A method, device, and medium for generating road feel for steer-by-wire based on time-series attribution
By acquiring vehicle status information to calculate driver steering activity and response residuals, and combining time-series correlation to generate virtual road feel torque, the problem of recognizability and consistency of road feel simulation in steer-by-wire systems is solved, improving the accurate recognition of driver response and driving experience.
Patent Information
- Application Number
- CN202611579603.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-17
AI Technical Summary
In existing steer-by-wire systems, the virtual road feel simulation scheme fails to accurately distinguish between driver-related and non-driver-related responses, resulting in insufficient road feel identifiability and time consistency, which can easily lead to false prompts or human-machine interaction.
By acquiring vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, the driver's steering activity and expected yaw rate are calculated, the yaw and lateral response residuals are determined, and based on temporal correlation and consistency, driver-related responses and non-driver-related responses are distinguished to generate normal responses and disturbance warning road feel torques.
It achieves accurate differentiation of driver-related responses, avoids false prompts and human-machine conflicts, ensures the temporal consistency and recognizability of virtual road sense, and improves the driving experience.
Smart Images

Figure CN123104772A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of vehicle steer-by-wire chassis technology, and in particular to a method, device and medium for generating steer-by-wire road feel based on time-series attribution. Background Technology
[0002] In traditional mechanical steering systems, lateral tire forces, self-centering torque, and road impacts are transmitted to the steering wheel along the mechanical drive chain, allowing the driver to perceive the vehicle's steering response. With steer-by-wire systems, the mechanical connection between the steering wheel and the steering wheels is eliminated, and the road feel is actively reconstructed by the road feel motor based on the vehicle's status.
[0003] Existing virtual road feel typically consists of basic stiffness, damping, friction, self-centering torque, and dynamic torque reflecting the vehicle's lateral movement. Yaw rate and lateral acceleration are used to reflect the vehicle's steering response or stability. However, yaw rate and lateral acceleration only describe the lateral movement the vehicle has generated, not its cause. The same magnitude of yaw rate or lateral acceleration could originate from normal driver steering, crosswinds, low-traction road surfaces, or intervention of the vehicle stability control system. Furthermore, vehicle speed, steering wheel input, yaw rate, and lateral acceleration may be provided by different controllers and sensors with varying sampling periods, communication delays, and filtering delays. After the driver turns the steering wheel, the vehicle's yaw response and lateral acceleration exhibit dynamic delays that vary with vehicle speed, tires, and vehicle parameters.
[0004] In existing technologies, road sense simulation schemes that fuse multi-source signals typically directly map or weight vehicle states such as vehicle speed, steering wheel angle, yaw rate, and lateral acceleration into feedback torque. Because they fail to distinguish between lateral responses caused by driver-driven steering and those caused by non-driver factors, and do not address the aforementioned dynamic delays, they are prone to misinterpreting normal steering as disturbance warnings, or outputting warning torques that conflict with the driver's corrective actions, resulting in insufficient road sense discernibility and temporal consistency, and potentially leading to human-machine conflict.
[0005] Therefore, how to accurately distinguish between driver-related responses and non-driver-related responses without adding a rack force sensor, so as to generate a recognizable and time-consistent virtual road feel, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention
[0006] In view of this, the present disclosure aims to provide a method, apparatus and medium for generating steer-by-wire road feel based on time-series attribution, in order to solve the problem that existing virtual road feel methods do not distinguish the source of the vehicle's lateral response, resulting in insufficient road feel identifiability and temporal consistency.
[0007] To achieve the above objectives, the technical solution disclosed herein is implemented as follows:
[0008] The first aspect of this disclosure provides a method for generating steer-by-wire road feel based on time-series attribution, comprising the following steps:
[0009] S1. Obtain vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration;
[0010] S2. Determine the driver's steering activity based on the steering wheel angle, the steering wheel angular velocity, and the driver's input torque;
[0011] S3. Calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and the steering wheel angle;
[0012] S4. Calculate the yaw response residual between the actual yaw rate and the desired yaw rate, and the lateral response residual between the actual lateral acceleration and the desired lateral acceleration;
[0013] S5. Within the delay search interval related to the vehicle's longitudinal speed, determine the first temporal correlation between the steering wheel angular velocity and the rate of change of yaw angular velocity, and the second temporal correlation between the steering wheel angular velocity and the rate of change of lateral acceleration, and combine directional consistency and amplitude consistency to determine the driver's attribution confidence.
[0014] S6. Determine driver-related and non-driver-related responses based on the expected yaw rate, the expected lateral acceleration, the yaw response residual, the lateral response residual, and the driver attribution confidence.
[0015] S7. Generate normal response road feel torque based on the driver-related response, and generate disturbance alert road feel torque based on the non-driver-related response;
[0016] S8. The base road feel torque, the normal response road feel torque, the disturbance alert road feel torque, and the return road feel torque are superimposed to obtain the target road feel torque, which is then output to the road feel motor.
[0017] Furthermore, S2 includes: normalizing the steering wheel angle, the steering wheel angular velocity, and the driver input torque respectively to obtain normalized values for the steering wheel angle, the steering wheel angular velocity, and the driver input torque.
[0018] The driver's steering activity is obtained by weighted summing the normalized values of the steering wheel angle, the steering wheel angular velocity, and the driver's input torque.
[0019] Furthermore, S3 includes: determining the steering gear ratio based on the vehicle's longitudinal speed, and dividing the steering wheel angle by the steering gear ratio to obtain a reference front wheel angle;
[0020] Based on the reference front wheel steering angle and the vehicle longitudinal speed, the steady-state expected yaw rate is calculated using the vehicle reference model.
[0021] A first-order dynamic element is applied to the steady-state desired yaw rate to obtain the desired yaw rate;
[0022] The desired lateral acceleration is calculated based on the product of the desired yaw rate and the vehicle's longitudinal speed.
[0023] Furthermore, in S5, determining the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, includes:
[0024] The delay search interval is determined based on the vehicle's longitudinal speed;
[0025] Within the delay search interval, calculate the first cross-correlation value between the steering wheel angular velocity sequence and the yaw rate of change sequence, and the second cross-correlation value between the steering wheel angular velocity sequence and the lateral acceleration rate of change sequence;
[0026] The first cross-correlation value is used as the first time-series correlation, and the second cross-correlation value is used as the second time-series correlation.
[0027] Furthermore, in S5, the driver's attribution confidence is determined by combining directional consistency and amplitude consistency, including:
[0028] Based on the directional relationship between the steering wheel angle, the actual yaw rate, and the actual lateral acceleration within a set time window, directional consistency is determined;
[0029] The driver's steering activity, the first temporal correlation, the second temporal correlation, the directional consistency, and the amplitude consistency are fused to obtain the driver's attribution confidence.
[0030] Amplitude consistency is determined based on the normalized relationship between the yaw response residual and the desired yaw angular velocity, and the normalized relationship between the lateral response residual and the desired lateral acceleration.
[0031] Furthermore, S6 includes: performing low-pass filtering on the yaw response residual and the lateral response residual respectively to obtain the yaw slow-varying residual and the lateral slow-varying residual;
[0032] The driver-related yaw response is obtained by adding the expected yaw rate to the product of the driver's attribution confidence and the yaw slow-varying residual.
[0033] The driver-related lateral response is obtained by adding the desired lateral acceleration to the product of the driver's attribution confidence and the lateral slow-varying residual.
[0034] Subtracting the driver-related yaw response from the actual yaw rate yields the non-driver-related yaw response.
[0035] Subtracting the driver-related lateral response from the actual lateral acceleration yields the non-driver-related lateral response;
[0036] The driver-related response includes the driver-related yaw response and the driver-related lateral response, and the non-driver-related response includes the non-driver-related yaw response and the non-driver-related lateral response.
[0037] Furthermore, S7 includes: generating the normal response road feel torque based on the driver-related yaw response and the driver-related lateral response after low-frequency mapping and mid-frequency mapping, respectively;
[0038] The disturbance warning road feel torque is generated after transient shaping based on the non-driver-related yaw response and the non-driver-related lateral response. The transient shaping includes at least one of bandpass filtering, finite duration pulse generation, peak hold followed by exponential decay, and piecewise function.
[0039] Furthermore, in step S8, the return-to-center torque is determined based on the steering wheel angle and the vehicle's longitudinal speed.
[0040] Furthermore, it also includes: calculating a stability risk index based on the yaw response residual and the lateral response residual;
[0041] A stability warning torque is generated based on the stability risk index;
[0042] The target road feel torque is obtained by superimposing the base road feel torque, the normal response road feel torque, the disturbance warning road feel torque, the stability warning torque, and the straightening road feel torque.
[0043] Furthermore, it also includes: determining a suggested correction direction based on the non-driver-related responses;
[0044] When the direction of the steering wheel angular velocity is consistent with the suggested correction direction, and the non-driver-related response decreases after the driver's operation, it is determined that the driver is making an effective correction;
[0045] When it is determined that the driver is making an effective correction and the prompt torque is obstructing the direction of steering wheel movement, the disturbance prompt torque and the stability prompt torque are attenuated to obtain the attenuated prompt torque.
[0046] The step of superimposing the base road feel torque, the normal response road feel torque, the disturbance warning road feel torque, the stability warning torque, and the homing road feel torque to obtain the target road feel torque includes:
[0047] When it is determined that the driver is making an effective correction and the prompt torque is obstructing the direction of steering wheel movement, the base road feel torque, the normal response road feel torque, the attenuated prompt torque, and the return-to-center road feel torque are superimposed to form the target road feel torque.
[0048] Furthermore, it also includes: applying amplitude and rate of change limits to the target road-feel torque;
[0049] When the available capacity of the road sense motor is insufficient, the components are reduced in the following order of priority: disturbance warning road sense torque, stability warning torque, normal response road sense torque, and basic road sense torque.
[0050] When the actual yaw rate or the actual lateral acceleration fails, the disturbance indication torque and the stability indication torque are smoothly attenuated to zero according to a preset slope, while the basic road feel torque and the straightening road feel torque are retained.
[0051] Furthermore, in step S1, after acquiring the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, the method further includes:
[0052] Based on at least one of the timestamp, message counter, and reception time, the vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration are resampled to a unified control cycle.
[0053] The resampled vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration are each subjected to low-pass filtering.
[0054] Perform at least one of the following checks on the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration: range check, update timeout check, signal freeze check, mutation rate check, and physical consistency check, to obtain a confidence index for the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration.
[0055] A second aspect of this disclosure provides a steer-by-wire road feel generation device, comprising:
[0056] The signal acquisition module is used to acquire vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate and actual lateral acceleration;
[0057] The driver status recognition module is used to determine the driver's steering activity based on the steering wheel angle, the steering wheel angular velocity, and the driver's input torque.
[0058] The reference vehicle response calculation module is used to calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and the steering wheel angle.
[0059] The residual calculation module is used to calculate the yaw response residual between the actual yaw rate and the expected yaw rate, and the lateral response residual between the actual lateral acceleration and the expected lateral acceleration;
[0060] The temporal attribution module is used to determine the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, within the delay search interval related to the vehicle's longitudinal speed, and to determine the driver's attribution confidence by combining directional consistency and amplitude consistency.
[0061] The response component estimation module is used to determine driver-related and non-driver-related responses based on the expected yaw rate, the expected lateral acceleration, the yaw response residual, the lateral response residual, and the driver attribution confidence.
[0062] The road feel torque generation module is used to generate a normal response road feel torque based on the driver-related response and to generate a disturbance warning road feel torque based on the non-driver-related response.
[0063] The torque superposition module is used to superimpose the basic road feel torque, the normal response road feel torque, the disturbance indication road feel torque, and the return road feel torque to obtain the target road feel torque, and output it to the road feel motor.
[0064] A third aspect of this disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the time-attribution-based steer-by-wire road feel generation method described in the first aspect.
[0065] Compared with the prior art, this disclosure can achieve the following beneficial effects:
[0066] By acquiring vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, the driver's steering activity is determined, and the desired yaw rate and desired lateral acceleration are calculated. Within the delay search interval related to the vehicle's longitudinal speed, the first and second temporal correlations between the steering wheel angular velocity and the rates of change of yaw rate and lateral acceleration are determined. The driver's attribution confidence is determined by combining directional consistency and amplitude consistency, thereby determining the driver-related response and non-driver-related response, and generating normal response road feel torque and disturbance warning road feel torque respectively. The target road feel torque is obtained by superimposing the base road feel torque, normal response road feel torque, disturbance warning road feel torque, and centering road feel torque. Therefore, it is possible to distinguish between driver-related responses and non-driver-related responses, avoid misinterpreting normal steering as a disturbance warning, avoid outputting warning torque for directional conflict when the driver has already corrected the direction, and compensate for the dynamic delay between driver steering and vehicle lateral response, so that the virtual road sense is consistent in time. This solves the problem that existing virtual road sense methods do not distinguish the source of vehicle lateral response, resulting in insufficient road sense identifiability and temporal consistency. Attached Figure Description
[0067] The accompanying drawings, which form part of this disclosure, are used to provide a further understanding of this disclosure. The illustrative embodiments of this disclosure and their descriptions are used to explain this disclosure and do not constitute an undue limitation of this disclosure. In the drawings:
[0068] Figure 1 A flowchart of a time-attribution-based steer-by-wire road feel generation method provided in an embodiment of this disclosure;
[0069] Figure 2 This is a schematic diagram of the generation of road feel torque by type and the suppression of human-machine conflict provided in the embodiments of this disclosure;
[0070] Figure 3 This is a schematic diagram of the structure of the steer-by-wire road feel generation device based on time-series attribution provided in an embodiment of this disclosure. Detailed Implementation
[0071] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this disclosure. However, those skilled in the art will readily recognize that some features may be omitted in different circumstances, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this disclosure are not shown or described in the specification. This is to avoid obscuring the core parts of this disclosure with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.
[0072] It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined to form various implementations. Furthermore, the steps or actions in the method description can be rearranged or adjusted in a manner readily apparent to those skilled in the art. Therefore, the various orders in the specification and drawings are merely for the clear description of a particular embodiment and do not imply a mandatory order, unless otherwise stated that a particular order must be followed.
[0073] In the description of this disclosure, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing this disclosure and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this disclosure. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this disclosure, unless otherwise stated, "a plurality of" means two or more.
[0074] In the description of this disclosure, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linkage" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art will understand the specific meaning of the above terms in this disclosure based on the specific circumstances.
[0075] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.
[0076] like Figures 1 to 2 As shown, the first aspect of this disclosure provides a method for generating steer-by-wire road feel based on time-series attribution, comprising the following steps:
[0077] S1. Obtain vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration.
[0078] In some embodiments, the time-attribution-based steer-by-wire road feel generation method is executed by a road feel controller. The road feel controller acquires vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration according to a preset control cycle.
[0079] The steering wheel angular velocity can be directly provided by the steering wheel angle sensor, or it can be obtained by differential and filtering from the steering wheel angle; the driver input torque can be measured by the steering wheel torque sensor, or it can be estimated by the road feel motor torque, steering column dynamics and friction model.
[0080] Specifically, the input vector is represented as:
[0081] ,
[0082] in, Represented as an input vector, Represented as the first The input vector for each control cycle. Expressed as the vehicle's longitudinal speed. Represented as the first Vehicle longitudinal speed per control cycle This is represented by the steering wheel angle. Represented as the first Steering wheel angle per control cycle Expressed as the angular velocity of the steering wheel. Represented as the first Steering wheel angular velocity per control cycle This is expressed as the driver's input torque. Represented as the first Driver input torque per control cycle Expressed as the actual yaw rate, Represented as the first The actual yaw rate of each control cycle Expressed as actual lateral acceleration, Represented as the first The actual lateral acceleration per control cycle. This is represented as the control cycle number. This is represented as a vector transpose.
[0083] Specifically, the control period can be from 5 milliseconds to 20 milliseconds. Preferably, the control period is 10 milliseconds. It is understood that the control period can be calibrated according to the response capability of the road sensor motor and the communication rate of the vehicle network, and this disclosure does not limit it.
[0084] In step S1, after obtaining the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, the following steps are also included:
[0085] Based on at least one of the timestamp, message counter, and reception time, the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration are resampled to a unified control cycle. Since the signals may originate from different electronic control units, the road sense controller resamples the signals to a unified control cycle based on the timestamp, message counter, or reception time. For signals with slower sampling periods, zero-order hold or linear interpolation is used; for signals with fixed communication delays, calibration delay is used for forward compensation. Other resampling methods such as nearest neighbor interpolation and cubic spline interpolation can also be used, and no limitations are imposed here.
[0086] Low-pass filtering is applied to the resampled vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration. For transient response extraction, high-pass or band-pass components can be calculated for yaw rate and lateral acceleration. Low-pass filtering can also employ Butterworth filtering, Kalman filtering, etc., and can be set according to actual needs; no limitations are imposed here.
[0087] For vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, at least one of the following checks is performed: range check, update timeout check, signal freeze check, abrupt change rate check, and physical consistency check, to obtain a reliability flag for each of these parameters. Specifically, a reliability flag is established for each signal simultaneously, and the reliability judgment includes at least range check, update timeout check, signal freeze check, abrupt change rate check, and physical consistency check. When the vehicle speed signal fails, the system adopts conservative fixed basic road sense parameters; when the yaw rate or lateral acceleration fails, the dynamic response road sense and disturbance warning road sense are smoothly attenuated to zero according to a preset slope, retaining only the basic stiffness, damping, friction, and return-to-center road sense; when the driver input torque fails, the system stops performing human-machine conflict attenuation, but can still generate basic road sense based on the steering wheel angle and steering wheel angular velocity. Reliability assessment may also include signal correlation checks, signal redundancy cross-checks, etc., which can be set according to actual needs and are not limited here.
[0088] S2. Determine the driver's steering activity based on the steering wheel angle, steering wheel angular velocity, and driver input torque.
[0089] Step S2 includes: normalizing the steering wheel angle, steering wheel angular velocity, and driver input torque respectively to obtain normalized values for the steering wheel angle, steering wheel angular velocity, and driver input torque; and weighted summing the normalized values for the steering wheel angle, steering wheel angular velocity, and driver input torque to obtain the driver's steering activity.
[0090] Specifically, the three types of signals are normalized:
[0091] ;
[0092] in, Represented as the normalized value of the steering wheel angle. Represented as the normalized value of the steering wheel angular velocity. This is represented as the normalized value of the driver's input torque. This is represented by the steering wheel angle. Expressed as the angular velocity of the steering wheel. This is expressed as the driver's input torque. This is represented as the reference value for steering wheel angle calibration. This is represented as the reference value for steering wheel angle calibration. This represents the driver's input torque calibration reference value. This means that the result is limited to the range of 0 to 1.
[0093] Pilot steering activity is expressed as:
[0094]
[0095] ;
[0096] in, This represents the pilot's steering activity level. The weights are represented as the normalized values of the steering wheel angle. The weights are represented as the normalized values of the steering wheel angular velocity. This represents the weight of the driver's input torque normalized value.
[0097] Specifically, to avoid misjudging steady-state turns with steering wheel angular velocity close to zero as a lack of driver involvement, the road feel controller uses steering wheel angle, steering wheel angular velocity, and driver input torque simultaneously to construct driver steering activity.
[0098] In some embodiments, a driver state machine with hysteresis is established based on the driver's steering activity, steering wheel angular velocity direction, driver input torque direction, and the duration of each state. This state machine includes at least the following states: release state, steering wheel hold state, active steering state, rapid steering state, reverse correction state, and emergency correction state. State transitions must meet entry thresholds, exit thresholds, and minimum durations to prevent frequent switching caused by sensor noise.
[0099] S3. Calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and steering wheel angle.
[0100] Step S3 includes: determining the steering ratio based on the vehicle's longitudinal speed; dividing the steering wheel angle by the steering ratio to obtain the reference front wheel angle; calculating the steady-state desired yaw rate using a vehicle reference model based on the reference front wheel angle and the vehicle's longitudinal speed; applying a first-order dynamic element to the steady-state desired yaw rate to obtain the desired yaw rate; and calculating the desired lateral acceleration based on the product of the desired yaw rate and the vehicle's longitudinal speed.
[0101] Specifically, for vehicles with a fixed steering ratio, the reference front wheel steering angle is...
[0102] ;
[0103] For variable ratio steer-by-wire vehicles, the steering ratio Determined based on vehicle longitudinal speed and driving mode:
[0104] ;
[0105] in, Indicated as the reference front wheel steering angle, This is represented by the steering wheel angle. This is expressed as the steering gear ratio. Expressed as the vehicle's longitudinal speed. This is represented as the control cycle number.
[0106] Preferably, a linear two-degree-of-freedom vehicle reference model is used to calculate the steady-state expected yaw rate:
[0107] ;
[0108] in, Represented as the first The steady-state expected yaw rate for each control cycle. Represented as the first Vehicle longitudinal speed per control cycle This refers to the vehicle's wheelbase. Represented as the vehicle stability factor, Represented as the first The reference front wheel steering angle for each control cycle.
[0109] Specifically, to reflect the establishment process of the vehicle's yaw response, a first-order dynamic element is added to the steady-state expected yaw rate:
[0110] ;
[0111] in, Expressed as a time constant related to the vehicle's longitudinal speed. Expressed as the rate of change of the desired yaw rate, Expressed as the desired yaw rate, Expressed as the steady-state expected yaw rate,
[0112] The desired lateral acceleration can be directly output from the two-degree-of-freedom model, and in a simplified implementation, it can be obtained by the following formula:
[0113] ;
[0114] in, Represented as the first The expected lateral acceleration per control cycle. This is expressed as the desired yaw rate of the control period.
[0115] In some embodiments, the reference model may also employ a two-dimensional lookup table of vehicle speed and steering wheel angle, a piecewise linear model, an online vehicle parameter identification model, or a hybrid model corrected for normal road surface data. This disclosure does not limit the specific form of the reference model, only requiring that the model be able to provide the expected yaw rate and expected lateral acceleration that a normal vehicle should produce based on the vehicle's longitudinal speed and steering wheel angle. Optionally, the first-order dynamic element may also employ a second-order dynamic element, a delay element, etc.; the reference model may also employ a neural network model, a support vector machine model, etc., and this disclosure does not limit its application in this regard.
[0116] S4. Calculate the yaw response residual between the actual yaw rate and the desired yaw rate, and the lateral response residual between the actual lateral acceleration and the desired lateral acceleration.
[0117] In some embodiments, the yaw response residual between the actual yaw rate and the desired yaw rate, and the lateral response residual between the actual lateral acceleration and the desired lateral acceleration are calculated:
[0118]
[0119] ;
[0120] in, Represented as the first Yaw response residuals for each control cycle Represented as the first The actual yaw rate of each control cycle Represented as the first The expected yaw rate for each control cycle, Represented as the first Lateral response residuals for each control cycle Represented as the first The actual lateral acceleration per control cycle. Represented as the first The desired lateral acceleration for each control cycle.
[0121] Specifically, the response residual is not directly equivalent to the external disturbance. To reduce the impact of slow changes in vehicle parameters, load variations, and reference model bias on disturbance judgment, the residual is decomposed into slow-varying components and transient components:
[0122] ;
[0123] ;
[0124] in, This is represented as a yaw rate slow-changing residual. Represented as the yaw transient residual, Represented as laterally slowly varying residuals. Represented as lateral transient residuals, It is represented as a low-pass filter.
[0125] Slowly varying residuals are primarily used to correct long-term deviations between the reference model and the actual vehicle, while transient residuals are mainly used to describe the fast vehicle response, which is weakly correlated with driver input. The filter cutoff frequency is calibrated with vehicle speed and sensor noise levels. Optionally, low-pass filtering can also be implemented using complementary filtering, wavelet decomposition, or a state observer, which can be set according to actual needs and is not limited here.
[0126] S5. Within the delay search interval related to the vehicle's longitudinal speed, determine the first temporal correlation between the steering wheel angular velocity and the rate of change of yaw rate, and the second temporal correlation between the steering wheel angular velocity and the rate of change of lateral acceleration. Combine this with directional consistency and amplitude consistency to determine the driver's attribution confidence.
[0127] In step S5, determining the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, includes: determining a delay search interval based on the vehicle's longitudinal speed; within the delay search interval, calculating the first cross-correlation value between the steering wheel angular velocity sequence and the yaw rate of change sequence, and the second cross-correlation value between the steering wheel angular velocity sequence and the lateral acceleration rate of change sequence; using the first cross-correlation value as the first temporal correlation and the second cross-correlation value as the second temporal correlation.
[0128] Specifically, the road sense controller does not only compare driver input and vehicle response at the same time, but also calculates the temporal correlation between the driver input sequence and the vehicle response sequence within a delay search interval related to the vehicle's longitudinal speed.
[0129] Specifically, the delay search interval is first determined based on the vehicle's longitudinal speed. Within a sliding window of length N control cycles, the normalized cross-correlation between the steering wheel angular velocity and the rate of change of yaw rate, and between the steering wheel angular velocity and the rate of change of lateral acceleration, are calculated respectively, and the delay corresponding to the maximum correlation value is selected:
[0130]
[0131]
[0132] in, This is expressed as the optimal delay between the rate of change of steering wheel angular velocity and yaw rate. Represented as the first cross-correlation value, This is expressed as the optimal delay between the steering wheel angular velocity and the rate of change of lateral acceleration. This is represented as the second cross-correlation value. Expressed as the angular velocity of the steering wheel. Expressed as the actual rate of change of yaw rate, Expressed as the actual rate of change of lateral acceleration, Represents cross-correlation operations. This represents the delay amount that makes the function reach its maximum value. , This represents the maximum value within the delayed search interval. Represented as the first Steering wheel angle per control cycle.
[0133] When the driver is holding the steering wheel, the steering wheel angular velocity is close to zero, making it unsuitable to use only dynamic correlation coefficients. Therefore, the system uses the steady-state consistency between steering wheel angle, driver holding torque, low-frequency yaw rate, and low-frequency lateral acceleration in the holding state; in active steering and rapid steering states, it mainly uses rate-of-change time-series correlation.
[0134] In step S5, the driver's attribution confidence is determined by combining directional consistency and amplitude consistency, including: determining directional consistency based on the directional relationship between the steering wheel angle and the actual yaw rate and the actual lateral acceleration within a set time window; fusing the driver's steering activity, the first time-series correlation, the second time-series correlation, directional consistency, and amplitude consistency to obtain the driver's attribution confidence; and determining amplitude consistency based on the normalized relationship between the yaw response residual and the expected yaw rate, and the normalized relationship between the lateral response residual and the expected lateral acceleration.
[0135] Specifically, directional consistency does not merely judge the sign of a single sampling point, but rather statistically analyzes whether the steering wheel input direction, yaw response direction, and lateral response direction satisfy the calibration relationship in the vehicle coordinate system within a set time window, and sets dead zones, hysteresis, and minimum durations. The directional consistency index can be denoted as... .
[0136] Specifically, the consistency of the amplitude is calculated based on the normalized residuals of the reference response and the actual response:
[0137]
[0138] ;
[0139] in, This indicates consistency in yaw amplitude. This is represented as lateral amplitude consistency. For the yaw response residual, For the lateral response residual, For the desired yaw rate, For the desired lateral acceleration, and This is a preset small quantity.
[0140] Specifically, the driver's steering activity, first-time-series correlation, second-time-series correlation, directional consistency, and amplitude consistency are fused to obtain the driver's attribution confidence score:
[0141]
[0142] in, This is represented as the driver's attribution confidence score. This represents the pilot's steering activity level. The first cross-correlation value, This is the second cross-correlation value. For consistency of direction, To ensure consistent yaw amplitude, For lateral amplitude consistency, , , , , , To preset weights, This means that the result is limited to the range of 0 to 1.
[0143] A larger value indicates that the vehicle's lateral response is primarily related to the driver's steering. A smaller value indicates the presence of significant non-driver-related components in the vehicle response. To avoid jumps in confidence levels, The filter is first-order filtered and different rate-of-change limits are set for rising and falling.
[0144] It is understandable that cross-correlation can also be implemented using dynamic time warping, mutual information, coherence functions, recursive least squares, Granger causality tests, Kalman filter innovation sequences, or trained time series classification models. It can be designed according to actual needs, and no restrictions are imposed here.
[0145] S6. Determine the driver-related and non-driver-related responses based on the expected yaw rate, expected lateral acceleration, yaw response residual, lateral response residual, and driver attribution confidence.
[0146] In step S6, the following steps are performed: low-pass filtering is applied to the yaw response residual and the lateral response residual to obtain the yaw slow-varying residual and the lateral slow-varying residual, respectively; the desired yaw rate is added to the product of the driver's attribution confidence and the yaw slow-varying residual to obtain the driver-related yaw response; the desired lateral acceleration is added to the product of the driver's attribution confidence and the lateral slow-varying residual to obtain the driver-related lateral response; the actual yaw rate is subtracted from the driver-related yaw response to obtain the non-driver-related yaw response; the actual lateral acceleration is subtracted from the driver-related lateral response to obtain the non-driver-related lateral response; wherein, the driver-related response includes the driver-related yaw response and the driver-related lateral response, and the non-driver-related response includes the non-driver-related yaw response and the non-driver-related lateral response.
[0147] Specifically, this disclosure does not use the method of simply multiplying the actual yaw rate by the attribution coefficient for decomposition, but instead uses the method of "reference response plus credible slow-varying residual correction" to determine the driver's relevant response:
[0148]
[0149] ;
[0150] in, This is represented as the driver-related yaw response. This is represented as the driver-related lateral response. Expressed as the desired yaw rate, Expressed as the desired lateral acceleration, This is represented as the driver's attribution confidence score. This is represented as a yaw rate slow-changing residual. It is represented as a laterally slowly varying residual.
[0151] Subtracting the driver-related response from the actual response yields the non-driver-related yaw response and the non-driver-related lateral response:
[0152]
[0153] ;
[0154] in, This is represented as a non-driver-related yaw response. This is represented as a non-driver-related lateral response. Expressed as the actual yaw rate, Expressed as actual lateral acceleration, This is represented as the driver-related yaw response. This is represented as the driver-related lateral response.
[0155] Specifically, when At higher levels, the reference model allows for the absorption of slowly varying residuals consistent with driver steering, making the driver's relevant response closer to the actual vehicle response; when At lower values, less of the slowly varying residuals enter the driver-related components, and the remaining response is retained as non-driver-related components. This approach avoids treating all model errors directly as disturbances, while also avoiding simply proportionally partitioning the actual response.
[0156] In some embodiments, based on driver status, , , , and Construct a state machine with hysteresis to identify at least the following states: normal active steering, steady-state cornering, non-driver disturbances, under-response, over-response, driver-induced correction, and signal degradation. The system will only... and Defined as a "non-driver-related response," it is not mandatory to further classify it as crosswind, road impact, or stability control intervention, thereby avoiding unreliable judgments of disturbance sources when additional sensors are unavailable.
[0157] It is understandable that response component estimation can also be achieved using extended Kalman filtering, unscented Kalman filtering, sliding mode observer, perturbation observer, unknown input observer, or wavelet decomposition. The design can be tailored to actual needs and is not limited here.
[0158] S7. Generate normal response road feel torque based on driver-related responses, and generate disturbance warning road feel torque based on non-driver-related responses.
[0159] In some embodiments, S7 includes: generating a normal response road feel torque based on the driver-related yaw response and the driver-related lateral response after low-frequency mapping and mid-frequency mapping, respectively; and generating a disturbance warning road feel torque based on the non-driver-related yaw response and the non-driver-related lateral response after transient shaping, wherein the transient shaping includes at least one of bandpass filtering, finite-duration pulse generation, peak hold followed by exponential decay, and piecewise function.
[0160] Specifically, the driver-related vehicle response road feel is generated by the driver-related lateral acceleration and yaw rate:
[0161] ;
[0162] in, This represents the normal response road feel torque. Expressed as the lateral acceleration gain related to the vehicle's longitudinal speed. The yaw rate gain is related to the vehicle's longitudinal speed. For low-frequency or slowly varying mappings, For intermediate frequency mapping, For driver-related lateral responses, For driver-related yaw response, This refers to the vehicle's longitudinal speed.
[0163] Used to demonstrate the gradual establishment of lateral load; Used to reflect the vehicle's steering response speed. Yaw rate and lateral acceleration are measured at different frequency bands or with different gains to avoid the same vehicle motion information being repeatedly amplified.
[0164] Specifically, the non-driver-related response is first converted into the original disturbance cue torque:
[0165] ;
[0166] in, This is represented as the original disturbance indication torque. This is expressed as the non-driver-dependent yaw response gain related to the vehicle's longitudinal speed. This is expressed as the non-driver-related lateral response gain in relation to the vehicle's longitudinal speed. This is represented as a non-driver-related yaw response. This is represented as a non-driver-related lateral response. This is expressed as the vehicle's longitudinal speed.
[0167] Subsequently, through a transient shaper Generate disturbance feedback for road feel:
[0168] ;
[0169] in, This is represented as the disturbance indication torque. Represented as a transient shaper, It is represented as a Laplace transform variable.
[0170] Specifically, the transient shaper can be a bandpass filter, a finite-duration pulse generator, a peak-hold exponential attenuator, or a piecewise function, so that the disturbance cue manifests as a short-time bias, a single decaying pulse, or a finite-time additional damping, rather than continuously replicating the entire yaw rate and lateral acceleration onto the steering wheel. Optionally, transient shaping can also employ wavelet transform, empirical mode decomposition, etc.; low-frequency mapping and mid-frequency mapping can also employ lookup table methods, polynomial fitting, neural networks, etc., which are not limited in this disclosure.
[0171] In some embodiments, disturbance alerts may also employ short-time bias, exponentially decaying pulses, bandpass torque, finite additional damping, or torque gradient changes; stability alerts may employ resistance plateaus, directional pulses, low-amplitude periodic alerts, or combinations thereof, which can be set according to actual needs and are not limited herein.
[0172] S8. The base road feel torque, normal response road feel torque, disturbance warning road feel torque, and return road feel torque are superimposed to obtain the target road feel torque, which is then output to the road feel motor.
[0173] Specifically, the baseline road-feel torque is used to simulate steering wheel stiffness, damping, and steering mechanism friction:
[0174] ;
[0175] in, Represented as basic road feel torque, This is expressed as stiffness related to the vehicle's longitudinal speed. This is expressed as damping related to the vehicle's longitudinal speed. This is represented by the steering wheel angle. Expressed as the angular velocity of the steering wheel. This is expressed as the frictional simulation torque, which includes static friction, Coulomb friction, or smooth friction models. This is expressed as the vehicle's longitudinal speed.
[0176] The return torque is determined based on the steering wheel angle and the vehicle's longitudinal speed.
[0177] ;
[0178] in, This is expressed as the return-to-center torque. This is expressed as the positivity gain related to the vehicle's longitudinal speed and the driver's steering activity. Expressed as the vehicle's longitudinal speed. This represents the pilot's steering activity level. This is represented by the steering wheel angle.
[0179] When the driver actively holds the steering wheel, the return-to-center gain is appropriately reduced to prevent the road feel motor from continuously resisting the driver; when the driver releases the steering wheel and the vehicle's lateral response is normal, the return-to-center gain is increased to smoothly return the steering wheel to the center position. At this time, the target road feel torque is superimposed according to the following formula:
[0180] ;
[0181] in, Expressed as target road feel torque, Represented as basic road feel torque, This represents the normal response road feel torque. This is represented as the disturbance indication torque. This is expressed as the return-to-center torque.
[0182] In some embodiments, the time-attribution-based steer-by-wire road feel generation method may further include: calculating a stability risk index based on yaw response residuals and lateral response residuals; and generating a stability warning torque based on the stability risk index.
[0183] Specifically, stability risk indicators are constructed based on the yaw response residuals, lateral response residuals, and their rates of change:
[0184] ;
[0185] in, This is represented as a stability risk indicator. Represented as the yaw response residual, Represented as lateral response residuals, Expressed as the rate of change of the yaw response residuals, This is represented as the normalized baseline value of the yaw response residual. Represented as the normalized baseline value of the lateral response residual. This is expressed as the normalized baseline value of the rate of change of the yaw response residuals. , , This is represented as the preset weight.
[0186] when No stability warning is output when the value is below the first threshold; when When the steering wheel resistance gradient is between the first and second thresholds, gradually increase it; when When the second threshold is exceeded, a restricted directional cue or a finite pulse is output. The stability cue torque is denoted as... .
[0187] In the implementation that adds stability indicators, the target road feel torque is superimposed according to the following formula:
[0188] ;
[0189] in, Torque is provided to indicate stability.
[0190] When there is a risk to stability, the stability indication torque allows the driver to perceive the vehicle's stability status through the steering wheel.
[0191] In some embodiments, the steer-by-wire road feel generation method based on time-series attribution may further include: determining a suggested correction direction based on non-driver-related responses; determining that the driver is making an effective correction when the steering wheel angular velocity direction is consistent with the suggested correction direction and the non-driver-related responses decrease after the driver's operation; and attenuating the disturbance road feel torque and stability feel torque when the driver is making an effective correction and the indication torque creates resistance in the steering wheel movement direction, to obtain the attenuated indication torque.
[0192] Specifically, the road sense controller determines recommended correction directions that help reduce abnormal lateral responses based on non-driver-related yaw responses and non-driver-related lateral responses:
[0193] ;
[0194] in, This indicates a suggested direction for correction. This is represented as a non-driver-related yaw response. This is represented as a non-driver-related lateral response. This represents the weight of the non-driver-related yaw response. Represented as the weights of non-driver-related lateral responses, Represents a symbolic function.
[0195] When the direction of the steering wheel angular velocity is... A valid correction is determined when the driver's input torque exceeds the correction threshold, and the non-driver-related response begins to decrease after the driver's operation. A correction is not considered valid if the steering wheel is reversed without improving the vehicle's response.
[0196] Specifically, the sum of the disturbance warning torque and the stability warning torque is defined as the warning torque. :
[0197] ;
[0198] When the driver is making effective corrections and indicates that the torque is resisting the direction of steering wheel movement, i.e.
[0199] ;
[0200] Furthermore, if the degree of obstruction exceeds a threshold, a human-machine conflict is determined to exist. The attenuation coefficient is calculated based on the driver's correction intensity, the degree of obstruction from the indicated torque, and the degree of improvement in vehicle condition. :
[0201] ;
[0202] in, It is represented as the attenuation coefficient. This represents the preset minimum attenuation coefficient. This is represented as the preset gain. Indicated as the degree of conflict, This indicates taking the maximum value.
[0203] The corrected indicated torque is The road feel controller only attenuates disturbances and stability cues that conflict with correct driver corrections, without reducing the base road feel or normal vehicle response road feel; when driver directional corrections are detrimental to vehicle recovery, Keep it at 1 or limit it only to the safety upper limit.
[0204] In the implementation method that increases human-machine conflict attenuation, the target road feel torque is superimposed according to the following formula:
[0205] ;
[0206] Compared to the aforementioned superposition formula, this embodiment replaces the sum of the disturbance warning torque and the stability warning torque with the attenuated warning torque. .
[0207] In some embodiments, the steer-by-wire road feel generation method based on time-series attribution may further include: applying amplitude and rate-of-change limits to the target road feel torque; when the available capacity of the road feel motor is insufficient, reducing each component in order of priority: disturbance warning road feel torque, stability warning torque, normal response road feel torque, and basic road feel torque; when the actual yaw rate or actual lateral acceleration fails, smoothly attenuating the disturbance warning road feel torque and stability warning torque to zero according to a preset slope, while retaining the basic road feel torque and the return-to-center road feel torque.
[0208] Specifically, when the road feel motor temperature is too high, the bus voltage is insufficient, or the available torque is reduced, the disturbance warning and stability warning components are reduced first, then the normal response road feel is reduced, and finally the basic damping and return torque that meet the minimum safe feel requirements are retained.
[0209] To avoid sudden engagement or disengagement of dynamic road sense, all state transitions employ ramp, hysteresis, or weighted gradual transitions. If an abnormal increase in response residual, unstable attribution results, or output torque exceeding capability limits is detected in multiple consecutive control cycles, the system restores the previous set of valid parameters or enters the basic road sense degradation mode, and reports the fault status to the vehicle controller.
[0210] Alternatively, safety constraints can also be implemented using model predictive control, sliding mode control, etc.; fault degradation can also be implemented using lookup table method, state machine, etc., which can be designed according to actual needs and are not limited here.
[0211] Specifically, before the torque is sent to the road sensor motor, the target road sensor torque is subject to amplitude limits, rate of change limits, steering wheel angular velocity-related limits, and road sensor motor availability limits:
[0212] ;
[0213] in, This is expressed as the target feedback torque output to the road feeler motor. Expressed as the maximum permissible torque related to the vehicle's longitudinal speed. This is expressed as the maximum permissible rate of torque change.
[0214] Target feedback torque The road feeler motor torque closed-loop controller converts the torque into motor current command, causing the road feeler motor to apply the corresponding feedback torque to the steering wheel.
[0215] like Figure 3 As shown, a second aspect of this disclosure provides a steer-by-wire road feel generation device, comprising: a signal acquisition module, a driver state recognition module, a reference vehicle response calculation module, a residual calculation module, a time-series attribution module, a response component estimation module, a road feel torque generation module, and a torque superposition module.
[0216] The signal acquisition module is used to acquire vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration.
[0217] The driver status recognition module is used to determine the driver's steering activity based on the steering wheel angle, steering wheel angular velocity, and driver input torque.
[0218] The reference vehicle response calculation module is used to calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and steering wheel angle.
[0219] The residual calculation module is used to calculate the yaw response residual between the actual yaw rate and the desired yaw rate, as well as the lateral response residual between the actual lateral acceleration and the desired lateral acceleration.
[0220] The temporal attribution module is used to determine the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, within the delay search interval related to the vehicle's longitudinal speed. It also combines directional consistency and amplitude consistency to determine the driver's attribution confidence.
[0221] The response component estimation module is used to determine driver-related and non-driver-related responses based on the expected yaw rate, expected lateral acceleration, yaw response residual, lateral response residual, and driver attribution confidence.
[0222] The road feel torque generation module is used to generate normal response road feel torque based on driver-related responses and disturbance warning road feel torque based on non-driver-related responses.
[0223] The torque superposition module is used to superimpose the basic road feel torque, normal response road feel torque, disturbance warning road feel torque, and return road feel torque to obtain the target road feel torque, which is then output to the road feel motor.
[0224] The aforementioned modules can be implemented by software programs within the same controller processor, or they can be implemented collaboratively by multiple controllers. Each controller includes a processor and a memory; the memory stores the program, and the processor executes the program to implement the method of this invention.
[0225] Optionally, the steer-by-wire road feel generation device may further include a stability warning module, a human-machine conflict management module, and a safety constraint module. The stability warning module calculates a stability risk index based on the yaw response residual and the lateral response residual, and generates a stability warning torque based on the stability risk index. The human-machine conflict management module determines a suggested correction direction based on non-driver-related responses. When the steering wheel angular velocity direction is consistent with the suggested correction direction, and the non-driver-related response decreases after driver operation, it determines that the driver is making an effective correction. Furthermore, when it determines that the driver is making an effective correction, and the warning torque creates resistance in the steering wheel movement direction, it attenuates the disturbance warning torque and the stability warning torque to obtain the attenuated warning torque. The safety constraint module is used to limit the amplitude and rate of change of the target road feel torque. When the available capacity of the road feel motor is insufficient, the components are reduced in order of priority: disturbance warning road feel torque, stability warning torque, normal response road feel torque, and basic road feel torque. When the actual yaw rate or actual lateral acceleration fails, the disturbance warning road feel torque and stability warning torque are smoothly decayed to zero according to the preset slope, while the basic road feel torque and the alignment road feel torque are retained.
[0226] A third aspect of this disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the time-attribution-based steer-by-wire road feel generation method of the first aspect described above.
[0227] In some embodiments, the computer-readable storage medium may be a read-only memory (ROM), random access memory (RAM), optical disc (CD-ROM), magnetic tape, floppy disk, and optical data storage device, etc. When the computer program is executed by the processor, it implements all or part of the steps in the time-series attribution-based steer-by-wire road sense generation method of the first aspect described above. Optionally, the computer-readable storage medium may also be flash memory, hard disk, or solid-state drive, and can be designed according to actual needs, without any limitation herein.
[0228] Although embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this disclosure.
[0229] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Any other corresponding changes and modifications made based on the technical concept of this disclosure should be included within the scope of protection of the claims of this disclosure.
Claims
1. A method for generating steer-by-wire road feel based on temporal attribution, characterized in that, Includes the following steps: S1. Obtain vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration; S2. Determine the driver's steering activity based on the steering wheel angle, the steering wheel angular velocity, and the driver's input torque; S3. Calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and the steering wheel angle; S4. Calculate the yaw response residual between the actual yaw rate and the desired yaw rate, and the lateral response residual between the actual lateral acceleration and the desired lateral acceleration; S5. Within the delay search interval related to the vehicle's longitudinal speed, determine the first temporal correlation between the steering wheel angular velocity and the rate of change of yaw angular velocity, and the second temporal correlation between the steering wheel angular velocity and the rate of change of lateral acceleration, and combine directional consistency and amplitude consistency to determine the driver's attribution confidence. S6. Determine driver-related and non-driver-related responses based on the expected yaw rate, the expected lateral acceleration, the yaw response residual, the lateral response residual, and the driver attribution confidence. S7. Generate normal response road feel torque based on the driver-related response, and generate disturbance alert road feel torque based on the non-driver-related response; S8. The base road feel torque, the normal response road feel torque, the disturbance alert road feel torque, and the return road feel torque are superimposed to obtain the target road feel torque, which is then output to the road feel motor.
2. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, S2 includes: The steering wheel angle, the steering wheel angular velocity, and the driver input torque are normalized respectively to obtain normalized values for the steering wheel angle, the steering wheel angular velocity, and the driver input torque. The driver's steering activity is obtained by weighted summing the normalized values of the steering wheel angle, the steering wheel angular velocity, and the driver's input torque.
3. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, S3 includes: The steering ratio is determined based on the vehicle's longitudinal speed, and the steering wheel angle is divided by the steering ratio to obtain the reference front wheel angle. Based on the reference front wheel steering angle and the vehicle longitudinal speed, the steady-state expected yaw rate is calculated using the vehicle reference model. A first-order dynamic element is applied to the steady-state desired yaw rate to obtain the desired yaw rate; The desired lateral acceleration is calculated based on the product of the desired yaw rate and the vehicle's longitudinal speed.
4. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, In step S5, determining the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, includes: The delay search interval is determined based on the vehicle's longitudinal speed; Within the delay search interval, calculate the first cross-correlation value between the steering wheel angular velocity sequence and the yaw rate of change sequence, and the second cross-correlation value between the steering wheel angular velocity sequence and the lateral acceleration rate of change sequence; The first cross-correlation value is used as the first time-series correlation, and the second cross-correlation value is used as the second time-series correlation.
5. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, In step S5, the driver's attribution confidence is determined by combining directional consistency and amplitude consistency, including: Based on the directional relationship between the steering wheel angle, the actual yaw rate, and the actual lateral acceleration within a set time window, directional consistency is determined; The driver's steering activity, the first temporal correlation, the second temporal correlation, the directional consistency, and the amplitude consistency are fused to obtain the driver's attribution confidence. Amplitude consistency is determined based on the normalized relationship between the yaw response residual and the desired yaw angular velocity, and the normalized relationship between the lateral response residual and the desired lateral acceleration.
6. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, S6 includes: The yaw response residual and the lateral response residual are respectively subjected to low-pass filtering to obtain the yaw slow-varying residual and the lateral slow-varying residual; The driver-related yaw response is obtained by adding the expected yaw rate to the product of the driver's attribution confidence and the yaw slow-varying residual. The driver-related lateral response is obtained by adding the desired lateral acceleration to the product of the driver's attribution confidence and the lateral slow-varying residual. Subtracting the driver-related yaw response from the actual yaw rate yields the non-driver-related yaw response. Subtracting the driver-related lateral response from the actual lateral acceleration yields the non-driver-related lateral response; The driver-related response includes the driver-related yaw response and the driver-related lateral response, and the non-driver-related response includes the non-driver-related yaw response and the non-driver-related lateral response.
7. The method for generating steer-by-wire road feel based on time-series attribution according to claim 6, characterized in that, S7 includes: Based on the driver-related yaw response and the driver-related lateral response, the normal response road feel torque is generated after low-frequency mapping and mid-frequency mapping, respectively. The disturbance warning torque is generated after transient shaping based on the non-driver-related yaw response and the non-driver-related lateral response. The transient shaping includes at least one of bandpass filtering, finite-duration pulse generation, peak hold followed by exponential decay, and piecewise function.
8. The method for generating steer-by-wire road feel based on time-series attribution according to claim 7, characterized in that, In step S8, the return-to-center torque is determined based on the steering wheel angle and the vehicle's longitudinal speed.
9. The method for generating steer-by-wire road feel based on time-series attribution according to claim 8, characterized in that, Also includes: The stability risk index is calculated based on the yaw response residual and the lateral response residual; A stability warning torque is generated based on the stability risk index; The target road feel torque is obtained by superimposing the base road feel torque, the normal response road feel torque, the disturbance warning road feel torque, the stability warning torque, and the straightening road feel torque.
10. The method for generating steer-by-wire road feel based on time-series attribution according to claim 9, characterized in that, Also includes: The suggested correction direction is determined based on the aforementioned non-driver-related responses; When the direction of the steering wheel angular velocity is consistent with the suggested correction direction, and the non-driver-related response decreases after the driver's operation, it is determined that the driver is making an effective correction; When it is determined that the driver is making an effective correction and the prompt torque is obstructing the direction of steering wheel movement, the disturbance prompt torque and the stability prompt torque are attenuated to obtain the attenuated prompt torque. The step of superimposing the base road feel torque, the normal response road feel torque, the disturbance warning road feel torque, the stability warning torque, and the homing road feel torque to obtain the target road feel torque includes: When it is determined that the driver is making an effective correction and the prompt torque is obstructing the direction of steering wheel movement, the base road feel torque, the normal response road feel torque, the attenuated prompt torque, and the return-to-center road feel torque are superimposed to form the target road feel torque.
11. The method for generating steer-by-wire road feel based on time-series attribution according to claim 9, characterized in that, Also includes: The target road-feel torque is subject to amplitude and rate of change limits. When the available capacity of the road sense motor is insufficient, the components are reduced in the following order of priority: disturbance warning road sense torque, stability warning torque, normal response road sense torque, and basic road sense torque. When the actual yaw rate or the actual lateral acceleration fails, the disturbance indication torque and the stability indication torque are smoothly attenuated to zero according to a preset slope, while the basic road feel torque and the straightening road feel torque are retained.
12. The method for generating steer-by-wire road feel based on time-series attribution according to claim 1, characterized in that, In step S1, after obtaining the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration, the method further includes: Based on at least one of the timestamp, message counter, and reception time, the vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration are resampled to a unified control cycle; The resampled vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration are all subjected to low-pass filtering. Perform at least one of the following checks on the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration: range check, update timeout check, signal freeze check, mutation rate check, and physical consistency check, to obtain a confidence index for the vehicle's longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate, and actual lateral acceleration.
13. A steer-by-wire road feel generation device, characterized in that, include: The signal acquisition module is used to acquire vehicle longitudinal speed, steering wheel angle, steering wheel angular velocity, driver input torque, actual yaw rate and actual lateral acceleration; The driver status recognition module is used to determine the driver's steering activity based on the steering wheel angle, the steering wheel angular velocity, and the driver's input torque. The reference vehicle response calculation module is used to calculate the desired yaw rate and desired lateral acceleration based on the vehicle's longitudinal speed and the steering wheel angle. The residual calculation module is used to calculate the yaw response residual between the actual yaw rate and the expected yaw rate, and the lateral response residual between the actual lateral acceleration and the expected lateral acceleration; The temporal attribution module is used to determine the first temporal correlation between the steering wheel angular velocity and the yaw rate of change, and the second temporal correlation between the steering wheel angular velocity and the lateral acceleration rate of change, within the delay search interval related to the vehicle's longitudinal speed, and to determine the driver's attribution confidence by combining directional consistency and amplitude consistency. The response component estimation module is used to determine driver-related and non-driver-related responses based on the expected yaw rate, the expected lateral acceleration, the yaw response residual, the lateral response residual, and the driver attribution confidence. The road feel torque generation module is used to generate a normal response road feel torque based on the driver-related response and to generate a disturbance warning road feel torque based on the non-driver-related response. The torque superposition module is used to superimpose the basic road feel torque, the normal response road feel torque, the disturbance indication road feel torque, and the return road feel torque to obtain the target road feel torque, and output it to the road feel motor.
14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steer-by-wire road feel generation method based on time-series attribution as described in any one of claims 1 to 12.
