamically adjusted based on the vehicle's real-time driving speed. The aiming time domain is the time required for the vehicle to travel the aiming distance at the current speed, i.e., the aiming time domain equals the aiming distance divided by the vehicle's real-time driving speed. The control lead is the reserved time from when the system generates the feedforward control command to when the wheel contacts the target road surface. The aiming time domain includes two sub-intervals: data processing and decision time and control lead. The control lead is maintained within the range of 100ms to 500ms to match the response delay of the steer-by-wire system and the transmission delay of the road surface excitation. When the vehicle speed is 60km / h (approximately 16.67m / s) and the control lead is 200ms, the aiming distance is approximately 30m; when the vehicle speed is 120km / h (approximately 33.33m / s) and the control lead is 300ms, the aiming distance is approximately 100m. Through this dynamic adjustment mechanism, the system can maintain sufficient control lead at different vehicle speeds, ensuring the timing accuracy of feedforward control.

[0048] Step 202: Acquire the road surface image data of the pre-aiming area using an image acquisition device.

[0049] In this step, visual information of the road surface within a pre-aimed area is acquired using an image acquisition device, providing an image data foundation for subsequent road surface semantic recognition and feature extraction. The image acquisition device is preferably a forward-facing binocular camera, installed in the rearview mirror position inside the vehicle's windshield, possessing a resolution of over 8 megapixels, a frame rate of 30fps, and a 120-degree field of view, capable of recognizing road surface image information within a 150m range ahead. After processing the road surface image data acquired by the camera using a semantic segmentation algorithm, the resulting semantic segmentation of the road surface can identify different types of road surface features such as speed bumps, potholes, seams, puddles, and ice / snow, as well as their spatial distribution.

[0050] Step 203: Obtain the three-dimensional point cloud data of the road surface in the pre-aiming area using lidar.

[0051] In this step, a lidar is used to collect three-dimensional spatial information of the road surface within the pre-aiming area, providing accurate three-dimensional point cloud data for subsequent road surface elevation modeling. The lidar is preferably a 128-line automotive-grade semi-solid-state lidar, installed on the front of the vehicle roof, with a point cloud frequency of 10Hz, a ranging range of 200m, and a ranging accuracy of ±2cm. The three-dimensional point cloud data of the road surface collected by the lidar includes the spatial coordinates (X, Y, Z) and reflection intensity information of each reflection point, accurately characterizing the three-dimensional morphology and elevation changes of the road surface, providing indispensable three-dimensional geometric data support for the construction of the road surface digital elevation pre-aiming model.

[0052] Step 204: Synchronize the road surface image data with the road surface 3D point cloud data in time and / or in space.

[0053] In this step, spatiotemporal synchronization eliminates the temporal and spatial discrepancies between camera image data and LiDAR point cloud data, achieving precise alignment of the two heterogeneous sensor data and providing a spatiotemporally consistent data foundation for subsequent multimodal fusion. Since the sampling frequencies, data formats, and coordinate reference systems of the camera and LiDAR are different, without spatiotemporal synchronization, the two types of data cannot be effectively correlated and fused.

[0054] In terms of time synchronization, the system uses the vehicle's PTP clock synchronization signal to unify the timestamps of camera image frames and LiDAR point cloud frames to the same time base, achieving a time synchronization accuracy within 1ms. For spatial synchronization, the system uses pre-calibrated camera intrinsic parameter matrices and LiDAR extrinsic parameter matrices to transform and unify the camera pixel coordinate system and LiDAR world coordinate system to the vehicle coordinate system, achieving pixel-level and point cloud-level spatial alignment. After spatiotemporal synchronization, the road surface semantic information recognized by the camera (such as the outline position of speed bumps) and the road surface three-dimensional information measured by the LiDAR (such as the height and distance of speed bumps) can accurately correspond, providing a reliable data foundation for multimodal fusion.

[0055] Step 205: Merge the road surface image data and the road surface 3D point cloud data to obtain the road surface perception data.

[0056] In this step, the spatiotemporally synchronized visual semantic information and 3D point cloud elevation information are fused using a multimodal method to generate fused road surface perception data. This fusion process fully leverages the complementary advantages of the two sensors: the camera excels in semantic recognition, capable of distinguishing road surface types and identifying obstacle categories; the LiDAR excels in 3D measurement, accurately acquiring road surface geometry and elevation information. Through multimodal fusion, the system obtains road surface perception data that includes both semantic classification information and precise 3D geometric information, significantly outperforming single-sensor solutions in terms of perception accuracy and reliability. For example, in identifying speed bumps, the camera provides the outline and planar position of the speed bump, while the LiDAR provides its height, length, and precise distance from the vehicle's front wheels. The fusion of these two parameters outputs a speed bump perception result containing complete information such as type, location, size, and distance.

[0057] An exemplary embodiment of this disclosure also provides a steer-by-wire control method, wherein the process of using this method to extract road feature parameters based on the road surface perception data and construct a road surface digital elevation prediction model based on the road surface feature parameters is as follows: Figure 3 As shown, it includes: Step 301: Extract the road feature parameters from the road surface perception data.

[0058] In this step, structured road feature parameters are extracted from the fused road surface full-domain perception data, transforming the original perception data into feature vectors that can be used for modeling and control decisions.

[0059] The road surface characteristic parameters include at least one or more of the following parameters: road surface geometric features, road surface paving and adhesion material features, and road surface damage features.

[0060] The road surface geometry features include parameters such as road surface unevenness, the geometric dimensions (height, width, length) of protrusions or depressions, and their spatial location (distance from the vehicle's front axle, lateral offset). These parameters directly determine the intensity and temporal characteristics of road surface excitation. Road pavement and adhesion material characteristics include parameters such as pavement type (asphalt pavement, concrete pavement, unpaved pavement, etc.), road adhesion coefficient, and road material uniformity. These parameters affect the interaction characteristics between the tire and the road surface, thus affecting the force distribution on the steering system. Road surface damage features include information on road surface damage such as cracks, potholes, ruts, and joints, as well as their severity. These parameters help the system assess road surface conditions and make corresponding adjustments to control strategies.

[0061] Step 302: Establish the vehicle coordinate system.

[0062] The vehicle coordinate system has the center point of the front axle of the vehicle as the origin, the direction of travel as the X-axis, the left and right lateral directions of the vehicle as the Y-axis, and the vertical height of the road surface as the Z-axis.

[0063] In this step, a unified vehicle coordinate system is established to provide a consistent spatial reference benchmark for the spatial mapping of road surface feature parameters and the construction of the road surface digital elevation prediction model.

[0064] According to one exemplary embodiment, the vehicle coordinate system uses the center point of the vehicle's front axle as its origin, ensuring a direct geometric correspondence between the coordinate system and the vehicle's steering system. This facilitates the subsequent conversion of road surface excitation into force analysis of the steering system. The X-axis, along the vehicle's forward direction, describes the distribution of road surface features along the travel direction; the Y-axis, along the vehicle's left-right lateral direction, describes the lateral offset of road surface features relative to the vehicle's trajectory; and the Z-axis, along the vertical height of the road surface, describes the elevation changes of the road surface. This vehicle coordinate system is dynamically updated as the vehicle moves in real time, ensuring that the coordinate system always reflects the real-time spatial relationship between the vehicle and the road surface.

[0065] Step 303: Map the road surface feature parameters to the vehicle coordinate system, and construct the road surface digital elevation prediction model covering the prediction area along the positive X-axis.

[0066] The road surface digital elevation prediction model outputs the road surface excitation input time sequence in the prediction time domain.

[0067] In this step, the extracted road surface feature parameters are spatially mapped in the vehicle coordinate system to construct a digital elevation prediction model of the road surface covering the entire prediction area, and the road surface excitation input time series in the prediction time domain is output. This model is stored in the form of a rasterized elevation map, with a preferred grid resolution of 0.1m × 0.1m, covering the entire spatial range from the front axle of the vehicle to the leading edge of the prediction area along the positive X-axis. Each grid cell in the model stores the road surface elevation value, road surface type label, and confidence information at that location.

[0068] Based on this digital elevation prediction model of the road surface, the system can predict the road surface excitation characteristics that a vehicle will traverse within the prediction time domain. Specifically, the system calculates the sequence of path points that the wheels will pass through within the prediction time domain based on the vehicle's current speed and trajectory. It extracts the road surface elevation values ​​and road surface type information at each path point from the elevation model, and combines this with the tire model and the vehicle dynamics model to predict the vertical road surface excitation and the axial impact of the steering system that the wheels will experience. Finally, it outputs the road surface excitation input time series sequence. This time series sequence contains key parameters such as the time node of the excitation occurrence, duration, amplitude, and frequency characteristics, providing core data input for subsequent vibration filtering decisions and road feel simulation decisions.

[0069] Furthermore, the system can classify road surfaces based on extracted road feature parameters, resulting in classifications such as smooth paved roads, bumpy roads, speed bumps, potholes, unpaved roads, and low-adhesion roads with ice, snow, or water accumulation. These classifications can be used to access pre-calibrated control parameter sets for the corresponding road condition type, providing parameter support for subsequent control decisions and improving the matching accuracy and response speed of control strategies.

[0070] An exemplary embodiment of this disclosure provides a steer-by-wire control method, wherein the process of generating feedforward compensation commands using this method based on the road excitation input timing sequence and the real-time vehicle driving state is as follows: Figure 4 As shown, it includes: Step 401: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact.

[0071] The vehicle driving state includes at least one or more of the following parameters: vehicle motion dynamics parameters, steering system self-state parameters, driving mode calibration parameters, and intelligent driving state parameters.

[0072] In this step, based on the excitation timing sequence output by the digital elevation prediction model of the road surface, axial impact events caused by road surface excitation are identified within the prediction area in front of the vehicle, and the vibration frequency of each impact event is determined. The axial impact refers to the axial force impact generated on the steering rack after road surface unevenness is transmitted to the steering system through the tires and suspension. Different road surface features will produce axial impacts with different frequency characteristics: for example, speed bumps typically produce high-frequency impact vibrations of 30 to 50 Hz, road joints produce higher-frequency, small-amplitude vibrations, and unpaved roads produce broadband continuous vibrations.

[0073] The vehicle driving state parameters provide the basis for overall vehicle status for control decisions. Vehicle motion dynamics parameters include vehicle speed, yaw rate, and lateral acceleration, reflecting the overall motion state of the vehicle. Steering system state parameters include steering wheel angle, steering wheel torque, steering motor speed, rack displacement, and rack force, reflecting the current operating state of the steering system. Driving mode calibration parameters include the current driving mode (Comfort mode, Sport mode, Autonomous driving mode, and Human-Machine Co-driving mode), determining the selection of calibration parameters for the road feel mapping model. Intelligent driving state parameters include the intelligent driving takeover state (manual driving or autonomous driving), affecting the switching of system control strategies. These parameters collectively constitute the input condition space for control decisions, ensuring that the generation of feedforward compensation commands fully considers the actual driving state of the vehicle.

[0074] Step 402: If the vibration frequency meets the preset harmful vibration conditions, generate the feedforward compensation command.

[0075] The feedforward compensation command includes a steering system damping coefficient and rack force compensation to offset the corresponding axial impact.

[0076] Harmful vibration conditions include harmful vibration frequency bands and harmful vibration amplitudes.

[0077] In this step, the harmful vibration frequency band can be used as a threshold condition to determine whether a feedforward compensation command needs to be generated for the identified axial impact. The system only triggers the generation and execution of the feedforward compensation command when the vibration frequency of the axial impact falls within the harmful vibration frequency band and / or the vibration amplitude exceeds the harmful vibration level. If the vibration frequency does not fall within the harmful vibration frequency band and / or the vibration amplitude does not reach the harmful vibration level, the conventional steering control strategy is maintained, and no additional vibration filtering feedforward control is triggered. This threshold determination mechanism avoids the system treating all road surface excitations indiscriminately, reducing unnecessary control intervention while ensuring vibration reduction effects, thus improving the overall efficiency of the system and the driving experience.

[0078] Low-frequency vibrations are beneficial road feedback, but when bumps and undulations are too large or the vibration amplitude exceeds the preset harmful vibration amplitude, they will turn into harsh impacts and large shaking, affecting ride comfort. Therefore, vibration amplitude is also included in the reference dimension for determining whether harmful vibrations occur.

[0079] According to one exemplary embodiment, the harmful vibration frequency band includes road impact vibrations above a preset frequency threshold (e.g., 20Hz) and low-frequency bump impacts above a preset amplitude. High-frequency vibrations above 20Hz, when transmitted to the steering wheel, cause the driver to experience noticeable high-frequency shaking and a jolting sensation, severely affecting driving comfort and operational precision; therefore, they are classified as harmful vibrations. While low-frequency vibrations (typically in the range of 1 to 20Hz) are not necessarily harmful in themselves, when their amplitude exceeds a preset threshold, they can also adversely affect driving comfort; therefore, low-frequency bump impacts above a preset amplitude are also included in the harmful vibration frequency band determination.

[0080] When the feedforward compensation command is triggered, the system calculates the steering system damping coefficient and rack force compensation amount required to counteract the axial impact based on the vehicle dynamics model. The damping coefficient is adjusted by changing the output characteristics of the steering actuator motor, and the rack force compensation amount is achieved by applying a counterforce to the steering rack. The feedforward compensation command includes the specific values ​​of the target damping coefficient and rack force feedforward compensation amount, as well as the execution time of the command (i.e., the control lead time), ensuring that the steering mechanism completes the adjustment of damping parameters before the wheels contact road obstacles.

[0081] An exemplary embodiment of this disclosure also provides a steer-by-wire control method, the flow of which executes feedforward control commands using this method is as follows: Figure 5 As shown, it includes: Step 501: Obtain road surface perception data of the pre-aiming area in the vehicle's driving direction. The road surface perception data includes at least road surface image data and / or road surface three-dimensional point cloud data of the pre-aiming area.

[0082] Step 502: Extract road feature parameters based on the road surface perception data, and construct a road surface digital elevation prediction model based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area.

[0083] Step 503: Based on the road excitation input timing sequence and the real-time vehicle driving status, generate a feedforward compensation command. The feedforward control command is used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

[0084] The implementation principles of steps 501 to 503 are the same as those of steps 101 to 103, and will not be repeated here.

[0085] Step 504: At the control lead time before the wheel contacts the axial impact, execute the feedforward control command.

[0086] Specifically, at the predicted moment when the vehicle comes into contact with the axial impact, the feedforward control command is compensated in a closed loop based on the real-time feedback signal from the rack force sensor.

[0087] In this step, at the control lead time before the wheel actually contacts the road obstacle, a feedforward control command is executed. The steering mechanism adjusts the vibration reduction parameters in advance, achieving active feedforward filtering of harmful vibrations. Simultaneously, at the predicted moment of wheel contact with the obstacle, the system combines the real-time feedback signal from the rack force sensor to perform closed-loop compensation on the feedforward control command, forming a composite vibration reduction control mechanism of feedforward plus feedback, further optimizing the vibration reduction effect.

[0088] According to one exemplary implementation, when executing feedforward control commands, the steering controller receives vibration filtering feedforward control commands and adjusts the output damping, assist characteristics, and rack force compensation of the steering actuator motor in real time. At the control lead time before the wheel contacts the target road surface, the system increases the damping coefficient of the steering system to a preset multiple of the normal operating condition, while simultaneously applying reverse rack force feedforward compensation to establish a preload force to counteract the upcoming road impact. When the wheel passes over the target road surface, the feedforward control has already completed damping adjustment and force compensation in advance, significantly offsetting the impact vibration from the road surface. Subsequently, the system combines the real-time feedback signal from the rack force sensor to calculate the remaining vibration and dynamically adjust the motor output, performing closed-loop compensation to achieve suppression of harmful vibrations across the entire time domain.

[0089] Taking a typical driving condition of a vehicle traveling at 60 km / h on an urban road with a speed bump 30 m ahead as an example: the system begins to execute feedforward control commands 150 ms before the wheels contact the speed bump (i.e., when the vehicle is about 5 m away from the speed bump), increasing the steering system damping coefficient to three times that under normal conditions, while simultaneously applying reverse rack force compensation. When the wheels roll over the speed bump, the peak rack force generated by the road impact is approximately 200 N. After feedforward compensation, the actual rack force transmitted to the steering system is reduced to approximately 30 N, achieving a vibration attenuation rate of 85%. The steering wheel vibration amplitude decreases from approximately 0.8 N·m to approximately 0.05 N·m, and the driver experiences almost no steering wheel kickback.

[0090] An exemplary embodiment of this disclosure also provides a steer-by-wire control method, the process of which transmits beneficial vibrations to enhance the user's driving experience as follows: Figure 6 As shown, it includes: Step 601: Obtain road perception data of the pre-aiming area in the vehicle's driving direction. The road perception data includes at least road image data and / or three-dimensional point cloud data of the road surface in the pre-aiming area.

[0091] Step 602: Extract road feature parameters based on the road surface perception data, and construct a road surface digital elevation prediction model based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area.

[0092] The implementation principles of steps 601 to 602 are the same as those of steps 101 to 102, and will not be repeated here.

[0093] Step 603: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact.

[0094] In this step, based on the road surface excitation input time sequence, axial impact events and their vibration frequencies within the pre-aiming area are identified, providing a basis for subsequent differentiation between beneficial and harmful vibrations. The implementation principle of this step is the same as that of step 401, and will not be repeated here.

[0095] Step 604: If the vibration frequency meets the preset beneficial vibration frequency band, generate a road feel torque feedforward command.

[0096] The road feel torque feedforward command includes the steering wheel feedback torque base value and dynamic compensation torque that match the corresponding axial impact.

[0097] In this step, the beneficial vibration frequency band is used as a threshold criterion to identify the beneficial road feel information that needs to be transmitted to the driver, and corresponding road feel torque feedforward commands are generated. Beneficial vibrations refer to vibration information that can help the driver perceive road conditions and assist driving decisions, including low-frequency torque changes caused by changes in road pavement type, steering wheel torque feedback corresponding to lateral forces in curves, and torque fluctuations caused by changes in the road adhesion coefficient. This beneficial vibration information is usually concentrated in a low frequency range and will not cause discomfort to the driver.

[0098] The road feel torque feedforward command includes two core parameters: the steering wheel feedback torque base value and the dynamic compensation torque. The feedback torque base value is determined by a pre-calibrated road feel torque mapping model based on road surface type and vehicle driving state, providing a basic level of road feel feedback. The dynamic compensation torque is dynamically calculated based on the specific characteristics of road surface excitation (such as unevenness amplitude and frequency), and is used to superimpose real-time torque changes that match road surface characteristics on the base torque, enabling the driver to experience a realistic road feel synchronized with the actual road surface.

[0099] The road feel torque mapping model is a nonlinear mapping model based on pre-planned road condition calibration. Its input parameters include pre-planned road surface unevenness, road surface type, adhesion coefficient, vehicle speed, steering wheel angle, and steering rate. The output parameters are the base value of the steering wheel feedback torque, the dynamic compensation value, and the vibration retention frequency band. The calibration parameters of the mapping model are different in different driving modes: in Comfort mode, high-frequency vibrations are filtered to retain basic road feel and prioritize driving comfort; in Sport mode, a wider frequency range (such as 1 to 15 Hz) of road feel information is retained to improve the clarity of road feedback and meet driving control requirements; in Autonomous Driving mode, all road feel vibrations are blocked, and only the steering limit torque is retained to reduce steering wheel disturbance during autonomous driving.

[0100] Step 605: At the predicted moment when the wheel contacts the axial impact, execute the road feel torque feedforward command.

[0101] In this step, at the predicted moment when the wheel contacts the road obstacle, a road feel torque feedforward command is executed. A feedback torque matching the predicted road surface characteristics is applied to the steering wheel via the road feel feedback motor, allowing the driver to experience a realistic road feel consistent with the road surface characteristics as the vehicle actually contacts the obstacle. This execution process is synchronized with the feedforward vibration filtering control in Example 5, and the two work together: vibration filtering control is responsible for suppressing the transmission of harmful vibrations to the steering wheel, while road feel torque control is responsible for transmitting beneficial road feel information to the steering wheel, together achieving a precise balance between harmful vibration filtering and beneficial road feel transmission.

[0102] When executing the road feel torque feedforward command, the road feel controller receives the command and controls the road feel feedback motor to output the corresponding feedback torque. The torque output employs a gradual transition strategy, using a preset torque transition duration (e.g., 200ms) to achieve smooth establishment and decay of the torque output, avoiding sudden changes in road feel that could cause driver discomfort. Taking a vehicle passing over a speed bump as an example: at the moment the wheels contact the speed bump, the road feel feedback motor begins to output a base torque (e.g., 0.8 N·m), followed by a dynamic compensation torque (e.g., 0.3 N·m) over 50 to 150ms, bringing the total feedback torque to 1.1 N·m, simulating the real road feel when passing over a speed bump; within 150 to 300ms, the torque gradually decays back to normal levels, completing the smooth transition of the road feel feedback. Throughout this process, high-frequency harmful vibrations (greater than 20Hz) are completely filtered out by vibration filtering control, and the driver only experiences clear but not abrupt low-frequency road feel information.

[0103] According to one exemplary implementation, a fail-safe mechanism is configured. Specifically, when the multi-sensor fusion pre-aiming module malfunctions (e.g., camera image loss, abnormal LiDAR point cloud), the confidence level of the perception data falls below a preset threshold (e.g., confidence level below 0.70), or the execution module responds abnormally, the system automatically exits the pre-aiming feedforward control mode and seamlessly switches to the traditional steering rack force feedback control mode. Simultaneously, it reports fault information to the vehicle controller and instrument system to ensure the functional safety of the steer-by-wire system. When only the camera malfunctions, the system can switch to a pure LiDAR pre-aiming mode; when only the LiDAR malfunctions, the system can switch to a pure camera pre-aiming mode; when both sensors malfunction simultaneously, the system completely exits the pre-aiming function, relying on traditional feedback control to maintain basic steering functionality. The entire fail-safe mechanism complies with ISO 26262 automotive-grade functional safety ASIL-D level requirements.

[0104] An exemplary embodiment of this disclosure also provides a steer-by-wire control method, the process of which uses this method to classify and determine road conditions and apply corresponding control strategies is as follows: Figure 7 As shown, it includes: Step 701: Classify the road surface working conditions according to the road surface characteristic parameters.

[0105] The working condition classification includes any one or more of the following types: smooth paved road, bumpy road, speed bump, pothole road, unpaved road, and low-adhesion road surface with ice / snow / water accumulation.

[0106] In this step, based on the structured feature parameters output by the road feature extraction unit, the road conditions within the current target area are classified into working conditions. The continuously changing road features are discretized into a limited number of typical working condition types, providing an index for the rapid matching of subsequent control strategies and the retrieval of pre-calibrated parameter sets. The road working condition classification comprehensively considers multi-dimensional information such as road geometric features (e.g., unevenness, bump / indentation size), road paving and adhesion material features (e.g., paving type, adhesion coefficient), and road damage features (e.g., the severity of cracks and potholes). Through preset classification rules or machine learning classification models, the road surface is divided into typical working condition types such as smooth paved roads, bumpy roads, speed bumps, potholes, unpaved roads, and icy / waterlogged low-adhesion roads. For example, when road surface feature parameters show that there is a protrusion with a height greater than 3cm and a length in the range of 20 to 40cm, and the protrusion is located on the vehicle's driving trajectory, the system classifies it as a "speed bump"; when the road surface shows continuous unevenness and the adhesion coefficient is less than 0.3, the system classifies it as an "ice / snow / water low adhesion road surface".

[0107] Step 702: Based on the results of the working condition classification, call the pre-calibrated control parameter set for the corresponding working condition type.

[0108] In this step, based on the working condition classification results obtained in step 701, a pre-calibrated control parameter set matching the working condition type is retrieved from the system's pre-stored control parameter library. This provides an initial parameter benchmark for subsequent vibration filtering and road feel simulation decisions. Different working condition types correspond to different road excitation characteristics, thus requiring differentiated control strategies and parameter configurations. The pre-calibrated control parameter set is obtained through bench tests and road tests during the system development phase. For each typical working condition type, an optimized combination of control parameters is stored, including key parameters such as the target damping coefficient base value, rack force compensation base value, vibration filtering frequency band range, road feel retention frequency band range, feedback torque base value, and dynamic compensation torque gain.

[0109] For example, when the working condition classification result is "speed bump", the system calls the pre-calibrated control parameter set corresponding to the speed bump working condition: the target damping coefficient base value is set to 3 times that of the normal working condition, the rack force compensation base value is set to the reverse compensation value required to offset the impact of the speed bump, the vibration filtering frequency band is set to 20 to 100 Hz, the road feel retention frequency band is set to 1 to 15 Hz, the feedback torque base value is set to 0.8 N·m, and the dynamic compensation torque gain is set to 0.3 N·m. When the working condition classification result is "smooth paved road", the system calls the pre-calibrated control parameter set corresponding to the smooth paved road working condition: the target damping coefficient base value is maintained at 1 times that of the normal working condition (i.e., no additional damping is added), the rack force compensation base value is set to 0, the vibration filtering frequency band is set to greater than 50 Hz (only filtering extremely high frequency vibrations), the road feel retention frequency band is set to 1 to 20 Hz, the feedback torque base value is set to 0.5 N·m, and the dynamic compensation torque gain is set to 0.1 N·m. When the working condition classification result is "unpaved road", the system calls the pre-calibrated control parameter set corresponding to the unpaved road working condition: the target damping coefficient base value is set to twice that of the normal working condition, the rack force compensation base value is dynamically calculated based on the amplitude of continuous unevenness, the vibration filtering frequency band is set to greater than 20Hz, the road feel retention frequency band is set to 1 to 15Hz, the feedback torque base value is set to 1.0N·m, and the dynamic compensation torque gain is set to 0.5N·m.

[0110] Step 703: Generate feedforward compensation command and / or road feel torque feedforward command based on the pre-calibrated control parameter set and the vehicle driving state.

[0111] In this step, the pre-calibrated control parameter set called in step 702 is used as the initial parameter benchmark. Combined with real-time vehicle driving conditions (including vehicle speed, driving mode, and intelligent driving takeover status), the parameters are corrected and optimized in real time, ultimately generating feedforward compensation commands and / or road feel torque feedforward commands. The pre-calibrated control parameter set provides optimized parameter benchmarks for specific operating conditions, but in actual control, dynamic adjustments are still needed based on the vehicle's real-time status. For example, under the same speed bump condition, when the vehicle speed is 60 km / h, the control lead is 150 ms; when the vehicle speed is 30 km / h, the control lead is adjusted to 300 ms accordingly. The target damping coefficient and rack force compensation in the pre-calibrated parameter set also need to be proportionally corrected according to changes in vehicle speed. Furthermore, different driving modes also affect the final parameter values: in comfort mode, the system further reduces the road feel feedback intensity based on the pre-calibrated parameters; in sport mode, the system improves the clarity of the road feel feedback based on the pre-calibrated parameters; in autonomous driving mode, the system shields road feel vibrations, retaining only the steering limit torque.

[0112] Through the mechanism of classifying operating conditions and calling pre-calibrated parameter sets, the system can complete the entire link from road surface perception to control parameter matching within milliseconds. This avoids the latency issues caused by real-time online optimization calculations for different road conditions, ensuring the real-time performance and reliability of feedforward control. Meanwhile, the pre-calibrated parameter sets have undergone thorough experimental verification, guaranteeing the stability and effectiveness of the control strategy under different operating conditions.

[0113] Step 704: If the vehicle's driving state changes, update the operating condition classification and / or the pre-calibrated control parameter set.

[0114] In this step, when the vehicle's driving state changes (such as changes in vehicle speed, driving mode switching, or changes in intelligent driving takeover status), the system dynamically updates the operating condition classification results and / or calls to the pre-calibrated control parameter set to ensure that the control strategy always matches the current driving state. For example, when the vehicle decelerates from 60 km / h to 30 km / h, the pre-aiming distance is shortened accordingly. The system reassesses the road surface characteristics within the pre-aiming area, and may reclassify speed bumps that were originally far away as the current operating condition that needs to be handled, and call the corresponding pre-calibrated parameter set. When the driver switches from manual driving to autonomous driving mode, the system switches from the parameter set of sport or comfort mode to the parameter set of autonomous driving mode, shielding road vibrations and adjusting the vibration filtering strategy. This dynamic update mechanism ensures that the system maintains the optimal control state throughout the entire driving process, adapting to complex and ever-changing real-world driving scenarios.

[0115] According to one exemplary embodiment, the operating condition classification results determined in this disclosure can also be used for assisted driving mode recommendation. When the system continuously identifies low-adhesion operating conditions such as unpaved roads or icy / snowy roads, it can send a mode recommendation signal to the vehicle controller, suggesting that the driver switch to a more conservative driving mode to improve driving safety. When the system identifies continuous speed bumps or potholes ahead, it can send road condition warning information to the instrument system, prompting the driver to slow down in advance and reduce the impact of road impacts on the vehicle and passengers.

[0116] An exemplary embodiment of this disclosure also provides a steer-by-wire control device, the structure of which is as follows: Figure 8 As shown, it includes: The data pre-acquisition module 801 is used to acquire road surface perception data of a pre-aiming area in the direction of vehicle travel, wherein the road surface perception data includes at least road surface image data and / or road surface three-dimensional point cloud data of the pre-aiming area. The road surface excitation simulation module 802 is used to extract road surface feature parameters based on the road surface perception data, and construct a road surface digital elevation prediction model based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area. The vibration feedforward compensation module 803 is used to generate feedforward compensation commands based on the road excitation input timing sequence and the real-time vehicle driving status. The feedforward control commands are used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

[0117] The aforementioned devices can be integrated into intelligent driving domain controllers, steering control / actuators, road feel feedback mechanisms, and other equipment, enabling the corresponding functions to be implemented by the equipment.

[0118] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0119] An exemplary embodiment of this disclosure also provides a computer apparatus, including: processor; Memory used to store processor-executable instructions; The processor is configured to execute the steer-by-wire control method provided in the embodiments of this disclosure.

[0120] An exemplary embodiment of this disclosure also provides a non-transitory computer-readable storage medium, wherein instructions in the storage medium, when executed by a computer's processor, enable the computer to perform the steer-by-wire control method provided in embodiments of this disclosure.

[0121] This disclosure provides a steer-by-wire control method, apparatus, and computer device. First, road surface perception data of a pre-aiming area along the vehicle's driving direction is acquired. Based on the road surface perception data, road surface feature parameters are extracted, and a digital elevation model (DEM) for road surface pre-aiming is constructed based on these parameters. The DEM forecasting model outputs a road surface excitation input time sequence for the pre-aiming area. Then, based on the road surface excitation input time sequence and the real-time vehicle driving state, a feedforward compensation command is generated. This feedforward control command instructs the steering mechanism to suppress harmful vibrations caused by the pre-aiming area. By relying on multi-sensor fusion pre-aiming and feedforward control, the problem of lag in steer-by-wire systems is solved, harmful vibrations are attenuated in advance, and the driving experience and safety are improved. While reducing harmful vibrations, it simultaneously restores realistic road feel, balancing driving comfort and handling quality, adapting to advanced intelligent driving and human-machine co-driving scenarios, and demonstrating strong mass production feasibility.

[0122] By employing multi-sensor fusion pre-aiming technology combining cameras and lidar, the system acquires 3D elevation information and semantic features of the road surface ahead in advance. Based on a digital elevation pre-aiming model of the road surface, it predicts the timing of road excitation that the wheels will experience. Through feedforward control, it adjusts the steering system damping coefficient and rack force compensation before the wheels contact road obstacles, advancing the trigger time of vibration reduction control from the traditional "after the wheel has passed over" to "before the wheel makes contact." This breaks through the technical bottleneck of passive vibration reduction in traditional steer-by-wire systems, achieving proactive vibration filtering in advance and completely solving the inherent time delay problem of feedback control. Taking the scenario of traveling at 60 km / h over a speed bump as an example, the system begins feedforward control 150 ms before the wheels contact the speed bump, increasing the steering system damping to three times that of the normal scenario and applying reverse rack force compensation. This reduces the peak rack force generated by the road impact from approximately 200 N to approximately 30 N, achieving a vibration attenuation rate of 85%. The steering wheel vibration amplitude decreases from approximately 0.8 N·m to approximately 0.05 N·m, making the driver almost unaware of steering wheel kickback and significantly improving driving comfort.

[0123] Based on the real road surface features acquired through pre-aiming, a road feel torque mapping model is constructed. While filtering harmful vibrations, beneficial road feel information (such as changes in pavement type, lateral forces in curves, and changes in the coefficient of adhesion) is simultaneously transmitted to the steering wheel via road feel torque feedforward commands. This allows the driver to experience a realistic road feel consistent with the road surface characteristics the moment the vehicle actually contacts an obstacle. The road feel torque output employs a gradual transition strategy, achieving smooth torque build-up and decay through a preset torque transition duration, avoiding sudden road feel changes that could cause driver discomfort. Simultaneously, the system uses a dual-threshold judgment mechanism for beneficial and harmful vibration frequency bands to accurately distinguish between beneficial road feel to be transmitted and harmful vibrations to be filtered, achieving a balance between road feel realism and driving comfort. This achieves precise synchronization between road feel simulation and actual road surface conditions, solving the problem of unrealistic steering feel in traditional steer-by-wire systems.

[0124] By integrating semantic information from a forward-looking camera with 3D elevation information from a LiDAR sensor, the complementary advantages of both are fully utilized: the camera excels at road surface type recognition and obstacle classification, while the LiDAR excels at precise measurement of road surface geometry and elevation. Through time synchronization based on the vehicle's PTP clock and spatial synchronization based on intrinsic and extrinsic parameter matrices, spatiotemporal alignment at the pixel and point cloud levels is achieved. The fused road surface perception data significantly outperforms single-sensor solutions in terms of accuracy and reliability. Multi-sensor fusion ensures high reliability and high accuracy for forward-looking perception, enabling accurate identification of road surface features under various complex conditions such as day and night, rain, fog, and snow.

[0125] The anticipation time domain is the time required for the vehicle to travel the anticipation distance, while the control lead is a reserved execution window of 100ms to 500ms. The two are dynamically mapped through vehicle speed. The higher the vehicle speed, the longer the anticipation distance, but the control lead is always maintained within the effective range, ensuring that the system has sufficient perception-decision-execution time at different vehicle speeds. Through the layered design of the anticipation time domain and the control lead, the timing accuracy of the feedforward control is ensured, taking into account both the long-distance perception requirements of high-speed scenarios and the rapid response requirements of low-speed scenarios.

[0126] When a sensor malfunctions, the perceived data is abnormal, or the execution module responds abnormally, the system automatically exits the pre-aiming feedforward control mode and seamlessly switches to the traditional feedback control mode, ensuring that the basic functions of the steer-by-wire system are not affected. It supports degraded operation (pure camera pre-aiming or pure LiDAR pre-aiming) in case of a single sensor failure, and complete exit in case of a dual-sensor failure. The entire fail-safe mechanism meets the ISO 26262 automotive-grade functional safety ASIL-D level requirements, satisfying automotive-grade functional safety requirements and possessing extremely high reliability for engineering applications.

[0127] Those skilled in the art will also understand that the various illustrative logical blocks and steps listed in the embodiments of this disclosure can be implemented by electronic hardware, computer software, or a combination of both. Whether such functionality is implemented in hardware or software depends on the specific application and the overall system design requirements. Those skilled in the art can implement the described functionality using various methods for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of this disclosure.

[0128] 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.”

[0129] Similarly, although this disclosure 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 this specification and the accompanying drawings. This disclosure 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 specific features of this disclosure 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 claims, such terms are intended to be inclusive in a manner similar to the term “including.”

[0130] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure 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 disclosure are indicated by the following claims.

[0131] It should be understood that this disclosure is not limited to the precise structures 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 disclosure is limited only by the appended claims.

Claims

1. A steer-by-wire control method, characterized in that, include: Acquire road perception data of a pre-aimed area in the vehicle's driving direction, wherein the road perception data includes at least road image data and / or three-dimensional point cloud data of the pre-aimed area; Based on the road surface perception data, road surface feature parameters are extracted, and a road surface digital elevation prediction model is constructed based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area. Based on the road excitation input timing sequence and the real-time vehicle driving status, a feedforward compensation command is generated. The feedforward control command is used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

2. The steer-by-wire control method according to claim 1, characterized in that, The step of acquiring road perception data of the pre-aimed area in the vehicle's driving direction includes: The pre-aiming area is divided according to the vehicle's driving status in the driving direction; The road surface image data of the pre-aiming area is acquired using an image acquisition device; The three-dimensional point cloud data of the road surface in the pre-aiming area is obtained by lidar; Synchronize the road surface image data with the road surface 3D point cloud data in time and / or in space; The road surface image data and the road surface 3D point cloud data are fused to obtain the road surface perception data.

3. The steer-by-wire control method according to claim 1, characterized in that, The steps of extracting road feature parameters based on the road surface sensing data and constructing a road surface digital elevation prediction model based on the road surface feature parameters include: The road surface feature parameters are extracted from the road surface perception data, and the road surface feature parameters include at least any one or more of the following parameters: Road surface geometric features, road surface paving and adhesion material features, and road surface damage features; Establish a vehicle coordinate system with the center point of the front axle of the vehicle as the origin, the direction of travel as the X-axis, the left and right lateral directions of the vehicle as the Y-axis, and the vertical height of the road surface as the Z-axis. The road surface feature parameters are mapped to the vehicle coordinate system, and a road surface digital elevation prediction model covering the prediction area is constructed along the positive X-axis. The road surface digital elevation prediction model outputs the road surface excitation input time series sequence in the prediction time domain.

4. The steer-by-wire control method according to claim 1, characterized in that, The vehicle driving state includes at least one or more of the following parameters: Vehicle motion dynamics parameters, steering system state parameters, driving mode calibration parameters, and intelligent driving state parameters. The step of generating feedforward compensation commands based on the road excitation input time sequence and the real-time vehicle driving state includes: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact; When the vibration frequency meets the preset harmful vibration conditions, the feedforward compensation command is generated. The feedforward compensation command includes the steering system damping coefficient and rack force compensation amount to offset the corresponding axial impact.

5. The steer-by-wire control method according to claim 4, characterized in that, The method further includes: The feedforward control command is executed at the control lead time before the wheel contacts the axial impact.

6. The steer-by-wire control method according to claim 5, characterized in that, The step of executing the feedforward control command at the control lead time before the wheel contacts the axial impact includes: At the predicted moment when the vehicle is in contact with the axial impact, closed-loop compensation is performed on the feedforward control command based on the real-time feedback signal from the rack force sensor.

7. The steer-by-wire control method according to claim 1, characterized in that, The method further includes: Based on the road surface excitation input timing sequence, identify at least one axial impact caused by the road surface excitation and the vibration frequency of each axial impact; When the vibration frequency meets the preset beneficial vibration frequency band, a road feel torque feedforward command is generated. The road feel torque feedforward command includes a steering wheel feedback torque base value and a dynamic compensation torque that match the corresponding axial impact.

8. The steer-by-wire control method according to claim 7, characterized in that, The method further includes: At the predicted moment when the wheel contacts the axial impact, the road feel torque feedforward command is executed.

9. A steer-by-wire control device, characterized in that, include: The data pre-acquisition module is used to acquire road perception data of a pre-aiming area in the direction of vehicle travel. The road perception data includes at least road image data and / or three-dimensional point cloud data of the road surface in the pre-aiming area. The road surface excitation simulation module is used to extract road surface feature parameters based on the road surface sensing data, and to construct a road surface digital elevation prediction model based on the road surface feature parameters. The road surface digital elevation prediction model outputs the road surface excitation input time sequence of the prediction area. The vibration feedforward compensation module is used to generate feedforward compensation commands based on the road excitation input timing sequence and the real-time vehicle driving status. The feedforward control commands are used to instruct the steering mechanism to suppress harmful vibrations caused by the pre-aiming area.

10. A computer device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to perform the steer-by-wire control method as described in any one of claims 1 to 8.

Images