A steering control method for a distributed drive vehicle

Abstract

The application discloses a steering control method of a distributed drive vehicle, and relates to the technical field of vehicle steering control, and the method comprises the following steps: acquiring motion state information and road surface environment information of a target vehicle; determining a current adhesion coefficient of each wheel based on the motion state information and the road surface environment information; determining a target steering mode of the target vehicle and a center wheel corresponding to the target steering mode from a plurality of preset steering modes based on the current adhesion coefficient of each wheel, wherein the preset steering modes comprise a single center mode, a double center mode or a transition mode; and performing torque distribution control on each wheel based on the determined center wheel, so as to realize steering control of the target vehicle. The application realizes real-time dynamic adjustment of the steering mode and the center wheel, and effectively improves the steering stability and safety of the vehicle under complex road conditions.

Description

Technical Field

[0001] This application relates to the field of vehicle steering control technology, and in particular to a steering control method for a distributed drive vehicle. Background Technology

[0002] Distributed drive vehicles, by independently controlling the torque of each wheel, can achieve on-the-spot U-turns by locking a single wheel as the center, significantly reducing the turning radius. Currently, existing on-the-spot U-turn control technologies generally employ a fixed-center strategy, meaning that a pre-set wheel is always used as the rotation center during the U-turn. However, when the vehicle is traveling on low-traction surfaces such as water accumulation or ice / snow, the fixed-center wheel is prone to slippage due to insufficient traction, causing the vehicle's rotation trajectory to deviate or even become uncontrollable. In asymmetrical road conditions with significant differences in the coefficients of friction between the two sides, a fixed center also struggles to maintain a stable rotation axis, affecting the safety of the U-turn process. Therefore, a steering control method for distributed drive vehicles is urgently needed to solve the aforementioned technical problems. Summary of the Invention

[0003] The summary section introduces a series of simplified concepts, which will be further explained in detail in the detailed description section. This summary section is not intended to limit the key and essential technical features of the claimed technical solutions, nor is it intended to determine the scope of protection of the claimed technical solutions.

[0004] In a first aspect, this application provides a steering control method for a distributed drive vehicle, comprising: Acquire the target vehicle's motion status information and road environment information; Based on motion state information and road environment information, the current adhesion coefficient of each wheel is determined; Based on the current adhesion coefficient of each wheel, the target steering mode of the target vehicle and the center wheel corresponding to the target steering mode are determined from multiple preset steering modes. The preset steering modes include single center mode, double center mode or transition mode. Based on a defined center wheel, torque distribution control is applied to each wheel to achieve steering control of the target vehicle.

[0005] In some implementations, motion state information includes the speed difference between adjacent wheels and the lateral acceleration of the vehicle body, while road environment information includes road type and corresponding road adhesion confidence.

[0006] In some implementations, the current coefficient of adhesion for each wheel is determined based on motion state information and road surface environment information, including: The geometric correction factor for each wheel is determined based on the angle between each wheel and the line connecting the center of gravity of the target vehicle. Based on the road surface adhesion confidence level, the visual correction coefficient for each wheel is determined; Based on the speed difference, geometric correction coefficient, visual correction coefficient, and preset calibration coefficient, the current adhesion coefficient of each wheel is determined by the extended Kalman filter algorithm.

[0007] In some implementations, a target steering mode for the target vehicle is determined from a plurality of preset steering modes based on the current adhesion coefficient of each wheel, including: When the current adhesion coefficient of each wheel is greater than or equal to the first preset threshold, the single-center mode is determined as the target steering mode. When the current adhesion coefficient of any wheel is less than the second preset threshold, the dual-center mode is determined as the target steering mode. When the current adhesion coefficient of each wheel is greater than or equal to the second preset threshold and less than the first preset threshold, the transition mode is determined as the target steering mode.

[0008] In some implementations, the center wheel corresponding to the target steering pattern is determined based on the current adhesion coefficient of each wheel, including: When the target steering mode is single-center mode, the wheel with the highest current adhesion coefficient is determined as the center wheel; When the target steering mode is the dual-center mode, the wheel with the current adhesion coefficient less than the second preset threshold is identified as the low-adhesion wheel, and the two wheels that are not on the same diagonal as the low-adhesion wheel are identified as the center wheel. When the target steering mode is transition mode, two wheels are selected as the center wheels based on the current adhesion coefficient of each wheel.

[0009] In some implementations, based on the current adhesion coefficient of each wheel, two wheels are identified from among the wheels as the center wheels, including: Based on the current adhesion coefficient of each wheel, calculate the adhesion margin of each wheel. The adhesion margin is the difference between the current adhesion coefficient of that wheel and the minimum current adhesion coefficient among all wheels. The wheel with the largest attachment margin is designated as the first center wheel, and the wheel located on the same diagonal as the first center wheel is designated as the second center wheel, so that the first center wheel and the second center wheel are used as the center wheels.

[0010] In some implementations, the motion state information includes the vehicle body yaw rate, and also includes: Obtain the yaw rate of the target vehicle; Determine the target yaw rate based on the driver's steering instructions; Determine the angular velocity deviation based on the vehicle body yaw rate and the target yaw rate; When the angular velocity deviation is greater than the preset deviation threshold, the first center wheel with the largest adhesion margin is kept unchanged, and the wheel located on the same side of the vehicle body as the first center wheel is determined as the second center wheel.

[0011] In some implementations, torque distribution control is performed on each wheel based on a determined center wheel, including: When the target steering mode is single-center mode, the center wheel is locked. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the position of the central wheel, determine the single center distance from each non-central wheel to the central wheel; Based on the base torque and the distance to the single center, determine the target torque at the single center of each non-center wheel; Based on the single-center target torque of each non-center wheel, control the output torque of each non-center wheel.

[0012] In some implementations, torque distribution control is performed on each wheel based on a determined center wheel, including: When the target steering mode is the dual-center mode, the two center wheels are locked. The virtual steering axis is determined based on the positions of the two central wheels; The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual steering axis, determine the distance between the two centers of each non-center wheel and the virtual steering axis; Based on the base torque and the distance between the two centers, determine the target torque at the two centers for each non-center wheel; Based on the dual-center target torque of each non-center wheel, the output torque of each non-center wheel is controlled.

[0013] In some implementations, torque distribution control is performed on each wheel based on a determined center wheel, including: When the target steering mode is transition mode, perform a semi-locking action on the two center wheels; Based on the positions of the two central wheels, a virtual centerline is determined, where the virtual centerline is the perpendicular bisector of the line connecting the two central wheels. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual centerline, determine the transition distance from each non-center wheel to the virtual centerline; Based on the base torque and transition distance, determine the initial target torque for each non-center wheel; Based on the initial target torque of each non-center wheel, control the output torque of each non-center wheel.

[0014] In some implementations, the two central wheels are partially locked, including: Based on the target vehicle's current rotational linear velocity and wheel radius, determine the semi-lock target speed and control the two central wheels to rotate at the semi-lock target speed.

[0015] In some implementations, it also includes: When the angular velocity deviation is greater than the preset deviation threshold, the attitude correction coefficient is determined based on the vehicle body yaw rate and the target yaw rate. The initial target torque is corrected based on the attitude correction coefficient to obtain the corrected target torque; Based on the corrected target torque, the output torque of each non-center wheel is controlled.

[0016] In some implementations, it also includes: When switching the center wheel in transition mode, the speed transition control is performed smoothly between the original center wheel before replacement and the new center wheel after replacement, based on the preset transition time.

[0017] In some implementations, based on a preset transition time, a smooth transition control of the rotational speed is performed between the original center wheel before replacement and the new center wheel after replacement, including: Based on the preset transition time, the rotational speed of the original center wheel is linearly increased from the semi-locked target speed to the target speed of the non-center wheel. Based on the preset transition time, the rotational speed of the new center wheel is linearly reduced from the current speed to the semi-lock target speed.

[0018] Secondly, this application proposes a steering control device for a distributed drive vehicle, comprising: The data acquisition unit is used to acquire motion status information and road environment information of the target vehicle. The adhesion coefficient calculation unit is used to determine the current adhesion coefficient of each wheel based on motion state information and road environment information; The center wheel determination unit is used to determine the target steering mode of the target vehicle and the center wheel corresponding to the target steering mode from multiple preset steering modes based on the current adhesion coefficient of each wheel. The preset steering modes include single center mode, double center mode or transition mode. The torque steering control unit is used to control the torque distribution to each wheel based on a defined center wheel in order to achieve steering control of the target vehicle.

[0019] In summary, the steering control method for distributed drive vehicles provided in this application, by acquiring the target vehicle's motion state information and road environment information, can perceive the vehicle's current motion state and the road surface's adhesion conditions in real time, providing a data foundation for steering decisions. This allows the control strategy to make decisions based on real-time changes in vehicle dynamics and road conditions. Based on the acquired motion state information and road environment information, the current adhesion coefficient of each wheel is determined, enabling the assessment of the grip capability between each wheel and the ground. The stability limits of different wheels can be identified based on differences in adhesion coefficients. According to the current adhesion coefficient of each wheel, the target steering mode and its corresponding center wheel are determined from multiple preset steering modes, including single-center mode, double-center mode, or transition mode. This allows the steering strategy to dynamically match the most suitable control method according to real-time road adhesion conditions. For example, on high-adhesion roads, the single-center mode with the strongest adhesion is selected to ensure efficiency; on low-adhesion roads, double-center mode is used to distribute rotational torque; and on medium-adhesion or asymmetrical roads, a transition mode is entered to balance stability. This maintains the controllability of the vehicle body under different adhesion conditions. Based on a defined center wheel, torque distribution control is performed on each wheel, ensuring that the torque distribution is based on the real-time selected center wheel. This ensures that the torque output of each wheel matches the current road surface adhesion conditions and steering requirements, thereby achieving stable steering control of the target vehicle under complex and changing road conditions. Attached Figure Description

[0020] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit this specification. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic flowchart of a steering control method for a distributed drive vehicle provided in an embodiment of this application; Figure 2 This is a schematic diagram of torque distribution in a dual-center mode provided in an embodiment of this application; Figure 3 This is a schematic diagram of a steering control device for a distributed drive vehicle provided in an embodiment of this application. Detailed Implementation

[0021] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them.

[0022] The steering control method for distributed drive vehicles proposed in this application can be widely applied to various electric vehicles using wheel-side or hub motors for independent drive, and is particularly suitable for application scenarios involving U-turns or small-radius steering maneuvers in complex and variable road conditions. Specifically, when the vehicle is traveling on low-friction surfaces such as water accumulation or ice and snow, or encountering asymmetrical road conditions with significant differences in the friction conditions on both sides, the method of this application can dynamically adjust the steering control strategy by sensing the vehicle's motion state and road environment information in real time, thereby ensuring the vehicle's steering stability and safety under extreme conditions. This method can be integrated into the vehicle's existing chassis control system or intelligent driving assistance system as a core function of distributed drive vehicles, and is applicable to various practical driving scenarios such as urban commuting, off-road driving, and parking in tight spaces.

[0023] To illustrate the technical solution of this application, several specific terms appearing in this application will be explained.

[0024] The distributed drive vehicle of this application refers to a vehicle in which each wheel is driven and controlled by an independent motor, and has the characteristic that the torque of each wheel is independently controllable.

[0025] The central wheel in this application refers to the wheel selected as the rotation reference during vehicle steering, around which the vehicle rotates or around a virtual rotation center composed of multiple central wheels.

[0026] The current adhesion coefficient in this application refers to the quantitative value of the adhesion ability between each wheel and the road surface estimated in real time through sensor information, which is used to characterize the grip limit of the wheel at the current moment.

[0027] The single-center mode of this application refers to a steering control mode in which only a single wheel is locked as the center wheel on a high-adhesion road surface, and the remaining wheels rotate around the center.

[0028] The dual-center mode of this application refers to a steering control mode that uses two diagonally opposite wheels as rotation centers on low-adhesion road surfaces, with the two centers jointly bearing the rotational torque.

[0029] The transition mode of this application refers to a steering control mode that uses two wheels as the center and performs semi-lock control on medium-adhesion or asymmetrical road surfaces, and distributes torque according to the adjusted virtual rotation center.

[0030] The adhesion margin in this application is used to characterize the degree of excess current adhesion capability of a certain wheel, specifically the difference between the current adhesion coefficient of the wheel and the minimum adhesion coefficient among all wheels.

[0031] The virtual steering axis of this application refers to an imaginary straight line formed by the line connecting two locked diagonal center wheels in the dual-center mode, around which the non-center wheels distribute torque.

[0032] The virtual centerline in this application refers to the imaginary straight line formed by the perpendicular bisector of the line connecting the two semi-locked center wheels in the transition mode, which is used as a rotational reference for calculating the torque distribution distance of the non-center wheels.

[0033] The semi-locking in this application refers to applying a non-completely locked control state to the central wheel, causing it to rotate at a preset low-speed target speed to buffer the impact during mode switching.

[0034] The semi-lock target speed of this application refers to the small rotational speed maintained by the center wheel in the transition mode. This speed is determined by multiplying the ratio of the current vehicle body rotation linear velocity to the wheel radius by a preset coefficient.

[0035] The angular velocity deviation in this application refers to the difference between the actual yaw rate obtained by the vehicle body sensor and the target yaw rate determined according to the driver's operation, which is used to characterize the degree of deviation of the vehicle body attitude.

[0036] The attitude correction coefficient in this application refers to the torque adjustment coefficient calculated based on the angular velocity deviation, which is used to amplify and correct the torque of non-centered wheels when the vehicle body attitude deviates.

[0037] This application can be widely applied to various vehicles employing a distributed drive architecture, including but not limited to electric off-road vehicles, urban multi-purpose vehicles, special-purpose vehicles, and intelligent connected vehicles with U-turn capabilities. It is particularly suitable for applications requiring small-radius turning or U-turns under complex adhesion conditions such as flooded roads, icy roads, oily roads, and asymmetrical road surfaces that are partially dry and partially slippery. The above-mentioned terminology definitions are consistent throughout the various technical features of this application, providing a clear semantic foundation for understanding the technical solution of this application.

[0038] Please see Figure 1 The above is a schematic flowchart of a steering control method for a distributed drive vehicle provided in an embodiment of this application, including: S110. Obtain the target vehicle's motion status information and road environment information; For example, the real-time rotational speed of each wheel is collected by the wheel speed sensors configured in the vehicle, and the speed difference between adjacent wheels is calculated accordingly; the yaw rate and lateral acceleration of the vehicle body are collected by the inertial navigation module; and images of the road surface ahead are acquired by a high-definition vision camera, and the current road surface type and its corresponding road adhesion confidence level are obtained after image recognition processing. The above wheel speed information, vehicle attitude information, and road surface visual information together constitute the motion state information and road environment information of the target vehicle.

[0039] S120. Based on motion state information and road environment information, determine the current adhesion coefficient of each wheel; For example, based on the acquired motion state information and road environment information, the multi-source data is fused using an extended Kalman filter algorithm to estimate the current adhesion coefficient of each wheel. During this fusion process, the relative slippage trend of the wheels is reflected by the rotational speed difference between adjacent wheels, the lateral force state of the vehicle is reflected by the lateral acceleration of the vehicle body, geometric position corrections are introduced based on the angle between each wheel and the line connecting the vehicle's center of gravity, and visual perception corrections are introduced based on the road adhesion confidence level. Using these parameters, the current adhesion coefficient between each wheel and the current road surface is output.

[0040] S130. Based on the current adhesion coefficient of each wheel, determine the target steering mode of the target vehicle and the center wheel corresponding to the target steering mode from multiple preset steering modes, wherein the preset steering modes include single center mode, double center mode or transition mode. For example, the current road surface adhesion condition is determined based on the distribution of the current adhesion coefficients of each wheel. If the adhesion coefficients of all wheels are at a high level, a single-center mode is selected, with the wheel with the highest adhesion coefficient as the center wheel. If the adhesion coefficient of a wheel is too low, indicating that the wheel is in a low-adhesion area, a dual-center mode is activated, selecting two wheels diagonally opposite the low-adhesion wheel as center wheels, so that the rotational torque is shared by the two wheels. If the adhesion coefficients of all wheels are at a medium level or exhibit an asymmetrical distribution, a transition mode is entered, prioritizing the wheel with the largest adhesion margin and its diagonal wheel as center wheels, and adjusting the center combination according to changes in vehicle attitude to balance steering stability.

[0041] S140. Based on the determined center wheel, torque distribution control is performed on each wheel to achieve steering control of the target vehicle.

[0042] For example, depending on the selected target steering mode, locking or semi-locking control is applied to the center wheel, and the target torque is calculated and output to the non-center wheels. In single-center mode, the center wheel is fully locked, and the non-center wheels distribute torque according to their respective distances from the center. In dual-center mode, the two center wheels are fully locked to form a virtual steering axis, and the non-center wheels distribute reverse torque around this axis. In transition mode, the two center wheels are semi-locked and rotate at a preset low speed. The torque of the non-center wheels is calculated based on the virtual centerline, and a correction coefficient is introduced to adjust the torque output according to the vehicle body attitude deviation. At the same time, a smooth speed transition strategy is adopted when switching center wheels to avoid shock.

[0043] In summary, this application provides a data foundation for steering control decisions by acquiring the target vehicle's motion state information and road environment information, enabling the control strategy to be dynamically adjusted based on the vehicle's current actual motion state and road adhesion conditions. By determining the current adhesion coefficient of each wheel based on motion state information and road environment information, the grip capability between each wheel and the ground is assessed, and the stability limits of different wheels can be identified based on real-time changes in the adhesion coefficient. By determining the target steering mode of the target vehicle and the corresponding center wheel from multiple preset steering modes based on the current adhesion coefficient of each wheel, the preset steering modes include a single-center mode, a double-center mode, or a transition mode. This allows the steering strategy to actively match the most suitable control method according to real-time changes in road adhesion conditions. For example, on high-adhesion roads, the single-center mode with the strongest adhesion is selected to ensure steering efficiency; on low-adhesion roads, a double-center mode is used to distribute rotational torque and prevent single-wheel slippage; and on medium-adhesion or asymmetrical roads, a transition mode is entered to balance stability, maintaining the controllability of the vehicle body under different adhesion conditions. By controlling the torque distribution of each wheel based on a determined center wheel, the torque distribution can be based on the real-time selected center wheel, ensuring that the torque output of each wheel matches the current road surface adhesion conditions and steering requirements. This achieves stable steering control of the target vehicle under complex and variable road conditions, improving the steering stability and trajectory accuracy of the vehicle in complex scenarios such as water accumulation, ice and snow, and asymmetrical road surfaces.

[0044] In some instances, motion state information includes the speed difference between adjacent wheels and the lateral acceleration of the vehicle body, while road environment information includes the road surface type and the corresponding road adhesion confidence level.

[0045] For example, wheel speed sensors installed at each wheel collect the rotational angular velocity of each wheel in real time, and calculate the rotational speed difference between adjacent wheels based on the difference in angular velocity. This rotational speed difference directly reflects whether there is a relative slippage tendency between the wheels. When the rotational speed of a certain wheel is significantly higher than that of the adjacent wheels on the same axle or side, it indicates that the wheel may be over-rotating due to low traction on the road surface. At the same time, the lateral acceleration of the vehicle body is detected in real time by the inertial measurement unit built into the vehicle. This parameter characterizes the magnitude of the lateral inertial force experienced by the vehicle during steering, and can indirectly reflect the lateral dynamic state of the entire vehicle.

[0046] Furthermore, high-definition vision cameras deployed at the front of the vehicle capture real-time images of the road surface ahead of the vehicle's travel direction. The captured image data is then processed by a pre-trained deep learning semantic segmentation model. This model is trained on a large-scale image dataset containing various typical road surface scenarios, including dry asphalt, standing water, ice and snow, and oil stains. During training, pixel-level annotations enable the model to learn visual features such as texture, color, and reflectivity corresponding to different road surface types. In the inference phase, the model predicts the classification probability of each pixel in the input image, outputting the probability distribution of each pixel belonging to various road surface types. Then, by statistically analyzing and fusing the pixel probabilities in the area corresponding to the vehicle's travel trajectory, the specific type of the current road surface is identified. Simultaneously, the model outputs the normalized probability value corresponding to the identification result as the road surface adhesion confidence score. This confidence score, ranging from 0 to 1, characterizes the reliability of the visual perception module's judgment of the current road surface type. A higher confidence score indicates a more reliable identification result, providing a quantifiable visual correction basis for adhesion coefficient estimation.

[0047] In summary, this application utilizes the speed difference between adjacent wheels, lateral acceleration of the vehicle body, and road surface type with confidence level as motion state and road environment information to achieve perception of vehicle dynamic response and road surface adhesion conditions. The speed difference between adjacent wheels can keenly capture the precursory features of wheel slippage, providing a quantitative basis for the degree of wheel slippage in adhesion coefficient estimation, enabling the control strategy to identify risks before slippage occurs. Lateral acceleration of the vehicle body reflects the lateral force limit of the vehicle during steering, introducing vehicle dynamics constraints into the adhesion coefficient calculation, helping to distinguish between sideslip caused by insufficient road surface adhesion and lateral forces generated by normal steering. Road surface type with confidence level provides prior adhesion characteristics of the road surface ahead, allowing for early prediction of low-adhesion areas through visual recognition, enabling adhesion coefficient estimation to incorporate visual correction and reduce estimation errors caused by sensor noise or delay. The fusion of these multi-source information lays the data foundation for determining the current adhesion coefficient of each wheel, ensuring that the adhesion coefficient accurately reflects the real-time grip capability between each wheel and the road surface. This allows steering mode and center wheel decisions to be based on reliable adhesion information, improving vehicle steering stability in complex road conditions.

[0048] In some instances, the current coefficient of adhesion for each wheel is determined based on motion state information and road surface environment information, including: The geometric correction factor for each wheel is determined based on the angle between each wheel and the line connecting the center of gravity of the target vehicle. Based on the road surface adhesion confidence level, the visual correction coefficient for each wheel is determined; Based on the speed difference, geometric correction coefficient, visual correction coefficient, and preset calibration coefficient, the current adhesion coefficient of each wheel is determined by the extended Kalman filter algorithm.

[0049] For example, a geometric correction coefficient is determined based on the angle between each wheel and the line connecting the target vehicle's center of gravity. This angle is a fixed parameter determined by the vehicle chassis design, reflecting the positional relationship of each wheel relative to the vehicle's center of gravity in geometric space. Since the wheel adhesion limit is related to the direction of the lateral force, the geometric correction coefficient weights the projections of the vehicle's lateral acceleration onto different wheels, allowing the adhesion coefficient estimation to account for the influence of wheel position on the distribution of lateral force. Simultaneously, a visual correction coefficient is determined based on the road adhesion confidence level. The road adhesion confidence level is characterized by the normalized probability value of the visual perception module's identification of the road surface type ahead; a higher confidence level indicates a more reliable visual recognition result. The visual correction coefficient is dynamically adjusted according to the road adhesion confidence level. When low-adhesion road surfaces such as water accumulation or ice / snow are identified with a high confidence level, the visual correction coefficient is reduced accordingly, lowering the estimated value of the corresponding wheel in advance during adhesion coefficient estimation to avoid potential slippage risks.

[0050] Based on the speed difference between adjacent wheels, the vehicle's lateral acceleration, geometric correction coefficients, visual correction coefficients, and preset calibration coefficients, the current adhesion coefficient of each wheel is determined using an extended Kalman filter algorithm. The mathematical expression for this estimation process is μ. i =k1×Δω i +k2×a ×cos(θ i )+ε, where μ i Δω represents the current adhesion coefficient of the i-th wheel. i The difference in rotational speed between adjacent wheels is collected and calculated in real time by wheel speed sensors, directly reflecting the relative slippage trend of the wheels; a The lateral acceleration of the vehicle body is acquired by the inertial navigation module and represents the lateral force state of the entire vehicle; cos(θ) i () is the geometric correction factor, based on the angle θ between the line connecting the wheel and the center of gravity. i The calculation involves k1 and k2, which are preset calibration coefficients obtained through real-vehicle experiments to adapt to the dynamic characteristics of different vehicle models; ε is a visual correction coefficient, dynamically adjusted based on road adhesion confidence. The extended Kalman filter algorithm uses the above multi-source information as observation input, combines it with the vehicle dynamics model for state estimation, and filters out sensor noise through iterative prediction and update steps to output the current adhesion coefficient between each wheel and the road surface. This algorithm can achieve an estimation delay of less than 30 milliseconds, meeting real-time control requirements and ensuring that the adhesion coefficient output can respond promptly to changes in the road and vehicle states.

[0051] In summary, by introducing a geometric correction coefficient based on geometric position, the adhesion coefficient estimation in this application can take into account the angle difference between each wheel and the center of gravity line, improving the accuracy of the estimation results under different steering states. Introducing a visual correction coefficient based on road surface adhesion confidence allows prior information from visual perception to be effectively integrated into the estimation process. When a low-adhesion road surface is visually detected, the adhesion coefficient estimation value is proactively reduced through the visual correction coefficient, enabling early avoidance of slippage risks and enhancing the feedforward response capability to sudden road surface changes. By fusing the speed difference, vehicle lateral acceleration, geometric correction coefficient, visual correction coefficient, and preset calibration coefficient using an extended Kalman filter algorithm, the complementary advantages of multi-source information are utilized, and sensor noise and instantaneous disturbances are suppressed through a filtering mechanism, outputting a reliable current adhesion coefficient. This provides a quantitative basis for steering mode and center wheel decisions, supporting stable and controllable steering of the vehicle under complex road conditions.

[0052] In some instances, the target steering mode for the target vehicle is determined from multiple preset steering modes based on the current adhesion coefficient of each wheel, including: When the current adhesion coefficient of each wheel is greater than or equal to the first preset threshold, the single-center mode is determined as the target steering mode. When the current adhesion coefficient of any wheel is less than the second preset threshold, the dual-center mode is determined as the target steering mode. When the current adhesion coefficient of each wheel is greater than or equal to the second preset threshold and less than the first preset threshold, the transition mode is determined as the target steering mode.

[0053] For example, when the current coefficient of friction of all four wheels is greater than or equal to a first preset threshold, it indicates that all wheels are on a high-friction surface, such as dry asphalt or concrete. At this time, each wheel has sufficient grip to provide stable support for the vehicle. In this case, the single-center mode is determined as the target steering mode. This mode locks a single wheel as the center of rotation and uses the torque difference of the remaining wheels to drive the vehicle body to rotate around this center. It can achieve minimum radius steering while ensuring stability and fully utilize the advantages of high-friction surfaces.

[0054] When the current adhesion coefficient of any wheel is less than the second preset threshold, it indicates that the wheel is on a low-adhesion surface, such as a waterlogged, icy, or oily area. Its adhesion is insufficient to resist the rotational torque. If this wheel is used as the center, slippage is likely to occur, leading to loss of steering control. In this case, the dual-center mode is determined as the target steering mode. This mode locks two wheels located diagonally away from the low-adhesion wheel as the rotation center. By having the two centers share the vehicle's rotational torque, the load on individual wheels is distributed, preventing the vehicle's rotation axis from being disrupted due to slippage of a single low-adhesion wheel, thus maintaining the stability of the steering process under low-adhesion conditions.

[0055] When the current adhesion coefficient of all wheels is greater than or equal to the second preset threshold and less than the first preset threshold, it indicates that the vehicle is driving on a medium-adhesion road surface or encountering asymmetrical adhesion conditions, such as a semi-dry road surface after rain or a mixed road surface that is half dry and half wet. At this time, the adhesion capability of each wheel is between low and high adhesion. If a single-center mode is used, the center may slip due to insufficient adhesion. If a dual-center mode is used, the steering smoothness may be affected by excessive constraint. Therefore, the transition mode is determined as the target steering mode. This mode selects two wheels as candidate centers and performs semi-lock control. At the same time, torque is distributed based on the virtual centerline. It can balance steering stability under medium or asymmetrical adhesion conditions and provide a basis for a smooth transition for possible center switching. In this embodiment, the first preset threshold is 0.6 and the second preset threshold is 0.3.

[0056] In summary, this application embodiment achieves adaptive matching between steering mode and road surface adhesion conditions through the aforementioned hierarchical judgment mechanism based on the current adhesion coefficient and a preset threshold. When the adhesion coefficients of all wheels reach the first preset threshold, a single-center mode is activated, fully utilizing the grip potential of high-adhesion surfaces to achieve efficient steering. When low-adhesion wheels are present, a dual-center mode is activated, effectively avoiding the risk of slippage in a single-center mode by dispersing torque through dual centers, ensuring steering controllability under low-adhesion conditions. When the adhesion coefficient is in the medium range, a transition mode is activated, avoiding both slippage that may occur in a single-center mode and excessive constraint in a dual-center mode, providing a stable steering strategy for medium-adhesion and asymmetric road surfaces. This hierarchical decision-making method discretizes the continuous changes in road conditions into three modes with clear physical meanings, enabling steering control to actively adapt to the real-time perceived adhesion conditions, improving the vehicle's steering adaptability and robustness under complex and changing road conditions.

[0057] In some instances, the center wheel corresponding to the target steering pattern is determined based on the current adhesion coefficient of each wheel, including: When the target steering mode is single-center mode, the wheel with the highest current adhesion coefficient is determined as the center wheel; When the target steering mode is the dual-center mode, the wheel with the current adhesion coefficient less than the second preset threshold is identified as the low-adhesion wheel, and the two wheels that are not on the same diagonal as the low-adhesion wheel are identified as the center wheel. When the target steering mode is transition mode, two wheels are selected as the center wheels based on the current adhesion coefficient of each wheel.

[0058] For example, when the target steering mode is single-center mode, it indicates that the current coefficient of friction of all wheels is greater than or equal to a first preset threshold, meaning the vehicle is traveling on a high-friction surface such as dry asphalt or cement, and each wheel has sufficient grip. In this case, the wheel with the highest current coefficient of friction among the four wheels is determined as the center wheel. This wheel has optimal grip and can withstand the maximum static friction generated during vehicle rotation. By fully locking this center wheel as the center of rotation, the other non-center wheels distribute torque according to their distance from the center, generating driving torque to drive the vehicle to rotate smoothly around this center. This achieves minimum radius steering while ensuring stability, fully utilizing the advantages of high-friction surfaces.

[0059] When the target steering mode is the dual-center mode, the wheel with a current adhesion coefficient less than the second preset threshold is identified as a low-adhesion wheel. This wheel lacks sufficient traction due to being on low-adhesion surfaces such as water or ice, and cannot independently act as the center of rotation. To prevent this low-adhesion wheel from being selected as the center and causing slippage and loss of control, two wheels not on the same diagonal as the low-adhesion wheel are designated as the center wheels. These two wheels form a diagonal pair, avoiding the weak adhesion area. By fully locking these two center wheels to create a virtual steering axis, the vehicle's rotational torque is shared by the two diagonal wheels, distributing the load across individual wheels and effectively preventing steering axis deviation caused by low-adhesion wheel slippage, ensuring vehicle steering stability on low-adhesion surfaces.

[0060] When the target steering mode is transitional, it indicates that the current adhesion coefficient of all wheels is between the second preset threshold and the first preset threshold. This means the vehicle is traveling on an asymmetrical road surface, either partially dry after rain or partially wet. The adhesion of each wheel is insufficient to support complete locking in single-center mode, nor does it reach the low adhesion level required to activate dual-center mode. To achieve stable steering within this transitional range, two wheels are selected as center wheels based on the current adhesion coefficient of each wheel. This selection process considers the relative strength of each wheel's adhesion, prioritizing wheel combinations with stronger adhesion and the ability to form a stable axis of rotation. This lays the foundation for semi-lock control and torque distribution. The specific selection method can be determined by calculating parameters such as adhesion margin based on the relative magnitude of the adhesion coefficients. This ensures that the selected two center wheels can maintain vehicle controllability during rotation and provide a smooth transition for possible dynamic center adjustments.

[0061] In summary, this application achieves the matching of steering mode and center wheel by determining the center wheel based on the current adhesion coefficient. In single-center mode, the wheel with the highest adhesion coefficient is selected as the center, utilizing the grip potential of the high-adhesion road surface to ensure the center has anti-slip capability. In dual-center mode, by identifying the low-adhesion wheel and selecting two wheels not on the same diagonal as it as the center, weak adhesion areas are avoided, and the rotational torque is shared by the two diagonal wheels, reducing the load on individual wheels, suppressing the risk of slippage on low-adhesion roads, and improving the reliability of the steering process. In transition mode, two center wheels are selected from each wheel based on the adhesion coefficient, providing a more adaptable combination of rotation centers for medium adhesion and asymmetrical road conditions, enabling steering control to be dynamically adjusted according to real-time adhesion conditions, avoiding the failure problem of the fixed-center strategy when adhesion changes. The aforementioned method for determining the center of the circle discretizes the continuous changes in road surface adhesion conditions into a decision basis with physical meaning, ensuring that steering control always focuses on the wheel with the strongest current adhesion or the most reasonable wheel combination. This maintains the stable rotation trajectory of the vehicle body in complex scenarios such as water accumulation, ice and snow, and asymmetrical road surfaces, thereby improving the safety and adaptability of vehicle steering.

[0062] In some instances, based on the current adhesion coefficient of each wheel, two wheels are identified as the center wheels from among the wheels, including: Based on the current adhesion coefficient of each wheel, calculate the adhesion margin of each wheel. The adhesion margin is the difference between the current adhesion coefficient of that wheel and the minimum current adhesion coefficient among all wheels. The wheel with the largest attachment margin is designated as the first center wheel, and the wheel located on the same diagonal as the first center wheel is designated as the second center wheel, so that the first center wheel and the second center wheel are used as the center wheels.

[0063] For example, when determining two center wheels in transition mode, the adhesion margin of each wheel is calculated based on its current adhesion coefficient. The adhesion margin is defined as the difference between the current adhesion coefficient of that wheel and the minimum current adhesion coefficient among all wheels. This calculation quantifies the relative traction capacity of each wheel under current road conditions by comparing the adhesion coefficient of each wheel one by one with the global minimum. A larger adhesion margin indicates that the wheel has stronger grip compared to the wheel with the weakest adhesion, is less prone to slippage during rotation, and is therefore more suitable to serve as the center wheel. By calculating the adhesion margin, even when the adhesion coefficient is in a moderate range or asymmetrically distributed, the wheel with relatively strong adhesion can be identified from the four wheels, avoiding the bias that may result from judging solely by the absolute value of the adhesion coefficient, thus making the determination of the center wheel more accurate.

[0064] After calculating the adhesion margin, the wheel with the largest adhesion margin is designated as the first center wheel. This wheel has optimal adhesion under the current operating conditions, providing the most reliable grip point during rotation and ensuring sufficient anti-slip capability at the center of rotation. The wheel located on the same diagonal as the first center wheel is designated as the second center wheel. Choosing the diagonal wheel as the second center is based on considerations of vehicle geometry and rotational mechanics. The two diagonally positioned wheels are symmetrically distributed relative to the vehicle's center of gravity. When these two wheels serve as a common rotation reference, they form a virtual rotation axis passing through the region near the vehicle's center of gravity, allowing for a balanced distribution of rotational torque between the front and rear axles. Even if the adhesion margin of the second center wheel is not the second largest, the diagonal combination ensures the geometric symmetry of the rotation axis, enabling the vehicle to rotate stably around the center of gravity and avoiding vehicle deflection or rotational trajectory distortion caused by the center being biased to one side. In this way, the first and second center wheels are jointly designated as the two center wheels in the transition mode, providing execution targets for semi-lock control and torque distribution.

[0065] In summary, this embodiment calculates the adhesion margin and determines the first center wheel based on the margin size, achieving a relative ranking of wheel adhesion capabilities. This eliminates the reliance on absolute threshold judgments for center wheel selection, allowing dynamic identification of the wheel with the strongest adhesion under continuously changing adhesion coefficients. This provides a reliable grip point for the rotation center and reduces the risk of center slippage due to insufficient adhesion. By defining the wheel located on the same diagonal as the first center wheel as the second center wheel, the geometric symmetry of the diagonal position relative to the vehicle's center of gravity is utilized. This allows the two center wheels to jointly form a virtual rotation axis passing near the center of gravity, ensuring a balanced distribution of vehicle rotation torque between the front and rear axles. This avoids torque imbalance and rotation trajectory deviation caused by the center being concentrated on one side. The combination of adhesion margin calculation and diagonal selection enables the center wheel in the transition mode to have both the advantage of strong single-point adhesion capability and the structural characteristics of two-point geometric symmetry. In medium adhesion or asymmetrical road conditions, it provides a rotational reference with both stability and balance for steering control, improving the steering trajectory accuracy and vehicle attitude controllability in complex road conditions.

[0066] In some instances, motion state information includes the vehicle body yaw rate, and also includes: Obtain the yaw rate of the target vehicle; Determine the target yaw rate based on the driver's steering instructions; Determine the angular velocity deviation based on the vehicle body yaw rate and the target yaw rate; When the angular velocity deviation is greater than the preset deviation threshold, the first center wheel with the largest adhesion margin is kept unchanged, and the wheel located on the same side of the vehicle body as the first center wheel is determined as the second center wheel.

[0067] For example, during steering control in transition mode, motion state information includes the vehicle's yaw rate, which is collected in real time by the inertial navigation module. This parameter directly reflects the actual speed at which the vehicle rotates around its vertical axis. Simultaneously, based on the driver's input steering commands, such as steering wheel angle or steering mode selection signals, the desired vehicle rotation rate is analyzed and determined as the target yaw rate. The real-time collected yaw rate is compared with this target yaw rate, and the difference between the two is calculated and its absolute value is taken to obtain the angular velocity deviation. This deviation quantifies the degree of deviation between the vehicle's actual rotation state and the driver's desired rotation state. When the deviation exceeds a preset deviation threshold, it indicates that the current vehicle attitude has significantly deviated, requiring adjustment of the center wheel assembly to correct the attitude.

[0068] When the angular velocity deviation exceeds a preset deviation threshold, the original first center wheel remains unchanged. This wheel has the largest current adhesion margin and strong grip, providing a stable rotational reference for attitude correction. Based on this, the wheel located on the same side of the vehicle as the first center wheel is designated as the new second center wheel, replacing the original second center wheel which might have been diagonally opposite. Here, "same side of the vehicle" refers to the left or right side of the vehicle where the first center wheel is located. For example, if the first center wheel is the left front wheel, the wheel on the same side is the left rear wheel; if the first center wheel is the right front wheel, the wheel on the same side is the right rear wheel. By adjusting the second center wheel to the same side as the first center wheel, the line connecting the two center wheels will be closer to one side of the vehicle. The resulting virtual rotation axis will also shift accordingly, causing the torque distribution of the non-center wheels to generate a stronger corrective torque, thereby quickly correcting the vehicle's attitude and causing the actual yaw rate to converge towards the target value.

[0069] In summary, this embodiment of the application achieves adaptive correction of the center wheel combination in transition mode through the aforementioned adjustment based on angular velocity deviation. First, the vehicle body yaw rate is acquired and its deviation from the target value is calculated, providing a basis for quantifying attitude deviation and enabling the system to promptly detect changes in the vehicle body's rotation state, preventing loss of control due to accumulated attitude deviation. Second, maintaining the first center wheel with the largest adhesion margin unchanged ensures the grip stability of the rotation reference, avoiding the introduction of new slip risks due to changing the main center. Third, adjusting the second center wheel to be on the same side as the first center wheel alters the spatial position of the virtual rotation axis, creating an asymmetry in the lever arm distribution of the non-center wheels, forming a corrective torque against the current attitude deviation direction, thus achieving active correction of vehicle body rotation. This adjustment mechanism allows steering control in transition mode to not only adapt to road adhesion conditions but also dynamically optimize the rotation center configuration based on real-time vehicle body feedback, effectively suppressing attitude deviation caused by road asymmetry or external disturbances, and improving the vehicle's ability to maintain steering trajectory and vehicle stability under medium adhesion and complex road conditions.

[0070] In some instances, torque distribution control is performed on each wheel based on a defined center wheel, including: When the target steering mode is single-center mode, the center wheel is locked. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the position of the central wheel, determine the single center distance from each non-central wheel to the central wheel; Based on the base torque and the distance to the single center, determine the target torque at the single center of each non-center wheel; Based on the single-center target torque of each non-center wheel, control the output torque of each non-center wheel.

[0071] For example, when the target steering mode is determined to be a single-center mode, it indicates that the current coefficient of friction of all wheels is at a high level, providing sufficient grip. In this case, a locking action is performed on the selected center wheel. This center wheel is the wheel with the highest current coefficient of friction among the four wheels, possessing optimal grip and capable of withstanding the maximum static friction generated during vehicle rotation. The locking action is specifically implemented by sending a zero-speed command to the motor controller of this wheel, causing its output torque to be zero and keeping it stationary, thus physically forming a fixed center of rotation and providing a stable geometric reference for the vehicle body to rotate around this center.

[0072] The calculation of the basic torque is based on the vehicle's inherent dynamic parameters and the driver's intention, as well as the vehicle's moment of inertia I. zThese are inherent constants determined during vehicle design, reflecting the magnitude of the vehicle body's inertia against rotational motion; the target angular acceleration α is derived in real-time from the driver's steering input via the steering wheel or function switches, characterizing the desired rate of change of the vehicle body's rotation speed; the wheel radius r is a fixed geometric dimension of the wheel. These parameters are then combined according to I... z The base torque is calculated by α / r, which reflects the torque required per unit lever arm to drive the vehicle body to rotate at the target angular acceleration under ideal rotation conditions.

[0073] After obtaining the base torque, a specific target torque needs to be assigned to each non-centered wheel, depending on the spatial position of each wheel relative to the centered wheel. Based on the geometric coordinates of the centered wheel, the single-center distance L from each non-centered wheel to that centered wheel is calculated. j This distance is the torque arm. According to the torque formula T... j =(I z ×α) / (r·L j The base torque is divided by the distance to the center of each non-center wheel to obtain the target torque for each non-center wheel. This formula shows that the farther the wheel is from the center, the smaller the torque required, and vice versa, thus ensuring that the driving torque generated by each wheel can work together to drive the vehicle body to rotate smoothly around the center. The calculated target torque for each center is then sent to the motor controller of the corresponding wheel to control each non-center wheel to output the corresponding torque, achieving steering motion around a fixed center.

[0074] In summary, this embodiment of the application, through torque distribution control in the single-center mode described above, uses the wheel with the highest current coefficient of adhesion as the rotation center, ensuring that the center has anti-slip capability and providing a fixed reference for the entire steering process. Introducing a base torque quantifies the driver's desired rotational acceleration and the vehicle's inherent inertia, enabling torque distribution to accurately respond to the driver's intentions. By calculating the distance from each non-center wheel to the center as the lever arm and distributing torque according to the inverse relationship of the lever arm, it ensures that after distributing torque to each wheel according to the inverse relationship with the distance to the center, the driving force output by each wheel can work synergistically around the center, jointly driving the vehicle body to rotate smoothly around the center. This avoids unnecessary translation or swaying of the vehicle body due to improper torque distribution, achieving the minimum radius U-turn function on high-adhesion surfaces and improving the accuracy and controllability of the steering trajectory.

[0075] In some instances, torque distribution control is performed on each wheel based on a defined center wheel, including: When the target steering mode is the dual-center mode, the two center wheels are locked. The virtual steering axis is determined based on the positions of the two central wheels; The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual steering axis, determine the distance between the two centers of each non-center wheel and the virtual steering axis; Based on the base torque and the distance between the two centers, determine the target torque at the two centers for each non-center wheel; Based on the dual-center target torque of each non-center wheel, the output torque of each non-center wheel is controlled.

[0076] For example, please refer to Figure 2 This is a schematic diagram of torque distribution in a dual-center mode provided in an embodiment of this application. When the target steering mode is determined to be a dual-center mode, a locking action is first performed on the two selected center wheels. Figure 2 As shown, based on the current adhesion coefficient distribution, the area where the left front wheel C is located is identified as a low-adhesion icy and snowy surface, and its current adhesion coefficient is less than the second preset threshold, so it is determined to be a low-adhesion wheel; the right front wheel A and the left rear wheel B, which are not on the same diagonal, are selected as the center wheels. By sending a zero-speed command to the motor controllers of the right front wheel A and the left rear wheel B, these two wheels are completely locked and kept stationary, physically forming two fixed rotational fulcrums, providing a geometric reference for torque distribution.

[0077] A virtual steering axis is determined based on the spatial positions of the two locking center wheels on the vehicle. For example... Figure 2 As shown, the virtual steering axis is a straight line connecting the right front wheel A and the left rear wheel B, and its spatial orientation is uniquely determined by the geometric coordinates of these two wheels. During vehicle steering, the vehicle body will rotate around this virtual steering axis, rather than around a single fixed point. The introduction of the virtual steering axis expands the center of rotation from a single point to a straight line, thereby distributing the rotational torque to the two diagonally opposite wheels and avoiding the risk of slippage caused by insufficient attachment to a single center. At the same time, this axis passes through the area near the vehicle's center of gravity, ensuring a balanced distribution of rotational torque between the front and rear axles.

[0078] To quantify the torque output of each non-central wheel, a base torque is calculated based on the vehicle's inherent parameters and the driver's intent. The base torque is determined by the vehicle's moment of inertia, target angular acceleration, and wheel radius. The formula is: base torque = vehicle moment of inertia multiplied by target angular acceleration divided by wheel radius. The vehicle's moment of inertia is a constant determined during vehicle design, representing the vehicle's resistance to rotational motion. The target angular acceleration is derived in real-time from the driver's steering input via the steering wheel or function switches, reflecting the desired rate of change of vehicle rotation. The wheel radius is a fixed geometric dimension. The base torque represents the torque requirement per unit lever arm required to drive the vehicle to rotate at the target angular acceleration under ideal rotational conditions.

[0079] After obtaining the base torque, a specific target torque needs to be assigned to each non-centered wheel, depending on the spatial position of each non-centered wheel relative to the virtual steering axis. For each non-centered wheel, its vertical distance to the virtual steering axis is calculated; this distance represents the lever arm length of that wheel in the current rotation mode. According to the torque distribution principle, the double-centered target torque for each non-centered wheel is obtained by dividing the base torque by its corresponding vertical distance, and a positive or negative sign is assigned according to the direction of rotation to characterize the direction of torque application. Figure 2 As shown, the left front wheel C is located to the left of the virtual steering axis. To make the vehicle rotate in the positive direction around the axis, a reverse torque of -150 Nm is applied to it. The right rear wheel D is located to the right of the axis, and a positive torque of +160 Nm is applied to it. The torques of the two non-centered wheels are in opposite directions, jointly driving the vehicle to rotate smoothly around the virtual steering axis. The calculated dual-centered target torques are sent to the motor controllers of the corresponding wheels, controlling each non-centered wheel to output the specified torque, thereby realizing steering control in dual-centered mode.

[0080] In summary, this application's embodiment improves steering stability on low-adhesion surfaces through torque distribution control in the aforementioned dual-center mode. First, based on the current adhesion coefficient, low-adhesion wheels are identified, and two wheels diagonally opposite to them are selected as centers and locked. This ensures the rotation center avoids weak adhesion areas, preventing steering loss due to slippage at a single center. Second, a virtual steering axis is constructed using the two locked center wheels, expanding the rotation center from a point to a line. This allows the vehicle's rotational torque to be shared by the two diagonally opposite wheels, distributing the load across individual wheels and suppressing the risk of slippage on low-adhesion surfaces. Third, the calculation of the base torque quantifies the driver's desired rotational acceleration and the vehicle's inherent inertia, enabling torque distribution to respond to operational intentions. Finally, torque distribution is performed based on the vertical distance from each non-center wheel to the virtual steering axis, and the torque is assigned a positive or negative sign according to the direction of rotation. This ensures that the torque output by each wheel matches its lever arm, jointly driving the vehicle body to rotate smoothly around the axis, avoiding additional translation or yaw caused by improper torque distribution. These technical features enable vehicles to maintain their rotation trajectory and body posture when making U-turns or small-radius turns on low-adhesion surfaces such as water accumulation and ice / snow, thus improving the safety and reliability of the steering process.

[0081] In some instances, torque distribution control is performed on each wheel based on a defined center wheel, including: When the target steering mode is transition mode, perform a semi-locking action on the two center wheels; Based on the positions of the two central wheels, a virtual centerline is determined, where the virtual centerline is the perpendicular bisector of the line connecting the two central wheels. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual centerline, determine the transition distance from each non-center wheel to the virtual centerline; Based on the base torque and transition distance, determine the initial target torque for each non-center wheel; Based on the initial target torque of each non-center wheel, control the output torque of each non-center wheel.

[0082] For example, when the target steering mode is transition mode, it indicates that the current adhesion coefficient of all wheels is within the range between the second preset threshold and the first preset threshold. This means the vehicle is traveling on an asymmetrical road surface, either partially dry after rain or partially wet. In this case, the adhesion capability of each wheel is insufficient to support complete locking in single-center mode, nor has it reached the low adhesion level required to activate dual-center mode. In this situation, a semi-locking action is performed on the two selected center wheels. Unlike the complete lock-up control in single-center and dual-center modes, semi-locking refers to applying a non-complete lock-up control state to the center wheels, keeping them at a small speed rather than completely stationary. This maintains the rotational reference while providing a buffer for possible center switching, avoiding vehicle impact caused by sudden changes in the lock state. Through semi-locking control, the two center wheels in transition mode can provide a certain degree of constraint on vehicle rotation while retaining the flexibility of speed adjustment.

[0083] The specific implementation of the semi-locking action on the two center wheels is as follows: the target semi-locking speed is determined based on the current linear velocity of the target vehicle's rotation and the wheel radius. The current linear velocity of rotation refers to the instantaneous linear velocity of the vehicle body around the center of rotation during steering. This value can be calculated from the vehicle body's yaw rate and rotation radius collected by the inertial navigation module. Dividing the current linear velocity of rotation by the wheel radius yields the theoretical wheel speed that matches the vehicle body's rotation. This is then multiplied by a preset semi-locking coefficient, such as 0.1, to finally determine the target semi-locking speed. This target semi-locking speed characterizes the small rotational speed that the center wheels should maintain in transition mode. This ensures that the center wheels are neither completely stationary nor freely rotating, but rather follow the vehicle body's rotation at a low speed proportional to the vehicle body's rotation. Thus, while maintaining the rotational reference, the small rotational speed buffers the impact that may occur during mode switching or center adjustment, making the vehicle body movement smoother.

[0084] After identifying the two center wheels and applying a semi-locking action, a virtual centerline is determined based on their spatial positions on the vehicle. Specifically, the virtual centerline is the perpendicular bisector of the line connecting the two center wheels, passing through the midpoint of that line and perpendicular to it. The virtual centerline is introduced because, in transition mode, the vehicle body does not rotate around a fixed single point or an axis defined by the two locked wheels, but rather around a dynamic rotational reference defined by the two semi-locked center wheels. This virtual centerline serves as the geometric reference for rotation, used for lever arm calculations during subsequent torque distribution to the non-center wheels, allowing the rotation center to dynamically adjust with changes in the center wheel position, adapting to variations in adhesion conditions under mid-adhesion and asymmetric road conditions.

[0085] Based on the established virtual centerline, torque distribution calculations are performed for the non-centered wheels. First, a base torque is determined based on the vehicle's inherent dynamic parameters and the driver's intent. This base torque is calculated from the vehicle's moment of inertia, target angular acceleration, and wheel radius. The vehicle's moment of inertia is a constant determined during vehicle design, the target angular acceleration is obtained from the driver's steering input, and the wheel radius is a fixed geometric dimension. The base torque reflects the torque required per unit lever arm to drive the vehicle body to rotate at the target angular acceleration. Then, based on the spatial position of the virtual centerline, the vertical distance from each non-centered wheel to the virtual centerline is calculated, and this distance is used as the transition distance. Finally, the initial target torque for each non-centered wheel is determined based on the base torque and the transition distance. Specifically, the base torque is divided by the corresponding transition distance for each non-centered wheel to obtain the torque value that each non-centered wheel should output. Since the vertical distances of each non-centered wheel relative to the virtual centerline are different, the allocated initial target torques also differ. Wheels farther from the virtual centerline receive smaller torques, and vice versa. This ensures that the driving torque output by each wheel can collaboratively drive the vehicle body to rotate smoothly around the virtual centerline, achieving steering control in the transition mode.

[0086] In summary, the embodiments of this application achieve stable steering under medium adhesion and asymmetric road conditions through the control method described above in the transition mode. First, a semi-locking action is performed on the two center wheels, and the target semi-locking speed is determined based on the current rotational linear velocity. This allows the center wheels to maintain a small rotational speed while keeping the rotational reference intact. This avoids the slippage risk that may occur under conditions of moderate adhesion when fully locked, and the preservation of the small rotational speed provides a buffer for center wheel switching or attitude adjustment, reducing the impact of sudden control actions on the vehicle body. Second, a virtual centerline is determined based on the positions of the two semi-locked center wheels, extending the rotational reference from a fixed point or fixed axis to a dynamically adjustable centerline. This allows the rotation center to adaptively adjust according to the real-time selected center wheel position, avoiding trajectory deviation caused by a mismatch between the fixed rotational reference and the current adhesion conditions. Furthermore, the initial target torque is calculated based on the base torque and the vertical distance from each non-centered wheel to the virtual centerline. This ensures that the torque distribution corresponds to the rotational geometry, guaranteeing that the torque output by each non-centered wheel matches its lever arm length. This collectively drives the vehicle body to rotate smoothly around the virtual centerline, avoiding body sway or uneven rotation caused by improper torque distribution. These technical features enable the vehicle to maintain a stable rotational trajectory and controllable body posture under conditions such as semi-dry roads after rain, half-dry roads with half-water accumulation, or asymmetrical road conditions, thus improving the adaptability of steering control.

[0087] In some instances, it also includes: When the angular velocity deviation is greater than the preset deviation threshold, the attitude correction coefficient is determined based on the vehicle body yaw rate and the target yaw rate. The initial target torque is corrected based on the attitude correction coefficient to obtain the corrected target torque; Based on the corrected target torque, the output torque of each non-center wheel is controlled.

[0088] For example, during steering control in transition mode, when the angular velocity deviation exceeds a preset deviation threshold, it indicates a significant deviation between the actual rotation rate of the vehicle body and the rotation rate expected by the driver. Adjustments to the torque output of the non-centered wheels are needed to correct the vehicle's attitude. Specifically, based on the real-time acquired yaw rate and the target yaw rate determined according to the driver's steering input, the absolute value of the difference between the two is calculated. The ratio of this difference to the target yaw rate is then incremented by 1 to obtain the attitude correction coefficient. This attitude correction coefficient is mathematically expressed as k. att =1+|ω r -ω target | / ω target , where ω r Let ω be the yaw rate of the vehicle body. targetThe target yaw rate is denoted as 1. The attitude correction coefficient is a value greater than or equal to 1. The larger the deviation between the actual yaw rate and the target value, the larger the coefficient becomes, resulting in a stronger adjustment effect in torque correction.

[0089] After obtaining the attitude correction coefficient, it is applied to correct the initial target torque of the non-centered wheels. For each non-centered wheel, the initial target torque calculated based on the base torque and transition distance is multiplied by the attitude correction coefficient to obtain the corrected target torque. This multiplication operation allows the torque output to be scaled proportionally according to the magnitude of the attitude deviation; the larger the deviation, the more significant the torque increase, thus generating a stronger corrective torque. The corrected target torque is sent as the final execution command to the motor controllers corresponding to each non-centered wheel, controlling these wheels to output drive torque according to the corrected torque value. By introducing the attitude correction coefficient to dynamically adjust the initial target torque, the torque distribution can respond in real time to changes in the vehicle's attitude, driving the vehicle's rotation rate to converge towards the target value until the angular velocity deviation is reduced to within a preset deviation threshold.

[0090] In summary, this embodiment of the application achieves active control of the vehicle's rotational attitude through the torque correction mechanism based on the attitude correction coefficient. First, the angular velocity deviation is calculated based on the vehicle's yaw rate and the target yaw rate, and this deviation is quantified and incorporated into the determination of the attitude correction coefficient, ensuring that the correction coefficient accurately reflects the degree of deviation between the actual and desired attitudes. Second, the attitude correction coefficient is multiplied by the initial target torque, causing the torque output to scale according to the magnitude of the deviation; the larger the deviation, the more significant the torque increment, thereby generating a corrective torque that suppresses the vehicle's yaw tendency caused by road asymmetry or external disturbances. Third, the corrected torque is directly applied to the non-centered wheels, adjusting the vehicle's rotational angular acceleration by changing the wheel drive torque, gradually bringing the actual yaw rate closer to the target value, thus improving the vehicle's resistance to attitude disturbances in transitional modes. These technical features enable steering control not only to select the centered wheels based on road adhesion conditions but also to dynamically optimize torque distribution based on real-time attitude feedback, enhancing the vehicle's steering stability and trajectory maintenance capabilities under medium adhesion and complex road conditions, and avoiding the risk of steering loss of control due to accumulated attitude deviations.

[0091] In some instances, it also includes: When switching the center wheel in transition mode, based on a preset transition time, smooth speed transition control is performed on the original center wheel before replacement and the new center wheel after replacement, including: Based on the preset transition time, the rotational speed of the original center wheel is linearly increased from the semi-locked target speed to the target speed of the non-center wheel. Based on the preset transition time, the rotational speed of the new center wheel is linearly reduced from the current speed to the semi-lock target speed.

[0092] For example, in the transition mode, when it is determined that the center wheel needs to be replaced based on factors such as vehicle posture deviation or changes in adhesion margin, a smooth transition control of the rotational speed is required to ensure the smoothness of the steering process. The core of this control is to avoid impact or vibration of the vehicle body caused by a sudden change in the state of the center wheel, thereby maintaining the stability of the vehicle posture. Specifically, when the switch occurs, the original center wheel will no longer serve as the center wheel, and its control state needs to change from a semi-locked state to a normal non-center wheel state; while the new center wheel needs to change from a normal non-center wheel state to a semi-locked state. Since the target rotational speeds of the wheels are different in the two states, a direct switch will cause a step change in rotational speed, which will lead to a sudden change in torque. Therefore, a preset transition time is needed to make the rotational speed change continuously to achieve a smooth transition.

[0093] Based on a preset transition time, the rotational speed of the original center wheel linearly increases from the current semi-locked target speed to the target speed of the non-center wheel. The rotational speed of the original center wheel in the semi-locked state is determined by multiplying the ratio of the current vehicle rotational linear velocity to the wheel radius by a preset semi-locking coefficient; this speed is much lower than the normal operating speed of the non-center wheel. During the switching process, the controller uses a preset transition time as a time window, starting from the semi-locked target speed at the beginning, and linearly increases the speed at a constant rate of change until the target speed of the non-center wheel is reached at the end of the transition time. Simultaneously, the rotational speed of the new center wheel linearly decreases from its current speed to the semi-locked target speed. Before the switch, the new center wheel operates as a normal non-center wheel, and its speed is determined by the target vehicle speed and wheel radius. Within the same preset transition time, the controller linearly decreases its speed from the current speed at a constant rate of change until it drops to the semi-locked target speed at the end of the transition time. Through the above bidirectional linear adjustment, the rotational speeds of both wheels change continuously within the transition time, avoiding torque shocks caused by speed jumps. The preset transition time is 50ms.

[0094] In summary, this embodiment of the application provides a time window for speed changes by introducing a preset transition period, giving the speed adjustment process a controllable time scale and avoiding abrupt changes in control commands caused by instantaneous switching. By linearly increasing the speed of the original center wheel from the semi-lock target speed to the target speed of the non-center wheel, the rotational speed of this wheel gradually synchronizes with the overall vehicle movement during the transition from the center role to the normal wheel role, preventing dragging or stagnation effects caused by sudden acceleration and maintaining the continuity of vehicle rotation. Similarly, by linearly decreasing the speed of the new center wheel from the current speed to the semi-lock target speed, the rotational speed of this wheel gradually matches the semi-lock requirement during the transition from the normal wheel role to the center role, avoiding slippage or impact caused by sudden deceleration. This linear change strategy ensures that the speeds of all four wheels are continuously changing during the switching process, maintaining a stable relative motion between each wheel and the ground, thereby suppressing vehicle vibration and improving ride comfort and trajectory accuracy during steering.

[0095] It should be noted that in the transition mode, when the center adjustment is triggered due to an angular velocity deviation exceeding a preset deviation threshold, only the second center wheel is switched, while the first center wheel with the largest adhesion margin remains unchanged. Specifically, when it is determined that the deviation between the actual yaw rate of the vehicle body and the target yaw rate exceeds a preset range, the controller keeps the wheel with the largest current adhesion margin as the first center wheel, which always serves as the primary reference for rotation due to its optimal adhesion capability. Based on this, the original second center wheel is replaced with a wheel located on the same side of the vehicle body as the first center wheel. This same-side wheel replaces the original diagonally positioned wheel as the new second center wheel. During the switching process, only the two second center wheels involved in the replacement undergo smooth speed transition control. That is, the speed of the original second center wheel linearly increases from the semi-locked target speed to the target speed of the non-center wheel, while the speed of the new second center wheel linearly decreases from its current speed to the semi-locked target speed, while the first center wheel remains in a semi-locked state. By switching only the second center while keeping the first center unchanged, the stability of the rotation reference is maintained by utilizing the wheel with the strongest adhesion, and the spatial position of the virtual centerline is changed by adjusting the second center to generate a corrective torque. This achieves the correction of the vehicle body attitude while ensuring that the rotation reference does not become unstable.

[0096] The technical solution of this application will be further described below with reference to specific embodiments.

[0097] Example 1: Dry asphalt road surface (single-center model) This embodiment of the application corresponds to a scenario where a vehicle performs a U-turn on a dry asphalt road surface. During the perception phase, wheel speed sensors collect the rotational speed of each wheel in real time. Calculations show that the speed difference between adjacent wheels is less than 2 rpm, indicating that there is no relative slippage trend among the wheels. The inertial navigation module detects that the vehicle's yaw rate is zero, confirming that the vehicle is currently stationary. The high-definition visual camera performs semantic segmentation and recognition on the road surface image ahead, determines that the road surface type is dry asphalt, and outputs a corresponding road surface adhesion confidence score of 0.95. This high confidence score indicates that the visual recognition result is reliable.

[0098] During the adhesion coefficient estimation stage, the controller, based on the acquired speed difference, vehicle lateral acceleration, geometric correction coefficients of the line connecting each wheel to the center of gravity, and visual correction coefficients determined by the road adhesion confidence level, performs multi-source data fusion processing using an extended Kalman filter algorithm. The calculated current adhesion coefficients for the four wheels are 0.85, 0.83, 0.87, and 0.84, respectively. It is determined that the current adhesion coefficients of all wheels are greater than or equal to the first preset threshold of 0.6, thus meeting the criteria for a high-adhesion road surface.

[0099] During the decision-making phase, the controller determines the single-center mode as the target steering mode and, based on the comparison of adhesion coefficients, identifies the right rear wheel with the highest current adhesion coefficient among the four wheels as the center wheel. During the execution phase, the controller sends a zero-speed command to the motor controller of the right rear wheel, causing its output torque to be zero and maintaining a fully locked state, thus forming a fixed rotation center. Simultaneously, based on the vehicle's rotational inertia, target angular acceleration, and wheel radius, the controller calculates the base torque. Then, according to the single-center distances from the left front wheel, left rear wheel, and right front wheel to the right rear wheel, the controller distributes the target torque to each non-center wheel in an inverse relationship between the base torque and the single-center distance. The calculated target torques are: left front wheel +300 Nm, left rear wheel -280 Nm, and right front wheel +290 Nm. Each non-center wheel outputs driving torque according to the aforementioned target torque, jointly driving the vehicle body to rotate smoothly around the fully locked right rear wheel, ultimately achieving a U-turn with a radius of only 1.2m, fully demonstrating the efficient and precise steering control capability of the single-center mode on high-adhesion road surfaces.

[0100] Example 2: Mixed ice and snow road surface (dual-center mode) like Figure 2As shown in the figure, this embodiment of the application corresponds to an asymmetric adhesion scenario where the vehicle is traveling on an icy or snowy road surface on the left and a dry road surface on the right. During the perception phase, the wheel speed sensor of the left front wheel detected a sudden increase in its rotational speed difference to 15 rpm, indicating that the wheel has a significant relative slipping trend; the inertial navigation module detected that the lateral acceleration of the vehicle body reached 1.2 m / s², reflecting that the vehicle is subjected to significant lateral inertial force; the high-definition vision camera identified the road surface image ahead, determined that the left area was an icy or snowy road surface and the right area was a dry road surface, and output a low adhesion confidence score of 0.92 for the area where the left front wheel is located.

[0101] During the adhesion coefficient estimation stage, the controller integrates the aforementioned multi-source information and estimates the current adhesion coefficient of each wheel using an extended Kalman filter algorithm. The calculated adhesion coefficient of the left front wheel is 0.25, which is less than the second preset threshold of 0.3, thus confirming it as a low-adhesion wheel. The adhesion coefficients of the other three wheels are 0.75, 0.72, and 0.78, respectively, all within the high-adhesion range. During the decision-making stage, due to the presence of low-adhesion wheels, the controller determines the dual-center mode as the target steering mode and selects the center based on the position of the low-adhesion wheels: the right front wheel and the left rear wheel, which are not on the same diagonal as the low-adhesion left front wheel, are identified as the two center wheels, thereby effectively avoiding the weak adhesion area.

[0102] During execution, the controller sends zero-speed commands to the motor controllers of the right front wheel and left rear wheel, completely locking these two diagonally positioned wheels. The line connecting them forms a virtual steering axis. Based on the vehicle's moment of inertia, target angular acceleration, and wheel radius, the base torque is calculated. The vertical distances from the left front wheel and right rear wheel to this virtual steering axis are then determined as lever arms. Torque is distributed according to the inverse relationship between the base torque and the lever arm, and a positive or negative sign is assigned based on the direction of rotation. The target torque for the left front wheel is calculated to be -150 Nm to cause it to rotate in the opposite direction, and the target torque for the right rear wheel is calculated to be +160 Nm to cause it to rotate in the forward direction. The opposing torques output by the two non-central wheels jointly drive the vehicle to rotate smoothly around the virtual steering axis formed by the right front wheel and left rear wheel. The measured trajectory deviation is less than 0.3 m, achieving stable steering control under low-adhesion and asymmetrical road conditions, avoiding the risk of steering loss due to slippage at a single center point.

[0103] Example 3: Semi-dry asymmetric road surface after rain (transition mode) This embodiment of the application corresponds to a mid-attachment asymmetric scenario where the left side of the vehicle is a semi-dry road surface after rain and the right side is a dry road surface. During the perception phase, the wheel speed sensors detect wheel speed differences of 5 rpm, 3 rpm, 4 rpm, and 2 rpm, respectively, reflecting different degrees of wheel slippage. The inertial navigation module detects the vehicle's yaw rate as 12 deg / s, while the target yaw rate determined according to the driver's steering operation command is 10 deg / s, with the actual value deviating from the target value by 20%, indicating a significant shift in vehicle attitude. The high-definition visual camera identifies the left side as a semi-dry road surface after rain and the right side as a dry road surface, outputting a mid-attachment confidence level of 0.90.

[0104] During the adhesion coefficient estimation stage, the controller integrates the above information to estimate the current adhesion coefficients of the four wheels as 0.42, 0.55, 0.48 and 0.58, respectively. It is determined that all adhesion coefficients are within the range between the second preset threshold of 0.3 and the first preset threshold of 0.6, thus meeting the activation conditions of the transition mode.

[0105] During the decision-making phase, the adhesion margin of each wheel is first calculated. This is achieved by subtracting the minimum adhesion coefficient of 0.42 from the current adhesion coefficient of each wheel, resulting in margin values ​​of 0, 0.13, 0.06, and 0.16, respectively. The right front wheel with the largest adhesion margin is designated as the first center wheel, and the left rear wheel is designated as the second center wheel based on the diagonal principle. Since the angular velocity deviation of 20% exceeds the preset deviation threshold of 5%, the center wheel combination is adjusted to correct the vehicle's attitude. The controller keeps the right front wheel with the largest adhesion margin unchanged and switches the second center wheel from the original diagonal position of the left rear wheel to the right rear wheel on the same side as the right front wheel. Simultaneously, the attitude correction coefficient k is calculated based on the actual yaw rate and the target yaw rate. att It is 1.2, used for torque correction adjustment.

[0106] During the execution phase, the controller performs a semi-locking action on the two central wheels, the right front wheel and the right rear wheel. Based on the ratio of the current vehicle rotational linear velocity of 0.8 m / s to the wheel radius of 0.35 m, multiplied by a preset semi-locking coefficient of 0.1, the target semi-locking speed is determined to be 0.23 r / s, and the two central wheels are controlled to rotate at this small speed. Simultaneously, a virtual centerline is determined based on the positions of the two central wheels, and the transition distance from the left front wheel to the virtual centerline is calculated to be 1.2 m, and the transition distance from the left rear wheel to the virtual centerline is calculated to be 1.0 m. Combining the base torque and attitude correction coefficient, the target torque for the left front wheel correction is 3571 Nm, expressed by the formula T1=I z ·α / (r·L1)×k att= 2500kg·m²×0.5rad / s² / (0.35m×1.2m)×1.2≈3571Nm, the target torque for the left rear wheel correction is 4286 Nm, and the calculation formula is T2=2500×0.5 / (0.35×1.0)×1.2≈4286Nm; During the switching of the center wheel, the controller adopts a speed smooth transition strategy. Within the preset transition time of 50ms, the speed of the left rear wheel of the original second center wheel is linearly increased from the semi-locked target speed of 0.23r / s to the target speed of the non-center wheel of 1.8r / s, so that it smoothly transforms into an ordinary non-center wheel; At the same time, the speed of the right rear wheel of the new second center wheel is linearly reduced from the current speed of 1.5r / s before the switch to the semi-locked target speed of 0.23r / s, so that it smoothly transforms into a semi-locked center wheel. After the switch was completed, the vehicle's yaw rate quickly returned to 10.5 deg / s, and the deviation from the target value was reduced to within 5%. The vehicle's attitude remained stable, and the entire switch process was free of shocks and vibrations, verifying the effectiveness of the dynamic adjustment of the center and smooth transition control strategy in the transition mode.

[0107] Please see Figure 3 The diagram below illustrates the structure of a steering control device for a distributed drive vehicle, as provided in an embodiment of this application. The device includes: Data information acquisition unit 21 is used to acquire the motion state information of the target vehicle and the road environment information; The adhesion coefficient calculation unit 22 is used to determine the current adhesion coefficient of each wheel based on motion state information and road environment information; The center wheel determination unit 23 is used to determine the target steering mode of the target vehicle and the center wheel corresponding to the target steering mode from multiple preset steering modes based on the current adhesion coefficient of each wheel. The preset steering modes include single center mode, double center mode or transition mode. The torque steering control unit 24 is used to control the torque distribution to each wheel based on a determined center wheel in order to achieve steering control of the target vehicle.

[0108] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it; although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features.

[0109] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications that fall outside the scope of this specification.

[0110] Obviously, those skilled in the art can make various modifications to this specification without departing from its spirit and scope. Therefore, this specification also intends to include any modifications that fall within the scope of the claims and their equivalents.

Claims

1. A steering control method for a distributed drive vehicle, characterized in that, include: Acquire the target vehicle's motion status information and road environment information; Based on the motion state information and the road surface environment information, the current adhesion coefficient of each wheel is determined; Based on the current adhesion coefficient of each wheel, the target steering mode of the target vehicle and the center wheel corresponding to the target steering mode are determined from multiple preset steering modes, wherein the preset steering modes include single center mode, double center mode or transition mode. Based on the determined center wheel, torque distribution control is performed on each wheel to achieve steering control of the target vehicle.

2. The method according to claim 1, characterized in that, The motion state information includes the speed difference between adjacent wheels and the lateral acceleration of the vehicle body, and the road environment information includes the road surface type and the corresponding road surface adhesion confidence level.

3. The method according to claim 2, characterized in that, The determination of the current adhesion coefficient for each wheel based on the motion state information and the road surface environment information includes: The geometric correction factor for each wheel is determined based on the angle between each wheel and the line connecting the center of gravity of the target vehicle. Based on the road surface adhesion confidence level, the visual correction coefficient for each wheel is determined; Based on the speed difference, the geometric correction coefficient, the visual correction coefficient, and the preset calibration coefficient, the current adhesion coefficient of each wheel is determined using the extended Kalman filter algorithm.

4. The method according to claim 1, characterized in that, The determination of the target steering mode of the target vehicle from multiple preset steering modes based on the current adhesion coefficient of each wheel includes: When the current adhesion coefficient of each wheel is greater than or equal to the first preset threshold, the single-center mode is determined as the target steering mode. When the current adhesion coefficient of any wheel is less than the second preset threshold, the dual-center mode is determined as the target steering mode; When the current adhesion coefficient of each wheel is greater than or equal to the second preset threshold and less than the first preset threshold, the transition mode is determined as the target steering mode.

5. The method according to claim 4, characterized in that, Based on the current adhesion coefficient of each wheel, the center wheel corresponding to the target steering mode is determined, including: When the target steering mode is the single-center mode, the wheel with the highest current adhesion coefficient is determined as the center wheel; When the target steering mode is the dual-center mode, the wheel with the current adhesion coefficient less than the second preset threshold is identified as the low-adhesion wheel, and the two wheels that are not on the same diagonal as the low-adhesion wheel are identified as the center wheel. When the target steering mode is the transition mode, two wheels are determined from each wheel as the center wheel based on the current adhesion coefficient of each wheel.

6. The method according to claim 5, characterized in that, The step of determining two wheels as the center wheels based on the current adhesion coefficient of each wheel includes: Based on the current adhesion coefficient of each wheel, the adhesion margin of each wheel is calculated, wherein the adhesion margin is the difference between the current adhesion coefficient of the wheel and the minimum current adhesion coefficient among all wheels. The wheel with the largest attachment margin is designated as the first center wheel, and the wheel located on the same diagonal as the first center wheel is designated as the second center wheel, so that the first center wheel and the second center wheel are used as the center wheels.

7. The method according to claim 6, characterized in that, The motion state information includes the vehicle body yaw rate, and also includes: Obtain the yaw rate of the target vehicle; Determine the target yaw rate based on the driver's steering instructions; The angular velocity deviation is determined based on the vehicle body yaw rate and the target yaw rate; When the angular velocity deviation is greater than the preset deviation threshold, the first center wheel with the largest attachment margin is kept unchanged, and the wheel located on the same side of the vehicle body as the first center wheel is determined as the second center wheel.

8. The method according to claim 5, characterized in that, The torque distribution control of each wheel based on the determined center wheel includes: When the target steering mode is the single-center mode, the center wheel is locked. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the position of the central wheel, determine the single center distance from each non-central wheel to the central wheel; Based on the base torque and the single-center distance, determine the single-center target torque for each non-center wheel; Based on the single-center target torque of each non-center wheel, control the output torque of each non-center wheel.

9. The method according to claim 5, characterized in that, The torque distribution control of each wheel based on the determined center wheel includes: When the target steering mode is the dual-center mode, the two center wheels are locked. The virtual steering axis is determined based on the positions of the two central wheels. The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual steering axis, determine the double center distance from each non-center wheel to the virtual steering axis; Based on the base torque and the distance between the two centers, determine the target torque between the two centers for each non-center wheel; Based on the dual-center target torque of each non-center wheel, the output torque of each non-center wheel is controlled.

10. The method according to claim 7, characterized in that, The torque distribution control of each wheel based on the determined center wheel includes: When the target steering mode is the transition mode, a semi-locking action is performed on the two central wheels; Based on the positions of the two central wheels, a virtual centerline is determined, wherein the virtual centerline is the perpendicular bisector of the line connecting the two central wheels; The base torque is determined based on the vehicle body's rotational inertia, target angular acceleration, and wheel radius; Based on the virtual centerline, determine the transition distance from each non-center wheel to the virtual centerline; Based on the base torque and the transition distance, the initial target torque for each non-centered wheel is determined; Based on the initial target torque of each non-center wheel, control the output torque of each non-center wheel.

11. The method according to claim 10, characterized in that, The partial locking action on the two central wheels includes: Based on the current rotational linear velocity of the target vehicle and the wheel radius, a semi-lock target rotational speed is determined, and the two central wheels are controlled to rotate at the semi-lock target rotational speed.

12. The method according to claim 10, characterized in that, Also includes: When the angular velocity deviation is greater than a preset deviation threshold, an attitude correction coefficient is determined based on the vehicle body yaw rate and the target yaw rate. The initial target torque is corrected based on the attitude correction coefficient to obtain the corrected target torque; Based on the corrected target torque, the output torque of each non-center wheel is controlled.

13. The method according to claim 11, characterized in that, Also includes: When switching the center wheel in the transition mode, the rotational speed of the original center wheel before replacement and the new center wheel after replacement are smoothly controlled based on a preset transition time.

14. The method according to claim 13, characterized in that, The method of smoothly transitioning the rotational speed of the original center wheel before replacement and the new center wheel after replacement, based on a preset transition time, includes: Based on the preset transition time, the rotational speed of the original center wheel is linearly increased from the semi-locked target rotational speed to the target rotational speed of the non-center wheel; Based on the preset transition time, the rotational speed of the new center wheel is linearly reduced from the current rotational speed to the semi-lock target rotational speed.

Images