A method and system for dynamic self-learning of a position in a steer-by-wire system

Abstract

The present application relates to the technical field of steering system, and particularly relates to a method and system for dynamic self-learning of neutral position in a steer-by-wire system.The specific method flow is as follows: (1) the lower steering execution system obtains various information of the whole vehicle; (2) the lower steering execution system starts the neutral self-learning function through the straight driving detection function module; (3) the neutral self-learning function module of the lower steering execution system reads the neutral learning value from the data storage function module and provides the value to the angle calculation module of the lower steering execution module; (4) the neutral self-learning function module and the RPS motor rotor position acquisition module calculate the angle of the current lower steering execution system for the use of the angle-related function module.Compared with the prior art, the steer-by-wire system generally considers the re-calibration of the steering angle neutral position under the straight line working condition, so that the vehicle provides more accurate angle signals, and the accurate control of the steer-by-wire related function is facilitated.

Description

Technical Field

[0001] This invention relates to the field of steering system technology, specifically to a method and system for mid-level dynamic self-learning in steer-by-wire systems. Background Technology

[0002] With the development of the automotive industry, people have increasingly higher demands for driving performance and convenience. To meet these needs, cars are becoming more intelligent and functionally diverse, incorporating features such as driver assistance, lane keeping assist, and valet parking. Simultaneously, to meet the overall vehicle requirements, electric power steering has been developed, making the steering system more intelligent. In electric power steering systems, both the assist torque and the active return torque are set based on the angle values ​​of the steering angle sensor and vehicle speed. Therefore, the consistency between the angle of the steering angle sensor and the actual steering wheel angle is crucial. The key to calculating the steering angle is determining the zero point of the steering angle (i.e., the center of the steering angle).

[0003] The mid-angle dynamic self-learning in the steer-by-wire system refers to automatically determining the mid-angle value of the steering angle during straight-line driving by calibrating the angle sensor value when the steering gear is in the middle position. This provides a more accurate angle basis for triggering and controlling the angle-related functions of the steer-by-wire EPS.

[0004] The commonly used solutions for steering wheel angle center position calibration in existing technologies are as follows: Currently, center position angle learning is achieved through manual calibration. The steering wheel is manually placed in the center position, and the angle signal value of the angle sensor at this position is recorded in the electric power steering system's storage space. Subsequent steering wheel angle values ​​are calculated based on this value. However, this manual calibration method has the following disadvantages when applied to steer-by-wire: First, it requires four-wheel alignment for accurate calibration, which places certain requirements on the calibration environment. Second, recalibration is required after replacing corresponding parts, increasing labor and maintenance costs. Third, it requires high precision in the components; otherwise, it can cause left-right asymmetry, affecting steering feel. In extreme cases, an incorrect steering wheel zero position can increase driving risks and cause loss of life and property. Fourth, assembly errors in various vehicle systems and wear during vehicle use can cause the center position provided by the steering wheel angle sensor to be inconsistent with the center position during driving, resulting in poor vehicle stability. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a method and system for mid-range dynamic self-learning in a steer-by-wire system. This steer-by-wire system typically considers recalibrating the mid-range of the steering angle under straight-line conditions, thereby enabling the vehicle to provide more accurate angle signals and facilitating precise control of steer-by-wire related functions.

[0006] To achieve the above objectives, a method and system for mid-level dynamic self-learning in a steer-by-wire system are designed, characterized by the following specific method flow:

[0007] (1) The lower steering execution system obtains information on the vehicle speed, wheel speed, yaw rate and lateral acceleration signals of the whole vehicle through the whole vehicle CAN, and obtains the steering wheel torque and steering wheel angle information of the upper steering road feel system through the private CAN, which are used for the determination of the straight detection function module of the lower steering execution system.

[0008] (2) The steering system activates the center position self-learning function through the straight-line detection function module: When the vehicle is at high speed, the TAS angle acquisition module frequently acquires the TAS angle signal, and the center position self-learning function module learns the center position angle through the algorithm; when the vehicle is at low speed, the learned center position angle is stored in the data storage function module.

[0009] (3) The median self-learning function module of the down steering execution system will read the median learning value from the data storage function module, compensate the current median learning value by low-pass filtering, and provide it to the angle calculation module of the down steering execution module.

[0010] (4) The mid-position self-learning function module and the RPS motor rotor position acquisition module calculate the current angle of the steering execution system for use by angle-related function modules.

[0011] To enable the median self-learning module to perform median learning in a relatively stable straight-line environment, straight-line detection of the vehicle's driving conditions is required. The method for straight-line detection is as follows:

[0012] S1 detects the upper steering wheel hand torque, vehicle speed, wheel speed and speed difference of the four wheels, speed change gradient, vehicle yaw rate, lower steering shaft speed, and upper steering wheel angle in the steer-by-wire system; and the current lower steering TAS sensor value.

[0013] S2, determine whether the following conditions are all satisfied:

[0014] 1) The difference between the current TAS sensor value and the previously learned angle value is less than the first preset threshold;

[0015] 2) The steering wheel torque is less than the second preset threshold;

[0016] 3) The vehicle speed exceeds the preset third threshold;

[0017] 4) The difference between the wheel speed of the left front wheel and the wheel speed of the left rear wheel is less than the fourth preset threshold;

[0018] 5) The difference between the wheel speed of the right front wheel and the wheel speed of the right rear wheel is less than the fifth preset threshold;

[0019] 6) The difference between the sum of the wheel speeds of two diagonal wheels is less than a sixth preset threshold;

[0020] 7)} The gradient of the change in the entire vehicle speed is less than a seventh preset threshold;

[0021] 8) The rotational speed of the lower steering shaft under steer-by-wire is less than an eighth preset threshold;

[0022] 9) The steering wheel angle of the upper steering under steer-by-wire is less than a ninth preset threshold;

[0023] 10) The yaw rate of the entire vehicle is less than a tenth preset threshold;

[0024] S3: If all the conditions in step S2 are satisfied and after a stable time range, the neutral position self-learning functional module in the lower steering execution system is triggered, the angle value of the TAS angle sensor is the learned neutral position value, and the learned neutral position value is recorded through an algorithm.

[0025] The mentioned frequency sampling strategy is specifically as follows:

[0026] (1) In order to learn the neutral position value relatively more accurately, during the process when the neutral position self-learning is enabled, the neutral position value wConvTASAngle1 is learned at the moment of enabling, then wConvTASAngle2 is learned after an interval of time t, and learning is continued continuously when the straight traveling detection is satisfied until the preset nth angle value wConvTASAngle n, the preset sampling number n can be determined according to the performance of learning the neutral position value, and n is 100-200;

[0027] (2) During high-speed straight traveling, in the self-learning process, the continuous learning time is t1, t2, t3...tx respectively, until n preset values are learned, where n is 100-200;

[0028] A. In the t1 (Δt<t1<2Δt) period: when the neutral position is enabled, start learning the neutral position angle wConvTASAngle1; then learn wConvTASAngle2 after an interval of time Δt;

[0029] B. In the t2 (5Δt<t2<6Δt) period: when the neutral position is enabled again, as in step 1, continuously learn the neutral position angles wConvTASAngle3, wConvTASAngle4, wConvTASAngle5, wConvTASAngle6, wConvTASAngle7, wConvTASAngle8;

[0030] C. In the t3 (0<t3<Δt) period: when the neutral position is enabled again, as in step 2, continuously learn the neutral position angle wConvTASAngle9;

[0031] D. Until the final tx (3Δt<tx) segment: when the neutral position is re-enabled, as in step 3, continuously learn the neutral position angles wConvTASAngle(n-2), wConvTASAngle(n-1), and wConvTASAngle n.

[0032] The aforementioned compensation strategy is specifically as follows: the neutral position value stored in the data storage function module at low speed will update the currently used neutral position value in time after storage. To avoid instantaneous change of the angle caused by instantaneously updating the neutral position value, the original current neutral position learning value is compensated through a low-pass filter to gradually compensate to the actual steering wheel neutral position value, that is, the entire process of neutral position angle self-learning is completed.

[0033] The aforementioned frequency sampling refers to continuous sampling at predetermined time intervals of 50-100ms when the straight-line driving detection condition is satisfied, until the required number of samples is reached.

[0034] A system for dynamic neutral position self-learning of a steer-by-wire system, characterized in that the system comprises an upper steering road feel system, a vehicle signal processing system, and a lower steering execution system, the upper steering road feel system and the lower steering execution system are connected via a private CAN, and the upper steering road feel system and the lower steering execution system are connected to the vehicle signal processing system via a vehicle CAN;

[0035] The aforementioned upper steering road feel system: includes a steering wheel torque acquisition module and a steering wheel angle acquisition module, configured to calculate the torque applied on the steering wheel and the angle of the upper steering road feel system, and transmit them to the lower steering execution system via the private CAN for the straight-line driving detection function module of the lower steering execution system;

[0036] The aforementioned vehicle signal processing system: includes a vehicle speed acquisition module, a wheel speed acquisition module, a yaw rate acquisition module, and a lateral acceleration acquisition module, configured to calculate the vehicle speed signal, wheel speed signal, yaw rate signal, and lateral acceleration signal, and transmit them to the lower steering execution system via the vehicle CAN for the straight-line driving detection function module of the lower steering execution system;

[0037] The aforementioned downward steering execution system includes a straight-line detection module, a center-position self-learning module, a TAS angle acquisition module, a fault handling module, a data storage module, a downward steering execution module angle calculation module, and an RPS motor rotor position acquisition module. The straight-line detection module determines the vehicle's straight-line conditions to activate the center-position self-learning function, involving vehicle speed signals, wheel speed signals, yaw rate signals, lateral acceleration signals, steering wheel torque signals, and steering wheel angle signals. The TAS angle acquisition module provides the learnable center-position angle. The center-position self-learning module learns the center-position angle at high speed, stores the learned value at low speed, and provides the learned value to the downward steering execution module angle calculation module. The RPS motor rotor position acquisition module provides the motor rotor angle position to the downward steering execution module angle calculation module and calculates the angle based on the center-position value. The data storage module stores the center-position self-learning value, ensuring the value remains valid even after the EPS is powered on or off, thus meeting the non-volatility requirement of the center-position self-learning value. The fault handling module stores the fault and executes relevant safety measures when unexpected faults are encountered during learning.

[0038] Compared with the prior art, the present invention provides a method and system for mid-range dynamic self-learning in a steer-by-wire system. This steer-by-wire system typically considers recalibrating the mid-range of the steering angle under straight-line conditions, thereby enabling the vehicle to provide more accurate angle signals and facilitating precise control of steer-by-wire related functions. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the steer-by-wire system.

[0040] Figure 2 This is a schematic diagram of the self-learning function in the steer-by-wire system of the present invention.

[0041] Figure 3 This is a flowchart of the mid-position self-learning process of the steer-by-wire system of the present invention.

[0042] Figure 4 This is a schematic diagram of the self-learning process of the mid-angle in this invention.

[0043] Figure 5 This is a schematic diagram of the high-speed self-learning process in this invention.

[0044] Figure 6 This is a schematic diagram of the low-speed self-learning storage process in this invention.

[0045] Figure 7 This is a schematic diagram of the self-learning low-pass filter compensation in this invention. Detailed Implementation

[0046] The present invention will now be further described with reference to the accompanying drawings.

[0047] like Figure 1 As shown, the steer-by-wire system mainly consists of an upper steering feel system and a lower steering actuation system. The upper steering feel system mainly includes a steering wheel 1, a TAS torque and angle sensor 2, an upper steering column assembly 3, an upper steering feel motor and ECU 10, and a turbine assembly 11. The lower steering actuation module mainly includes a tie rod assembly 4, a rack assembly 5, a turbine assembly 6, a TAS angle sensor 7, a motor position sensor 8, a lower steering actuation motor and ECU (Electric Control Unit) 9. The upper steering feel system and the lower steering actuation system communicate with each other via a private CAN bus, and both systems communicate with the vehicle via the CAN bus. The steering wheel 1 is used for driver control, and the input end of the steering column assembly 3 is connected to the steering wheel 1. The TAS torque sensor 2 of the upper steering road feel module detects the driver's steering torque and angle. The road feel ECU 10 is used to collect and process the steering torque and angle, and transmits them to the lower steering actuator ECU 9 via private CAN for straight-line detection conditions during the self-learning process. The road feel motor 12 is mounted on the upper steering column assembly and provides road feel torque to the driver through a reduction mechanism.

[0048] The lower power steering motor 11 is mounted on the steering gear and provides power assist torque to the driver through the reduction mechanism. The torque of the power steering motor is applied to the rack assembly 5 and the tie rod assembly 4 through the turbine assembly to achieve power steering. The TAS angle sensor 7 is used to detect the angle value of the steering shaft to determine the center value of the lower steering angle. The motor position sensor 8 is used to detect the rotor position angle of the power steering motor and feeds back the detected angle of the power steering motor to the ECU 9 for calculating the lower steering angle.

[0049] like Figure 2 As shown, the median dynamic self-learning function of this invention includes three functional systems:

[0050] (1) Upper steering road feel system: including steering wheel torque acquisition module and steering wheel angle acquisition module, used to calculate the torque applied to the steering wheel and the angle of the upper steering road feel system, and transmit it to the lower steering execution system through private CAN, for the straight-line detection function module of the lower steering execution system.

[0051] (2) Vehicle signal processing system: including vehicle speed acquisition module, wheel speed acquisition module, yaw rate acquisition module and lateral acceleration acquisition module, used to calculate vehicle speed signal, wheel speed signal, yaw rate signal and lateral acceleration signal, and transmit them to the lower steering execution system through the vehicle CAN, for the straight-line detection function module of the lower steering execution system.

[0052] (3) Downward Steering Execution System: Includes a straight-line detection module, a center position self-learning module, a TAS angle acquisition module, a fault handling module, a data storage module, a downward steering execution module angle calculation module, and an RPS motor rotor position acquisition module. The straight-line detection module determines the straight-line conditions of the vehicle to activate the center position self-learning function, involving vehicle speed signal, wheel speed signal, yaw rate signal, lateral acceleration signal, steering wheel torque signal, and steering wheel angle signal. The TAS angle acquisition module provides the center position learnable angle. The center position self-learning module is used for high-speed self-learning of the center position angle, low-speed storage of the center position learning value, and providing the learning value to the downward steering execution module angle calculation module. The RPS motor rotor position acquisition module provides the motor rotor angle position to the downward steering execution module angle calculation module and calculates the angle based on the center position. The data storage module stores the center position self-learning value and can retain the learning value after the EPS is powered on and off, satisfying the non-volatility of the center position self-learning value. The fault handling module stores the fault and executes relevant safety measures when unexpected faults are encountered during learning.

[0053] like Figure 3 As shown in the flowchart, the method and flowchart of the present invention include the following four steps:

[0054] (1) The lower steering execution system obtains information on the vehicle speed, wheel speed, yaw rate and lateral acceleration signals of the whole vehicle through the whole vehicle CAN, and obtains the steering wheel torque and steering wheel angle information of the upper steering road feel system through the private CAN, which are used for the determination of the straight detection function module of the lower steering execution system.

[0055] (2) The steering system activates the center position self-learning function through the straight-line detection function module: When the vehicle is at high speed, the TAS angle acquisition module frequently samples (under the condition that the straight-line detection is met, it continuously samples at a predetermined time interval, the time interval is generally 50-100ms, until the preset number of samples is met) the TAS angle signal, and the center position self-learning function module learns the center position angle through an algorithm; when the vehicle is at low speed, the learned center position angle is stored in the data storage function module.

[0056] (3) The median self-learning function module of the down steering execution system will read the median learning value from the data storage function module, compensate the current median learning value by low-pass filtering, and provide it to the angle calculation module of the down steering execution module.

[0057] (4) The mid-position self-learning function module and the RPS motor rotor position acquisition module calculate the current angle of the downward steering execution system for use by angle-related function modules, such as the up and down following control function and the end protection function.

[0058] To enable the median self-learning module to perform median learning in a relatively stable straight-line environment, straight-line detection of the vehicle's driving conditions is required. The method for straight-line detection is as follows:

[0059] S1 detects the upper steering wheel hand torque, vehicle speed, wheel speed and speed difference of the four wheels, speed change gradient, vehicle yaw rate, lower steering shaft speed, and upper steering wheel angle in the steer-by-wire system; and the current lower steering TAS sensor value.

[0060] S2, determine whether the following conditions are all satisfied:

[0061] 1) The difference between the current TAS sensor value and the previously learned angle value is less than the first preset threshold;

[0062] 2) The steering wheel torque is less than the second preset threshold;

[0063] 3) The vehicle speed exceeds the preset third threshold;

[0064] 4) The difference between the wheel speed of the left front wheel and the wheel speed of the left rear wheel is less than the fourth preset threshold;

[0065] 5) The difference between the wheel speed of the right front wheel and the wheel speed of the right rear wheel is less than the fifth preset threshold;

[0066] 6) The difference between the sum of the speeds of the two diagonal wheels is less than the sixth preset threshold;

[0067] 7) The gradient of the vehicle speed change is less than the seventh threshold.

[0068] 8) The steering shaft speed under drive-by-wire steering is less than the eighth preset threshold;

[0069] 9) The steering wheel angle of the drive-by-wire steering system is less than the ninth preset threshold.

[0070] 10) The yaw rate of the entire vehicle is less than the tenth preset threshold;

[0071] S3. If all the conditions in step S2 are met and within a stable time range, the median self-learning function module in the downward steering execution system is triggered. The angle value of the TAS angle sensor is the learned median value, and the learned median value is recorded by the algorithm.

[0072] When the vehicle is at low speed, the median value learned at high speed is immediately stored in the data storage module. On the one hand, this realizes the median self-learning function of high acquisition and low storage, avoiding instantaneous changes in the median during high-speed driving, which could lead to loss of angle control. On the other hand, it can provide the learned median value in real time and accurately, laying a solid foundation for the angle control of steer-by-wire.

[0073] The steering wheel rotation angle θ and the actual steering wheel angle are compensated through a low-pass filter until the actual steering wheel rotation angle is consistent with the target angle, that is, the neutral angle self-learning process of the rotation angle sensor is completed, avoiding unexpected fluctuations caused by instantaneous changes in the calculated angle.

[0074] As shown in Figure 4 , Figure 5 , Figure 6 , the frequency acquisition strategy for neutral dynamic self-learning of the present invention is specifically as follows:

[0075] When the vehicle is traveling straight at high speed, the neutral self-learning is started once the straight travel detection condition is satisfied.

[0076] During the neutral self-learning process, the learned TAS angle will slightly fluctuate up and down due to mechanical vibration.

[0077] (1) In order to learn the neutral value relatively more accurately, during the process when the neutral self-learning is enabled, the neutral value wConvTASAngle1 is learned at the moment of enabling, then wConvTASAngle2 is learned after an interval t, and continuous learning is performed when the straight travel detection is satisfied until the preset nth angle value wConvTASAngle n is reached. The preset acquisition number n can be determined according to the performance of learning the neutral value, and n is generally 100-200;

[0078] (2) During high-speed straight travel, in the self-learning process, the continuous learning time is respectively t1, t2, t3...tn until n preset values are learned, and n is generally 100-200;

[0079] A. In the t1 (Δt<t1<2Δt) period: when the neutral position is enabled, start to learn the neutral angle wConvTASAngle1; then learn wConvTASAngle2 after an interval Δt;

[0080] B. In the t2 (5Δt<t2<6Δt) period: when the neutral position is enabled again, as in step 1, continuously learn the neutral angles wConvTASAngle3, wConvTASAngle4, wConvTASAngle5, wConvTASAngle6, wConvTASAngle7, wConvTASAngle8;

[0081] C. In the t3 (0<t3<Δt) period: when the neutral position is enabled again, as in step 2, continuously learn the neutral angle wConvTASAngle9;

[0082] D. Until the final tx (3Δt < tx) phase: when the neutral position is re-enabled, as in step 3, continuously learn the neutral position angles wConvTASAngle(n-2), wConvTASAngle(n-1), and wConvTASAngle n.

[0083] The final neutral position learning value is continuously updated based on the real-time learned value of the neutral position, and the algorithm formula is: wConvTASAngle final =(wConvTASAngle final_Delay ×(n-1)+wConvTASAngle current ) / n;

[0084] The neutral position angle learning value does not need to store a large number of learning values to calculate the final learning value, and the learning value can be calculated in real time. When the vehicle is driving at a low speed, the high-speed learning value is stored in the data storage function module, that is, the originally learned neutral position value is updated.

[0085] As shown in Figure 7 , the self-learning value compensation strategy for dynamic neutral position self-learning of the present invention is specifically as follows: the neutral position value stored in the data storage function module at a low speed will update the currently used neutral position value in time after storage. To avoid an instantaneous change in angle caused by instantaneously updating the neutral position value, the original current neutral position learning value is compensated through a low-pass filter to gradually compensate to the actual steering wheel neutral position value, so that the entire process of neutral position angle self-learning is completed.

[0086] The compensation strategy is determined according to the following formula: D n =p*Δ+(1-p)*D n-1 ; where D n is the current neutral position angle compensation value; D n-1 is the previous neutral position angle compensation value; Δ is the difference between the compensation target neutral position angle and the actual neutral position angle; p is the compensation coefficient.

Claims

1. A method for mid-level dynamic self-learning in a steer-by-wire system, characterized in that: The specific method and process are as follows: (1) The lower steering execution system obtains information on the vehicle speed, wheel speed, yaw rate and lateral acceleration signals of the whole vehicle through the whole vehicle CAN, and obtains the steering wheel torque and steering wheel angle information of the upper steering road feel system through the private CAN, which are used for the determination of the straight detection function module of the lower steering execution system. (2) The steering execution system activates the center position self-learning function through the straight-line detection function module: When the vehicle is at high speed, the TAS angle acquisition module frequently acquires the TAS angle signal, and the center position self-learning function module learns the center position angle through the algorithm; when the vehicle is at low speed, the learned center position angle is stored in the data storage function module. (3) The median self-learning function module of the downstairs steering execution system will read the median learning value from the data storage function module, compensate the current median learning value by low-pass filtering, and provide it to the angle calculation module of the downstairs steering execution module; (4) The mid-position self-learning function module and the RPS motor rotor position acquisition module calculate the current angle of the steering execution system for use by angle-related function modules; In step (2), the frequency acquisition strategy in the TAS angle acquisition module is as follows: To learn the median value more accurately, during the median self-learning process, the median value wConvTASAngle1 is learned at the moment of activation, and then wConvTASAngle2 is learned at an interval of Δt. This learning continues as long as the straight-line detection is satisfied, until the preset nth angle value wConvTASAngle n is obtained. The preset number of samples n is determined based on the performance of the learned median value, and n is 100-200. During the high-speed straight-line process, the self-learning process continues for t1, t2, t3...tn, until the preset n values ​​are learned, where n is 100-200; In segment t1: when the median is activated, the median angle wConvTASAngle1 is learned; then, at interval Δt, wConvTASAngle2 is learned, Δt... <t1<2Δt; In segment t2: when the median reopens, continue learning the median angles wConvTASAngle3, wConvTASAngle4, wConvTASAngle5, wConvTASAngle6, wConvTASAngle7, wConvTASAngle8, and 5Δt. <t2<6Δt; In segment t3: when the median re-opens, continue learning the median angle wConvTASAngle9,0 <t3<Δt; Until the final tn segment: when the median is reactivated, continue learning the median angles wConvTASAngle(n-2), wConvTASAngle(n-1), wConvTASAngle n, 3Δt <tn; The final median learned value is continuously updated based on the median value learned in real time. The algorithm formula is wConvTASAngle. final =(wConvTASAngle final_Delay ×(n-1)+wConvTASAngle current ) / n; where wConvTASAngle final wConvTASAngle is the median angle learned value calculated through n iterations. final_Delay wConvTASAngle is the median angle learned value calculated after n-1 iterations. current The median angle is the learning value of the current nth learning iteration, where n is the number of times the median angle has been learned. The median angle does not need to store a large number of learning values ​​to calculate the final learning value. The learning value is calculated in real time. When the vehicle is traveling at low speed, the high-speed learning value is stored in the data storage module, which updates the original learned median value.

2. The method for mid-level dynamic self-learning in a steer-by-wire system according to claim 1, characterized in that: To enable the median self-learning module to perform median learning in a relatively stable straight-line environment, straight-line detection of the vehicle's driving conditions is required. The method for straight-line detection is as follows: S1 detects the steering wheel torque, vehicle speed, wheel speed and wheel speed difference of the four wheels, speed change gradient, yaw rate, steering shaft speed of the lower steering wheel, and steering wheel angle of the upper steering wheel in the steer-by-wire system; and the current value of the lower steering TAS sensor. S2, determine whether the following conditions are all satisfied: 1) The difference between the current TAS sensor value and the previously learned angle value is less than a first preset threshold; 2) The steering wheel torque is less than the second preset threshold; 3) The vehicle speed exceeds the preset third threshold; 4) The difference between the wheel speed of the left front wheel and the wheel speed of the left rear wheel is less than the fourth preset threshold; 5) The difference between the wheel speed of the right front wheel and the wheel speed of the right rear wheel is less than the fifth preset threshold; 6) The difference between the sum of the wheel speeds of the two sets of diagonal wheels is less than the sixth preset threshold; 7) The gradient of the vehicle speed change is less than the seventh threshold. 8) The steering shaft speed under drive-by-wire steering is less than the eighth preset threshold; 9) The steering wheel angle on the drive-by-wire system is less than the ninth preset threshold; 10) The yaw rate of the entire vehicle is less than the tenth preset threshold; S3. If all the conditions in step S2 are met and within a stable time range, the median self-learning function module in the downward steering execution system is triggered. The angle value of the TAS angle sensor is the learned median value, and the learned median value is recorded by the algorithm.

3. The method for mid-level dynamic self-learning in a steer-by-wire system according to claim 1, characterized in that: The compensation strategy is as follows: the median value is stored at low speed in the data storage module. After storage, the median value is updated in a timely manner. In order to avoid instantaneous changes in the angle caused by instantaneous updates of the median value, the original current median learning value is gradually compensated to the actual steering wheel median value through a low-pass filter, thus completing the entire process of self-learning the median angle.

4. The method for mid-level dynamic self-learning in a steer-by-wire system according to claim 1, characterized in that: The aforementioned frequency sampling involves continuously collecting data at predetermined time intervals, with the time interval being 50-100ms, until the required number of samples is met, provided that the straight-line detection conditions are met.

5. The method for mid-level dynamic self-learning in a steer-by-wire system according to claim 1, characterized in that: The system includes an upper steering road feel system, a vehicle signal processing system, and a lower steering actuation system. The upper steering road feel system and the lower steering actuation system are connected via a private CAN bus, and the upper steering road feel system and the lower steering actuation system are connected to the vehicle signal processing system via the vehicle CAN bus. The aforementioned upper steering road feel system includes a steering wheel torque acquisition module and a steering wheel angle acquisition module, which are used to calculate the torque applied to the steering wheel and the angle of the upper steering road feel system, and transmit them to the lower steering execution system via a private CAN bus for the straight-line detection function module of the lower steering execution system; The vehicle signal processing system includes a vehicle speed acquisition module, a wheel speed acquisition module, a yaw rate acquisition module, and a lateral acceleration acquisition module. These modules are used to calculate vehicle speed signals, wheel speed signals, yaw rate signals, and lateral acceleration signals, and transmit them to the lower steering execution system via the vehicle CAN bus for use in the straight-line detection function module of the lower steering execution system. The described downward steering execution system includes a straight-line detection module, a center-position self-learning module, a TAS angle acquisition module, a fault handling module, a data storage module, a downward steering execution module angle calculation module, and an RPS motor rotor position acquisition module. The straight-line detection module determines the vehicle's straight-line conditions to activate the center-position self-learning function, involving vehicle speed signals, wheel speed signals, yaw rate signals, lateral acceleration signals, steering wheel torque signals, and steering wheel angle signals. The TAS angle acquisition module provides the angle for center-position learning. The center position self-learning module is used for high-speed self-learning of the center position angle, low-speed storage of the center position learning value, and provision of the learning value to the angle calculation module of the downward steering execution module; the RPS motor rotor position acquisition module provides the motor rotor angle position to the angle calculation module of the downward steering execution module, and calculates the angle based on the center position; the data storage module serves as the function of storing the center position self-learning value, and retains the learning value after the EPS is powered on and off, satisfying the non-volatility of the center position self-learning value; the fault handling module stores the fault and executes relevant safety measures when unexpected faults are encountered during learning.

Images