Cooperative control method, device and equipment of distributed steering system, and storage medium

Abstract

The application relates to a cooperative control method and device of a distributed steering system, equipment and a storage medium. When a vehicle is in a preset safe working condition, a plurality of to-be-tested wheel end angle modules are sequentially subjected to limited perturbation excitation signals, and response signals of each to-be-tested wheel end angle module are collected, then the dynamic response characteristic parameters corresponding to each to-be-tested wheel end angle module are identified online, and the feedforward compensation coefficients corresponding to each to-be-tested wheel end angle module are calculated, the feedforward compensation coefficients corresponding to each to-be-tested wheel end angle module are injected into the control target of the corresponding to-be-tested wheel end angle module, so as to compensate the dynamic response difference between the wheel end angle modules. The dynamic differences of the wheel ends are actively perceived without interfering with driving, the gain and delay are quantified online, the individualized feedforward compensation coefficients are calculated and injected into the control target, and finally the response time dispersion of the multiple wheels is significantly reduced, and the high-speed dynamic consistency and cooperative control precision of the distributed steering system are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle chassis control technology, specifically to a collaborative control method, device, equipment, and storage medium for a distributed steering system. Background Technology

[0002] Distributed wheel-end angle module steering systems integrate steering, driving, braking, and suspension functions into independent modules at each wheel end. A central control unit coordinates and distributes the target steering angle or angular velocity to each wheel end, enabling various functions such as yaw control, low-speed large-angle maneuvering, crabbing, stationary steering, and high-dynamic obstacle avoidance. Compared to traditional mechanical or hydraulic steering systems, this system offers advantages such as compact structure, flexible control, and ease of integrated vehicle dynamics control, thus gradually becoming an important development direction in the field of drive-by-wire chassis.

[0003] In existing technologies, the control methods of distributed steering systems mainly rely on the upper-level control model to assign target steering angle or angular velocity commands to each wheel-end module, and assume that each module has similar or consistent dynamic response characteristics. However, in actual engineering applications, due to factors such as manufacturing tolerances, transmission chain clearances, differences in friction characteristics, changes in ambient temperature, and power supply fluctuations, the "control input-output response" relationship of each wheel-end module varies significantly.

[0004] Therefore, the lack of coordinated control due to the inconsistent dynamic characteristics of multi-wheel end-angle modules in existing distributed steering systems is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] One of the objectives of this invention is to provide a collaborative control method, apparatus, device, and storage medium for a distributed steering system, in order to solve the problem of collaborative control in existing distributed steering systems caused by inconsistent dynamic characteristics of multi-wheel end-angle modules.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A cooperative control method for a distributed steering system includes:

[0008] When the vehicle is in a preset safe operating condition, a limited perturbation excitation signal is sequentially applied to multiple wheel end angle modules to be tested, and the response signal of each wheel end angle module to be tested is collected.

[0009] For each wheel end angle module under test, based on the perturbation excitation signal and the response signal of the wheel end angle module under test, the dynamic response characteristic parameters corresponding to the wheel end angle module under test are identified online. The dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules.

[0010] The feedforward compensation coefficient corresponding to the wheel end angle module under test is calculated based on the dynamic response characteristic parameters corresponding to the wheel end angle module under test.

[0011] The feedforward compensation coefficient corresponding to each wheel end angle module under test is injected into the control target of the corresponding wheel end angle module under test in order to compensate for the dynamic response differences between the wheel end angle modules.

[0012] Based on the above technical means, without interfering with normal driving, the dynamic response differences of each wheel end are actively identified online and corresponding feedforward compensation coefficients are generated. By injecting the compensation coefficients, the response time dispersion of multiple wheels to the same steering command is significantly reduced, thereby improving the dynamic consistency and collaborative control accuracy of the distributed steering system.

[0013] Furthermore, the dynamic response characteristic parameters include equivalent dynamic gain and equivalent delay. Therefore, the step of identifying the dynamic response characteristic parameters corresponding to the wheel end angle module under test online based on the perturbation excitation signal and the response signal of the module under test includes:

[0014] For each wheel end angle module to be tested, feature extraction is performed on the response signal of the wheel end angle module to obtain response feature quantities;

[0015] The test wheel end angle module is modeled and fitted with the dynamic characteristics of the wheel end based on the response feature quantity, and the equivalent dynamic gain and equivalent delay corresponding to the test wheel end angle module are calculated.

[0016] Based on the above technical means, the dynamic response characteristic parameters are specified as equivalent dynamic gain and equivalent delay, so that the wheel end response capability and speed can obtain clear physical quantitative indicators, which facilitates the calculation of compensation coefficient and subsequent performance evaluation.

[0017] Furthermore, the step of calculating the feedforward compensation coefficient corresponding to the wheel end angle module based on the dynamic response characteristic parameters corresponding to the wheel end angle module includes:

[0018] A multi-round synchronization target is constructed based on the dynamic response characteristic parameters corresponding to each wheel end angle module under test. The multi-round synchronization target includes reference gain and reference delay.

[0019] For each wheel end angle module to be tested, the feedforward compensation coefficient corresponding to the wheel end angle module to be tested is calculated based on the deviation between the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested and the multi-wheel synchronization target.

[0020] Based on the above technical means, the feedforward compensation coefficient is calculated based on the deviation between the multi-wheel synchronization target (reference gain and reference delay) and each wheel end, so as to ensure that the compensation of each wheel end is coordinated with each other and effectively improve the accuracy of multi-wheel synchronization control.

[0021] Furthermore, the method also includes:

[0022] For each wheel end angle module to be tested, the reliability of the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested is verified, and the parameters that pass the verification are stored in the valid parameter set corresponding to the wheel end angle module to be tested.

[0023] Accordingly, the step of calculating the corresponding feedforward compensation coefficient for each wheel end angle module based on the dynamic response characteristic parameters of each module under test includes:

[0024] Based on the effective parameter set corresponding to each wheel end angle module under test, the corresponding feedforward compensation coefficient is calculated for each wheel end angle module under test.

[0025] Based on the above technical means, the reliability of the identified dynamic response characteristic parameters is verified. Only the parameters that pass the verification are stored in the valid parameter set and used for compensation coefficient calculation, which avoids abnormal data from polluting the control system and enhances the robustness and reliability of collaborative control.

[0026] Furthermore, the credibility verification of the dynamic response characteristic parameters corresponding to the wheel end angle module under test includes:

[0027] For each wheel end angle module to be tested, the fitting residual during the online identification process of the wheel end angle module to be tested is obtained, and it is determined whether the fitting residual exceeds a preset residual threshold.

[0028] Determine whether the change range of the dynamic response feature parameter corresponding to the wheel end angle module under test and the dynamic response feature parameter obtained from the previous online identification of the wheel end angle module under test exceeds a preset amplitude threshold.

[0029] Determine whether the vehicle is always within a preset safe operating condition during the online identification process of the wheel end angle module under test;

[0030] If all the above checks pass, the credibility check is considered successful.

[0031] Based on the above technical means, through triple verification of residuals, consistency, and environmental validity, invalid identification results caused by fitting errors, parameter mutations, or breaches of safety conditions are completely eliminated, ensuring the accuracy and stability of online identification parameters.

[0032] Furthermore, the method also includes:

[0033] For each wheel end angle module to be tested, the parameters are corrected and standardized based on the historical dynamic response characteristic parameters corresponding to the wheel end angle module and the operating environment information, so as to obtain the processed parameters.

[0034] Calculate the average offset, short-term fluctuation, and long-term drift trend corresponding to the processed parameters;

[0035] Based on the abnormal event records corresponding to the wheel end angle module under test acquired in real time, the number of abnormal events is counted, the length of continuous abnormal events is extracted, and the duration of abnormal events is encoded to obtain abnormal degradation items.

[0036] A health index is formed by fusing the average offset, the short-term fluctuation, the long-term drift trend, and the abnormal degradation items.

[0037] Based on the health index and the duration of the health index continuously falling within any of the preset multiple health levels, the wheel end status is classified into levels, and status records, maintenance prompts, function restriction instructions, fault alarm signals, or degradation control requests are output respectively.

[0038] Based on the above technical means, the average offset, short-term fluctuation and long-term drift are calculated after correction and standardization of historical parameters, and combined with abnormal event records to form a health index. According to the index size and duration, the status record, maintenance prompts, function restrictions, fault alarms or degradation control requests are output in a graded manner, realizing early warning and predictive maintenance of slow degradation of actuators.

[0039] Furthermore, the method also includes:

[0040] Receive the real-time status signal of the vehicle;

[0041] The vehicle is determined to be in a preset safe operating condition based on the real-time status signal and the preset multi-dimensional safety threshold.

[0042] Based on the above technical means, by receiving vehicle status signals in real time and comparing them with multi-dimensional safety thresholds, it is determined whether a preset safety condition has been entered, providing safety gating conditions for identification and triggering, and avoiding interference with driving and damage to vehicle stability from the source.

[0043] Furthermore, the real-time status signal includes vehicle speed, vehicle steering wheel angle, steering wheel angular velocity, steering wheel torque, yaw rate, and lateral acceleration. Therefore, determining whether the vehicle is in a preset safe operating condition based on the real-time status signal and a pre-set multi-dimensional safety threshold includes:

[0044] The vehicle is determined to be in a preset safe operating condition when all of the following conditions are met:

[0045] The absolute value of the vehicle steering wheel angle is less than a first threshold, the absolute value of the steering wheel angular velocity is less than a second threshold, the absolute value of the steering wheel torque is less than a third threshold, the absolute value of the vehicle lateral acceleration is less than a fourth threshold, the absolute value of the yaw rate is less than a fifth threshold, the vehicle speed is between a preset minimum speed and a maximum speed, and all of the above conditions are maintained continuously for a preset duration threshold.

[0046] Based on the above technical means, multiple parameters such as steering wheel angle, angular velocity, torque, lateral acceleration, yaw rate, and vehicle speed are determined in parallel, and the duration accumulation condition is added to ensure the accuracy and robustness of the safety window trigger and prevent false triggering.

[0047] Furthermore, the perturbation excitation signal is one or more combinations of positive and negative symmetrical double pulse signal, single pulse signal, pseudo-random sequence signal or small-amplitude sweep frequency signal.

[0048] Based on the above technical means, the perturbation excitation signal is limited to positive and negative symmetrical double pulse, single pulse, pseudo-random sequence or small-amplitude frequency sweep signal, providing flexible and diverse excitation forms, ensuring full excitation of wheel end dynamic response while achieving low perceptibility, and adapting to different engineering needs.

[0049] A cooperative control device for a distributed steering system, comprising:

[0050] An acquisition module is used to apply restricted perturbation excitation signals to multiple wheel end angle modules under test when the vehicle is in a preset safe operating condition, and to acquire the response signal of each wheel end angle module under test.

[0051] The online identification module is used to identify the dynamic response characteristic parameters of each wheel end angle module under test based on the perturbation excitation signal and the response signal of the wheel end angle module under test. The dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules.

[0052] The calculation module is used to calculate the feedforward compensation coefficient corresponding to the wheel end angle module under test based on the dynamic response characteristic parameters corresponding to the wheel end angle module under test;

[0053] The injection module is used to inject the feedforward compensation coefficient corresponding to each wheel end angle module under test into the control target of the corresponding wheel end angle module under test, so as to compensate for the dynamic response differences between the wheel end angle modules.

[0054] An automobile includes: a vehicle body, a storage unit disposed in the vehicle body, and an electronic control unit;

[0055] The storage unit stores computer-executed instructions;

[0056] The electronic control unit executes the computer execution instructions stored in the storage unit to implement the cooperative control method of the distributed steering system described above.

[0057] A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement a cooperative control method for a distributed steering system as described in any of the preceding claims.

[0058] A computer program product includes a computer program that, when executed by a processor, implements the cooperative control method for the distributed steering system described in any of the preceding claims.

[0059] The beneficial effects of this invention are:

[0060] (1) By introducing a safety window-based perturbation excitation injection, online parameter identification and multiple reliable verification mechanisms, adaptive measurement and compensation for the slow-changing characteristics of the actuator caused by manufacturing tolerances, wear, temperature rise and aging are achieved without interfering with the driver's normal operation.

[0061] (2) By establishing a collaborative compensation algorithm for multi-wheel synchronization targets, and calculating the feedforward compensation coefficient based on the equivalent dynamic gain and equivalent delay obtained by online identification, the response time dispersion of multiple wheel ends to the same steering command is reduced, thereby improving the dynamic consistency, handling stability and collaborative accuracy of the distributed steering system.

[0062] (3) Combining online identification results with historical statistics to form an interpretable health index is used for actuator performance degradation early warning and fault classification processing, realizing the leap from passively suppressing differences to actively sensing health status, and enhancing the reliability and maintainability of the system. Attached Figure Description

[0063] Figure 1 This is the application system architecture diagram for this solution;

[0064] Figure 2 Example 1 of the cooperative control method for the distributed steering system provided in this solution;

[0065] Figure 3 Example 2 of the cooperative control method for the distributed steering system provided in this solution;

[0066] Figure 4 Example 3 of the cooperative control method for the distributed steering system provided in this solution;

[0067] Figure 5 Example 4 of the cooperative control method for the distributed steering system provided in this solution;

[0068] Figure 6 Example 5 of the cooperative control method for the distributed steering system provided in this solution;

[0069] Figure 7 Example 6 of the cooperative control method for the distributed steering system provided in this solution;

[0070] Figure 8 A schematic diagram of the collaborative control device for the distributed steering system provided in this solution;

[0071] Figure 9 A schematic diagram of the vehicle structure provided for this solution. Detailed Implementation

[0072] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0073] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0074] Figure 1 This is the application system architecture diagram for this solution, such as... Figure 1 As shown, the system includes: a security identification and judgment module, a perturbation excitation and sampling module, a feature extraction and parameter identification module, a credibility verification and collaborative compensation module, and a health index and fault early warning module.

[0075] First, the safety identification and judgment module continuously receives real-time status signals such as vehicle steering wheel angle, torque, yaw rate, and longitudinal / lateral acceleration, and performs parallel comparisons and logical AND operations with a set of pre-calibrated multi-dimensional safety thresholds. Only when all status quantities meet the safety conditions (such as no obvious steering intention from the driver and stable vehicle dynamics) and remain so for a preset duration, does this module determine that the system has entered the "allow online identification" safety window. If the safety conditions are not met, all subsequent modules will not be activated.

[0076] Once the safety window is open, the system initiates the sequence identification protocol. The perturbation excitation and sampling module injects a current pulse excitation signal into the target wheel end angle module, while simultaneously acquiring the response information of the wheel end angle module at high speed. Subsequently, the feature extraction module processes the acquired excitation-response data pairs, extracting key time-domain or frequency-domain features characterizing the dynamic response through methods such as linear fitting.

[0077] Based on the extracted features, the parameter identification module uses a lightweight estimation algorithm (such as recursive least squares) to calculate the current dynamic gain value of the wheel-end module in real time. This gain value is then sent to the reliability verification module, undergoing triple verification based on residual verification, consistency checks between old and new results, and safety judgment signal verification. Only the effective gain that passes all verifications is adopted by the collaborative compensation module to calculate the feedforward compensation coefficient for that module in real time and smoothly apply it to the unified steering command issued by the central authority. In addition, the effective gain and its historical sequence are synchronously injected into the health index and fault warning module. This module performs hierarchical diagnosis by analyzing the long-term trend, short-term fluctuations, and consistency between modules of the gain: identifying performance degradation trends to trigger predictive maintenance warnings, detecting severe mismatches to provide functional alerts, and identifying anomalies to initiate fault-tolerant control.

[0078] Figure 2 Example 1 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 2 As shown, the method includes:

[0079] S201: When the vehicle is in a preset safe operating condition, a limited perturbation excitation signal is applied sequentially to multiple wheel end angle modules to be tested, and the response signal of each wheel end angle module to be tested is collected.

[0080] This step, without interfering with the driver's normal operation or compromising vehicle stability, actively acquires the response data of each wheel-end angle module to a known excitation signal, providing high-quality input-output samples for subsequent online identification of dynamic characteristic differences. Simultaneously, by applying excitations sequentially, it avoids mutual interference or excessive vehicle disturbance caused by simultaneous excitation of multiple wheel ends.

[0081] Specifically, the vehicle is first positioned under a preset safe operating condition. Each wheel end angle module is then processed sequentially according to a preset order (e.g., front left → front right → rear left → rear right). A low-disturbance, low-perceptibility, but sufficiently identifiable test stimulus is applied to the specified wheel end angle module, and the dynamic response data of that wheel end is simultaneously acquired. This serves to provide observable, calculable, and comparable input-output samples for parameter identification, while minimizing the impact on vehicle handling and multi-wheel coordination. The perturbation excitation signal is one or a combination of positive and negative symmetrical double-pulse signals, single-pulse signals, pseudo-random sequence signals, or small-amplitude frequency sweep signals.

[0082] According to the pre-calibrated excitation strategy, a set of amplitude-limited and duration-limited perturbation excitation signals are superimposed on the current target channel at the selected wheel end. A positive-negative symmetrical double-pulse form is preferred, but it can also be extended to a single pulse, pseudo-random micro-sequence, or small-amplitude frequency sweep excitation as needed. The specific formula for generating the perturbation excitation signal is as follows:

[0083] in, This is the excitation signal corresponding to the i-th wheel end angle module under test. To excite the amplitude, The pulse width is The unit rectangular impulse function, To incentivize the starting moment, The pulse width. The interval is a double pulse, and i is the module number of the wheel end angle to be measured. The excitation parameters can be adaptively adjusted based on the current vehicle speed, wheel end temperature, historical identification stability, and current wheel end workload, thereby reducing additional disturbances while ensuring identification visibility. After the excitation is applied, signals such as wheel end drive current, actual wheel end angular velocity, actual wheel end rotation angle, and temperature are synchronously acquired within a preset observation window. The sampled data are buffered in chronological order to form a complete response sequence for the current identification cycle. The sampling observation window formula is:

[0084] in, To identify the observation window length, if the safety identification conditions fail during sampling, subsequent sampling is immediately stopped and the current period's data is discarded; if sampling is complete and the data is intact, it is output. This "directional perturbation, synchronous sampling, and abnormal termination" approach establishes a direct correspondence between test inputs and the actuator's dynamic output, providing a reliable data foundation for subsequent feature construction and parameter identification, while ensuring that the identification process itself does not significantly impact driving safety or vehicle dynamic consistency.

[0085] S202: For each wheel end angle module under test, the dynamic response characteristic parameters corresponding to the wheel end angle module under test are identified online based on the perturbation excitation signal and the response signal of the wheel end angle module under test.

[0086] This step extracts key parameters (especially equivalent dynamic gain and equivalent delay) from the collected excitation-response data that can quantify the "response speed" and "response intensity" of the wheel end, providing a unified and comparable basis for subsequent calculation of difference compensation.

[0087] Specifically, for each wheel end angle module under test, the response signal of the wheel end angle module under test is feature extracted to obtain the response feature quantity. Based on the response feature quantity, the wheel end angle module under test is modeled and the wheel end dynamic characteristics are fitted. The equivalent dynamic gain and equivalent delay corresponding to the wheel end angle module under test are calculated.

[0088] It should be noted that the dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules, and mainly include wheel end angular velocity, wheel end rotation angle and drive current signal.

[0089] S203: Calculate the feedforward compensation coefficient corresponding to the wheel end angle module under test based on the dynamic response characteristic parameters corresponding to the wheel end angle module under test.

[0090] This step utilizes the differences in dynamic parameters of multiple wheel ends to generate a personalized feedforward compensation amount for each wheel end, enabling each wheel end to achieve a near-synchronous response when receiving the same target command, thereby reducing response time dispersion.

[0091] Specifically, a multi-wheel synchronization target is constructed based on the dynamic response characteristic parameters corresponding to each wheel end angle module under test. For each wheel end angle module under test, the feedforward compensation coefficient corresponding to the wheel end angle module under test is calculated based on the deviation between the dynamic response characteristic parameters corresponding to the wheel end angle module under test and the multi-wheel synchronization target.

[0092] S204: Inject the feedforward compensation coefficient corresponding to each wheel end angle module under test into the control target of the corresponding wheel end angle module under test, so as to compensate for the dynamic response differences between the wheel end angle modules.

[0093] This step applies the calculated feedforward compensation coefficient to the control command of each wheel end angle module under test, so that when each wheel end receives the same upper-level target, the actual current or angle command executed is differentiated and corrected, ultimately compensating for the asynchronous response caused by the difference in gain and delay.

[0094] Specifically, first, the type of control interface is determined. If the upper-level controller outputs the target angular velocity, the compensation coefficient is directly multiplied into the original target angular velocity to form the corrected wheel-end speed command. If the upper-level controller outputs the target steering angle, an equivalent compensation amount is constructed based on the deviation between the target steering angle and the actual steering angle, and injected into the target steering angle channel. Considering that the online identification parameters may change in different control cycles, to avoid sudden switching of the compensation coefficients causing new dynamic shocks, the old and new compensation coefficients are smoothly merged to make the compensation update process continuous and gradual. The compensation smooth update formula is:

[0095] in, This refers to the compensation coefficient actually used in the current cycle. As a smoothing factor, This is the compensation coefficient for the previous period. These are the newly calculated compensation coefficients for this round. After compensation is completed, the corrected control objective is sent to the wheel-end local controller for execution.

[0096] It should be noted that after the compensation takes effect, the actual response continues to be monitored. If there are new identification results, the compensation coefficient is recalculated to form a closed-loop adaptive system.

[0097] The cooperative control method for a distributed steering system provided in this application applies restricted perturbation excitation signals sequentially to multiple wheel-end angle modules under test when the vehicle is in a preset safe operating condition. The response signal of each wheel-end angle module is collected. For each wheel-end angle module, based on the perturbation excitation signal and the response signal, the dynamic response characteristic parameters corresponding to the wheel-end angle module are identified online. Based on the dynamic response characteristic parameters, the feedforward compensation coefficient corresponding to the wheel-end angle module is calculated. This feedforward compensation coefficient is then injected into the control target of the corresponding wheel-end angle module to compensate for the dynamic response differences between the modules. This method achieves proactive perception of dynamic differences between wheel-end angles without interfering with driving, online quantification of gain and delay, calculation of personalized feedforward compensation coefficients, and injection into the control target. Ultimately, it significantly reduces the multi-wheel response time dispersion and improves the high-speed dynamic consistency and cooperative control accuracy of the distributed steering system.

[0098] Figure 3 Example 2 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 3 As shown, the dynamic response characteristic parameters include equivalent dynamic gain and equivalent delay. Therefore, based on the above embodiment, step S202 specifically includes:

[0099] S301: For each wheel end angle module to be tested, feature extraction is performed on the response signal of the wheel end angle module to be tested to obtain the response feature quantity.

[0100] This step converts the original time-domain signals such as current, angular velocity, and rotation angle into key quantitative indicators that can characterize the dynamic response of the wheel end (such as response delay, rise slope, peak value, etc.), providing compact and physically meaningful input data for subsequent modeling and parameter fitting, while suppressing the effects of measurement noise and baseline drift.

[0101] Specifically, key feature quantities that can characterize the dynamic characteristics of the wheel end are extracted from the response signal of the wheel end angle module under test. Its role is to transform the original measurement signal into a more physically meaningful and compact description that is easier to identify and verify later, thereby improving the robustness and reliability of parameter identification.

[0102] First, the acquired wheel end angular velocity, wheel end rotation angle, and drive current signals are filtered and smoothed to suppress measurement noise and high-frequency interference, ensuring that subsequent slope, peak value, and delay extraction are not affected by instantaneous jitter. The original signal filtering formula is:

[0103] in This represents the original discrete sampled wheel-end angular velocity, current, and angle signals. This represents the filtered wheel end angular velocity, current, and angle signals. This represents the filter coefficients.

[0104] Subsequently, using a steady-state window before the excitation was applied as the baseline, baseline elimination processing was performed on the response signal within the current observation window to eliminate the static bias and obtain the incremental signal that only reflects the changes caused by the perturbation excitation. The specific formula for baseline elimination is as follows:

[0105] in, The number of samples used to calculate the mean before the stimulus. To provide the discrete time corresponding to the start of the excitation. The baseline mean values ​​of wheel end angular velocity, current, and angle signals.

[0106] The specific formula for the incremental signal is:

[0107] in, This is the incremental response after baseline removal. After baseline removal, the response start time is further identified within the current observation window to determine the time point when the wheel end transitions from rest or steady state to dynamic response. The formula for extracting the response start time is:

[0108] in, Let i be the start time of the response at the i-th wheel end. To meet the detection threshold, the rise slope, peak increment, and earliest moment when the set threshold is reached of the angular velocity response are further extracted to quantify the speed and intensity of the wheel-end response establishment. The formula for the rise slope is:

[0109]

[0110] in, For discrete slope, For the maximum upward slope, The sampling period is [value]. The formula for extracting the peak increment is:

[0111]

[0112] in, This represents the peak increment of angular velocity. This represents the peak increment at the turning point. The formula for extracting the threshold arrival time is:

[0113] in, Let i be the earliest time when the i-th wheel end reaches the threshold angular velocity. This is the actual angular velocity. For reference angular velocity, This is the threshold scaling factor. After quantifying the speed of wheel-end response buildup, the equivalent delay factor characterizing the difference in wheel-end response buildup speed is calculated, and its formula is:

[0114] in, This represents the equivalent delay of the i-th wheel end relative to the excitation start point. Simultaneously, the angular velocity signal can be integrated or cross-compared with the rotation angle signal to determine if the angular velocity establishment and rotation angle output are consistent, thereby aiding in the identification of transmission backlash, friction thresholds, or local nonlinearities. After the above processing, features such as response delay, rise slope, threshold arrival time, peak increment, and fitting residual correlation are organized into the response feature quantities of the current wheel end and output. This implementation chain of "filtering—baseline removal—start point detection—slope and peak extraction—threshold time extraction—feature encapsulation" transforms complex raw time-series signals into high-value dynamic features, serving as a crucial bridge connecting the raw sampling data and parameter identification results.

[0115] S302: Based on the response feature quantity, model the wheel end angle module under test and fit the dynamic characteristics of the wheel end, and calculate the equivalent dynamic gain and equivalent delay corresponding to the wheel end angle module under test.

[0116] This step utilizes the features extracted in the previous step to quantitatively estimate the equivalent dynamic gain and equivalent delay of the wheel end by establishing a simplified dynamic model of the wheel end angle module. These two parameters represent the wheel end's "ability" and "response speed" in converting current into angular velocity, respectively, and are the core quantitative basis for subsequent collaborative compensation.

[0117] Specifically, the wheel end angle module is approximated as a first-order equivalent dynamic object under small disturbance conditions to describe the dynamic mapping relationship between the drive current and the angular velocity response. The specific formula is as follows:

[0118] in, The equivalent response time constant, Angular velocity, For equivalent dynamic response gain, The driving current is used. Subsequently, an identification sample set is constructed based on the input and output increment samples within the current observation window. The equivalent dynamic gain of the wheel end is calculated using least squares or equivalent lightweight fitting methods, thus characterizing the wheel end's ability to convert input excitation into angular velocity response. The formula for the input-output increment relationship is:

[0119] in, For the m-th output increment sample, For the m-th input incremental sample, The least squares estimation gain formula is used to fit the residuals.

[0120] Where M is the number of valid samples, The equivalent dynamic response gain is estimated. Simultaneously, combining the threshold arrival time and excitation start time obtained during the feature extraction stage, the equivalent delayed response time at the wheel end is calculated to reflect the speed of its response build-up.

[0121] Optionally, after identification, the current parameter results and fitting residuals are saved together for subsequent determination of whether the current parameters can be written into the valid parameter set. Through this implementation method of "based on a small perturbation dynamic model and with lightweight parameter fitting as the core", the online quantification of dynamic differences at each wheel end is realized, which is the core data generation unit of the entire collaborative compensation link.

[0122] The cooperative control method for a distributed steering system provided in this application extracts features from the response signal of each wheel-end angle module under test, obtaining response feature quantities. Based on these response feature quantities, the wheel-end angle module is modeled and fitted with wheel-end dynamic characteristics, calculating the equivalent dynamic gain and equivalent delay corresponding to the wheel-end angle module. This method achieves online quantification of the dynamic characteristics of each wheel-end angle module, transforming response differences that are difficult to compare directly into gain and delay parameters with clear physical meaning, laying a data foundation for subsequent cooperative compensation and health management.

[0123] Figure 4 Example 3 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 4 As shown, based on the above embodiments, step S203 specifically includes:

[0124] S401: Construct a multi-wheel synchronization target based on the dynamic response characteristic parameters corresponding to each wheel end angle module under test.

[0125] After obtaining the equivalent dynamic gain and equivalent delay of each wheel-end angle module, a unified reference benchmark needs to be established so that all wheel ends align with this benchmark. This benchmark is the "multi-wheel synchronization target." The core function of constructing the synchronization target is to provide a common ideal response benchmark for each wheel end. Then, based on the deviation of each wheel end's actual parameters from this benchmark, their respective compensation coefficients are calculated, thereby achieving consistent convergence of multi-wheel response speed and intensity. The multi-wheel synchronization target includes a reference gain and a reference delay.

[0126] Specifically, firstly, the dynamic response characteristic parameters corresponding to each wheel end angle module under test are read, including the dynamic gain and equivalent delay of each wheel end. Then, the reference gain and reference delay are calculated across multiple wheels to establish a vehicle-wide synchronization target. The formula for the reference synchronization parameters is:

[0127]

[0128] in, For the whole vehicle reference gain, The reference delay is for the entire vehicle, and N is the number of wheel ends participating in the coordinated control.

[0129] S402: For each wheel end angle module under test, the feedforward compensation coefficient corresponding to the wheel end angle module under test is calculated based on the deviation between the dynamic response characteristic parameters of the wheel end angle module under test and the multi-wheel synchronization target.

[0130] This step, for each wheel end, uses the deviation between its actual dynamic parameters and the synchronization target (gain too small / too large, delay too long / too short) to quantitatively calculate the feedforward compensation coefficient that needs to be applied to the control target of that wheel end. This coefficient will be directly multiplied or superimposed into the upper-level command, so that the actual response of each wheel end after compensation approaches the synchronization target, thereby reducing the response time dispersion between multiple wheels.

[0131] Specifically, based on the deviation of each wheel-end parameter from the overall vehicle reference value, feedforward compensation coefficients are constructed. Compensation is appropriately increased for wheel-ends with smaller gains and larger delays, and appropriately decreased for wheel-ends with faster responses. This ensures that multiple wheel-ends do not individually pursue local maxima, but rather converge towards a unified synchronization establishment time. The synchronization error objective is designed based on this, and the compensation objective is to reduce this error, achieving multi-wheel compensation and coordinated synchronization. The formula for the synchronization error objective function is:

[0132] in, The prediction setup time for the i-th round end, For reference, the establishment time, The target is the synchronization error. The formula for the feedforward compensation coefficient is:

[0133] in, Let be the feedforward compensation coefficient for the i-th wheel end. For the amplitude limiting function, This is the weight for delay compensation.

[0134] The cooperative control method for a distributed steering system provided in this application constructs a multi-wheel synchronization target based on the dynamic response characteristic parameters corresponding to each wheel-end angle module under test. For each wheel-end angle module under test, the feedforward compensation coefficient corresponding to the wheel-end angle module under test is calculated based on the deviation between the dynamic response characteristic parameters corresponding to the wheel-end angle module under test and the multi-wheel synchronization target. The above method quantizes the discrete dynamic parameters of each wheel-end angle module into a unified synchronization target and generates personalized feedforward compensation coefficients based on the deviation. Compared with traditional feedback control (such as sliding mode control, which can only correct after the error occurs), feedforward compensation can actively cancel the difference before the command is issued, thereby significantly improving the synchronization performance of multi-wheel response, and is especially suitable for yaw stability control under high dynamic conditions (such as emergency obstacle avoidance and high-speed lane change).

[0135] Figure 5 Example 4 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 5 As shown, based on the foregoing embodiments, the method further includes:

[0136] S51: For each wheel end angle module to be tested, the reliability of the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested is verified, and the parameters that pass the verification are stored in the valid parameter set corresponding to the wheel end angle module to be tested.

[0137] During online identification, the dynamic response characteristic parameters obtained from a single identification may be unreliable due to factors such as sensor noise, instantaneous disturbances, and sudden changes in operating conditions. Directly using unreliable parameters to calculate the feedforward compensation coefficients would actually worsen the multi-wheel synchronization performance. This step performs multiple independent verifications on the identification results of each wheel end, storing only the parameters that pass the verification into the valid parameter set for that wheel end, thereby ensuring the high reliability of the parameters used for subsequent collaborative compensation.

[0138] Specifically, this includes:

[0139] S511: For each wheel end angle module to be tested, obtain the fitting residual during the online identification process of the wheel end angle module to be tested, and determine whether the fitting residual exceeds the preset residual threshold.

[0140] S512: Determine whether the change amplitude of the dynamic response characteristic parameters corresponding to the wheel end angle module under test exceeds the preset amplitude threshold compared with the dynamic response characteristic parameters obtained from the previous online identification of the wheel end angle module under test.

[0141] S513: Determine whether the vehicle is always within the preset safe operating conditions during the online identification process of the wheel end angle module under test.

[0142] S514: If all the above checks pass, the credibility check is considered successful.

[0143] For each wheel end angle module under test, residual verification (S511), consistency verification (S512), and environmental validity verification (S513) are performed sequentially. Only when all three verifications pass (S514) are the parameters obtained in this identification stored in the valid parameter set of that wheel end; otherwise, the result is discarded, and the valid parameter set remains unchanged (the value from the previous cycle is continued). This AND logic ensures the strictness of parameter updates.

[0144] When identifying the gain using fitting algorithms such as the least squares method, a fitting residual (i.e., the difference between the measured angular velocity and the model-estimated angular velocity) is obtained. An excessively large residual indicates a poor match between the excitation-response data and the adopted first-order dynamic model, possibly caused by nonlinear friction, intermittent disturbances, or data quality issues. In this case, the identification result should be rejected.

[0145] Specifically, first, obtain information such as the current round's dynamic gain, equivalent delay, and fitting residuals, and calculate the residual index for this round of identification to determine whether the current input-output fitting quality meets the requirements. If the residual is too large, it indicates a poor match between the identified sample and the model, and the result of this round will be directly determined to be unreliable. The average residual verification formula is:

[0146] in, The mean residual index is M, where M is the number of valid samples. This is the fitting residual.

[0147] The dynamic response characteristic parameters (gain, delay) of the wheel end angle module are typically slowly time-varying, drifting gradually over hours or even days due to wear or temperature changes. If the identification result shows a significant jump compared to the valid parameters of the previous cycle (e.g., gain abruptly changes from 0.9 to 0.6), it is highly likely due to transient disturbances or identification errors and should not be adopted. Consistency checks prevent occasional anomalies from contaminating the valid parameter set.

[0148] Specifically, the current identification result is compared with the previously confirmed valid historical parameters to construct a consistency evaluation metric, which is used to determine whether there are any abnormal abrupt changes between the old and new results. If the change exceeds a preset threshold, it indicates that the current result may be affected by occasional disturbances and will not be included in the valid parameter set. The consistency verification formula is:

[0149] in, The current identification value, The valid identification value from the previous round, To prevent the gain coefficient from being too small in the denominator.

[0150] The online identification process must be conducted within preset safe operating conditions (e.g., no driver intention to steer, vehicle in a steady state). However, in actual identification, the driver may suddenly turn the steering wheel, or road impacts may cause the vehicle to yaw. If the collected data is used for identification under these circumstances, the results will be unreliable and may interfere with driving safety. This verification ensures that the safety conditions are met throughout the entire identification window; otherwise, it is considered invalid.

[0151] Specifically, the system combines safety judgment signals to determine whether the current identification occurs within the permissible comparison range; if it is in an unsafe state, the current parameter update is also suspended. The current dynamic parameters are written into the valid parameter set only if the residual check, consistency check, and safety judgment signal check all pass simultaneously; otherwise, the original valid parameters remain unchanged. This "triple verification, optimal writing, and failure retention" approach ensures that the system's online update process is not contaminated by a single anomaly, which is a crucial step in maintaining long-term control stability and the reliability of health assessments.

[0152] Accordingly, based on the dynamic response characteristic parameters of each wheel end angle module under test, the corresponding feedforward compensation coefficients are calculated for each wheel end angle module under test, including:

[0153] Based on the effective parameter set corresponding to each wheel end angle module under test, the corresponding feedforward compensation coefficient is calculated for each wheel end angle module under test.

[0154] This step is implemented in the same way as step S203 in the previous embodiment, except that the data in the effective parameter set has undergone residual verification, consistency verification, and environmental validity verification, which can truly reflect the current dynamic characteristics of the wheel end and avoid errors in the compensation coefficient due to a single abnormal identification. When an identification fails the verification, the effective parameter set remains unchanged (using the effective value of the previous wheel), thereby maintaining the continuity of the compensation coefficient and preventing torque shocks caused by sudden changes. The effective parameter set is gradually updated as the vehicle slowly degrades, enabling the feedforward compensation coefficient to adaptively track the time-varying characteristics of the actuator while rejecting transient interference.

[0155] The cooperative control method for a distributed steering system provided in this application verifies the reliability of the dynamic response characteristic parameters corresponding to each wheel-end angle module under test. The parameters that pass the verification are stored in the valid parameter set corresponding to the wheel-end angle module. Specifically, for each wheel-end angle module under test, the method obtains the fitting residual during the online identification process, determines whether the fitting residual exceeds a preset residual threshold, determines whether the change amplitude of the dynamic response characteristic parameters corresponding to the wheel-end angle module under test compared to the dynamic response characteristic parameters obtained in the previous online identification exceeds a preset amplitude threshold, and determines whether the vehicle remains within a preset safe operating condition throughout the online identification process. When all the above verifications pass, the reliability verification is considered successful. This method effectively solves the engineering pain points of "data being easily contaminated and reliability being difficult to guarantee" in online identification.

[0156] Figure 6 Example 5 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 6 As shown, based on the foregoing embodiments, the method further includes:

[0157] S601: For each wheel end angle module under test, the parameters are corrected and standardized based on the historical dynamic response characteristic parameters corresponding to the wheel end angle module under test, combined with the operating environment information, to obtain the processed parameters.

[0158] The original historical dynamic response characteristic parameters are affected by environmental factors (especially temperature), resulting in normal, non-degradable drift. Without correction, this drift can obscure the true performance degradation trend. This step compensates for the drift caused by environmental factors and normalizes the parameters to a uniform benchmark scale, facilitating subsequent cross-vehicle comparisons and long-term statistics.

[0159] Specifically, the system first receives the parameter sequence corresponding to historical dynamic response characteristic parameters, along with vehicle operating environment information and wheel-end abnormal event records. Then, the historical parameters are corrected and standardized based on the environmental information. For example, normal drift caused by temperature is corrected, and incomparable environmental samples are removed or downweighted. The processed parameters are then converted into normalized offsets relative to factory or calibration reference values. The temperature correction formula is:

[0160]

[0161] in, This is the temperature-corrected gain. This is the temperature-corrected delay. This is a temperature correction factor. For reference temperature, The current temperature. The normalized offset formula is:

[0162]

[0163] in, For gain normalization offset, To delay the normalization offset, As the reference gain, The baseline delay is used.

[0164] It should be noted that the obtained historical parameter sequence can also be obtained from a set of valid parameters. This application does not impose specific limitations.

[0165] S602: Calculate the average offset, short-term volatility, and long-term drift trend corresponding to the processed parameters.

[0166] This step extracts three dimensions of degradation features from the standardized parameter sequence:

[0167] Average offset: Reflects the degree of overall performance degradation over a long period of time.

[0168] Short-term fluctuations: reflect parameter instability; increased fluctuations may indicate intermittent failures or deterioration of lubrication.

[0169] Long-term drift trend: reflects the slope of performance changes over time, either positive drift (getting worse) or negative drift.

[0170] These statistics can distinguish between "stable degradation" and "volatile degradation," providing a more detailed basis for health diagnosis.

[0171] Specifically, the average deviation, short-term volatility, and long-term drift trend of the parameters are calculated to determine whether a particular wheel end has experienced stable deterioration, intermittent increased volatility, or continuous slow degradation. The effective sample size is:

[0172] in, The number of valid samples, L represents the number of samples, serving as a valid indicator for environmental samples.

[0173] The formula for calculating the average offset is:

[0174]

[0175] in, This is the average gain offset. The average offset due to delay. For gain normalization offset, This is the delayed normalized offset.

[0176] The formula for calculating volatility is:

[0177]

[0178] in, This is the gain fluctuation. This represents the delayed fluctuation.

[0179] The formula for calculating drift trend is:

[0180]

[0181] in, This represents the gain drift trend. To delay the drift trend, For the fitting intercept, This represents the drift slope.

[0182] S603: Based on the abnormal event records corresponding to the wheel end angle module under test acquired in real time, perform abnormal number statistics, continuous abnormal length extraction and abnormal duration encoding to obtain abnormal degradation items.

[0183] Besides parameter drift, various abnormal events may occur in real-world systems (such as identification failure, verification failure, motor overheating, communication timeout, etc.). Although these events are not directly reflected in the gain / delay values, their frequent or continuous occurrence is a significant indicator of performance degradation. This step transforms these abnormal events into quantifiable degradation items and incorporates them into the health index.

[0184] Specifically, the system maintains an abnormal event log for each wheel end, recording the type, occurrence time, and duration of each abnormality. Abnormal events include, but are not limited to: confidence verification failure, response signal timeout or no response, excitation current over-limit protection, abnormal sensor readings, and motor over-temperature alarm.

[0185] The system performs anomaly count, consecutive anomaly length extraction, and anomaly duration encoding to form a degradation term that reflects the frequency and persistence of anomalies. The formula for the anomaly degradation term is:

[0186] in, For abnormal event degradation items, This represents the total number of anomalies that occurred within the historical window. The weights are the number of times and the duration, where L is the number of samples. For the duration of the abnormality, This refers to the parameter update cycle.

[0187] S604: A health index is formed by combining average offset, short-term volatility, long-term drift trend and abnormal degradation items.

[0188] This step integrates the aforementioned multidimensional features (average offset, short-term fluctuations, long-term drift, and abnormal degradation) into a health index, which intuitively reflects the current health status of the wheel. A lower health index indicates better performance, while a higher index indicates more severe degradation. This index can be used for tiered early warning and control decisions.

[0189] Specifically, the formula for calculating the health index is:

[0190] in, The health index of the i-th wheel end. For current gain and reference gain, For current delay and reference delay, This is the gain fluctuation. This refers to the slip delay fluctuation. This represents the gain drift trend. To delay the drift trend, For abnormal event degradation items, This is a weighting coefficient. The higher the health index, the more likely the wheel end is in a state of degradation or failure.

[0191] S605: Based on the health index and the duration of the health index continuously falling within any of the preset health levels, classify the wheel end status into levels and output status records, maintenance prompts, function restriction commands, fault alarm signals, or degradation control requests respectively.

[0192] This step classifies the wheel-end status based on the magnitude of the health index and the duration of its persistence at the same level, and outputs corresponding operation instructions. This tiered processing avoids false alarms for instantaneous fluctuations while enabling timely responses to continuously deteriorating trends, achieving a progressive safety strategy from "recording" to "degradation control."

[0193] Specifically, this step does not rely solely on a single health index, but combines the health index with various sub-features to classify the current condition as closer to normal, mildly deteriorated, significantly deteriorated, or suspected faulty, thereby improving diagnostic interpretability. Finally, the wheel-end status is graded based on the magnitude and duration of the health index, and records, maintenance prompts, functional limitations, fault alarms, or degraded control requests are output accordingly. Through this implementation method of "parameter preprocessing—fluctuation and drift analysis—anomaly accumulation—health fusion—type discrimination—graded output," the online identification results are extended from the immediate control level to the long-term health management level, enabling the same set of parameter links to simultaneously possess the dual functions of collaborative control and predictive maintenance.

[0194] For example, the preset health level can be divided into 4 to 5 levels, such as:

[0195] Level 0 (Healthy): H<H1, no action is performed or only recording is performed.

[0196] Level 1 (Mild degradation): H1≤H<H2, a maintenance prompt is output.

[0197] Level 2 (Significant degradation): H2≤H<H3, a function restriction command is output.

[0198] Level 3 (Severe degradation): H3≤H<H4, a fault alarm signal is output.

[0199] Level 4 (Failure): H≥H4, a degraded control request is output (e.g., disabling the independent steering of the wheel and switching to the safe mode). Wherein, H is the health index.

[0200] Then a timer is introduced, and timing starts when the health index enters a certain level for the first time. The level is only confirmed to be valid if the health index stays continuously within this level for more than a preset duration (e.g., 10 seconds or 3 consecutive driving cycles). If the health index falls back to a lower level, the timer is reset to zero. This avoids misoperation caused by short-term spikes.

[0201] Wherein, the output content includes:

[0202] Status recording: the health index, level and timestamp are stored in a non-volatile memory for offline analysis.

[0203] Maintenance prompt: the driver or fleet is reminded to arrange maintenance through an instrument or a telematics system.

[0204] Function restriction command: the maximum steering angle, angular velocity or torque output of the wheel end is actively restricted to prevent further degradation.

[0205] Fault alarm signal: the fault indicator light is turned on, and a buzzer may be triggered.

[0206] Degraded control request: the vehicle control unit is requested to restrict the vehicle from entering high-dynamic working conditions such as crab walking and pivot steering, or to switch to a fault-tolerant control mode.

[0207] The collaborative control method for a distributed steering system provided in this application, for each wheel-end angle module under test, corrects and standardizes the parameters based on the historical dynamic response characteristic parameters corresponding to the module and the operating environment information to obtain processed parameters. It then calculates the average offset, short-term fluctuation, and long-term drift trend corresponding to the processed parameters. Based on real-time acquired abnormal event records corresponding to the wheel-end angle modules under test, it performs anomaly count, continuous anomaly length extraction, and anomaly duration encoding to obtain anomaly degradation items. A health index is formed by fusing the average offset, short-term fluctuation, long-term drift trend, and anomaly degradation items. Based on the health index and the duration of the health index continuously falling within any of a preset range of health levels, the wheel-end state is classified into levels, and status records, maintenance prompts, function restriction commands, fault alarm signals, or degradation control requests are output respectively. This method implements a progressive safety strategy from status recording to proactive degradation. It can not only compensate for dynamic differences between multiple wheels in real time but also predict the degradation trend of actuators, significantly improving the reliability, maintainability, and safety of the distributed steering system.

[0208] Figure 7 Example 6 of the cooperative control method for the distributed steering system provided in this solution, such as Figure 7 As shown, based on the foregoing embodiments, the method further includes:

[0209] S701: Receives real-time status signals of the vehicle.

[0210] S702: Determine whether the vehicle is in a preset safe operating condition based on real-time status signals and pre-set multi-dimensional safety thresholds.

[0211] The foundation for determining safe operating conditions is obtaining the vehicle's current real-time operating status. By collecting real-time signals from multiple sensors, a data source is provided for subsequent comparisons with preset safety thresholds. These signals encompass the driver's active operational intentions and the vehicle's dynamic stability, ensuring that identification and triggering are only permitted when the driver has no steering intention and the vehicle is in a steady state.

[0212] Specifically, real-time status signals may include, but are not limited to, vehicle speed, vehicle steering wheel angle, steering wheel angular velocity, steering wheel torque, yaw rate, and lateral acceleration.

[0213] Then, using a pre-calibrated set of safety thresholds (multidimensional), the collected real-time status is compared in parallel to determine whether the vehicle currently meets conditions such as "the driver has no intention to actively steer, the vehicle's dynamics are stable, and the speed is appropriate." Only when all threshold conditions are met simultaneously and continuously for a preset duration is the online identification process allowed. This ensures that the identification process does not interfere with normal driving or introduce additional disturbances when the vehicle is unstable.

[0214] It receives state variables such as steering wheel angle, steering wheel angular velocity, steering wheel torque, lateral acceleration, yaw rate, and vehicle speed, and compares them with pre-calibrated safety thresholds to determine whether the driver has a clear intention to actively steer, whether the vehicle is in a strong yaw or lateral dynamic state, and whether the vehicle speed is within the range that allows for triggering identification.

[0215] Specifically, the vehicle is considered to be in a preset safe operating condition when all of the following conditions are met:

[0216] The absolute value of the vehicle steering wheel angle is less than the first threshold, the absolute value of the steering wheel angular velocity is less than the second threshold, the absolute value of the steering wheel torque is less than the third threshold, the absolute value of the vehicle lateral acceleration is less than the fourth threshold, the absolute value of the yaw rate is less than the fifth threshold, the vehicle speed is between the preset minimum speed and the maximum speed, and all conditions are maintained continuously for a duration that reaches the preset duration threshold.

[0217] It should be noted that identification is not triggered immediately at a single sampling moment. Instead, the duration of states that meet the threshold condition is accumulated. Only after the safety condition is maintained continuously for the set duration is the identification trigger signal output to avoid accidentally entering the identification process due to short-term, accidental stability. The specific implementation formula is as follows:

[0218] in For steering wheel angle, This is the steering wheel angle threshold. Steering wheel speed, The steering wheel speed threshold. Steering wheel torque, Steering wheel torque threshold, For lateral acceleration, Here, r is the lateral acceleration threshold, and r is the yaw rate. The yaw rate threshold. For vehicle speed, These are the minimum and maximum threshold values ​​for vehicle speed. The duration for which the condition is met continuously. To maintain the time threshold.

[0219] When the judgment condition is met, an excitation signal is output, and the parameter identification link is notified to enter a standby state. When the judgment condition is violated during the identification process, such as when the driver suddenly turns the steering wheel, the vehicle yaw disturbance increases, or the vehicle speed exceeds the limit range, an abort signal is immediately issued to stop the micro-perturbation excitation, terminate the current data acquisition, cancel the current round of parameter updates, and mark the current identification result as invalid, thereby preventing data from polluting system parameters under abnormal operating conditions. This implementation method of "judging first, then continuously verifying, then triggering, and immediately aborting when the condition fails" is equivalent to establishing a dynamic safety gating mechanism for the entire online identification link, which is a key preliminary link to ensure the availability of system engineering and driving safety.

[0220] The cooperative control method for a distributed steering system provided in this application receives real-time vehicle status signals and determines whether the vehicle is in a preset safe operating condition based on the real-time status signals and pre-set multi-dimensional safety thresholds. This method first collects multi-source status signals in real time, and then accurately determines whether to allow entry into the online identification mode through multi-dimensional thresholds and duration constraints. This is a key prerequisite for actively sensing dynamic differences at the wheel ends without interfering with normal driving.

[0221] Figure 8 A schematic diagram of the cooperative control device for the distributed steering system provided in this solution is shown below. Figure 8 As shown, the cooperative control device 80 of the distributed steering system specifically includes:

[0222] The acquisition module 81 is used to apply limited perturbation excitation signals to multiple wheel end angle modules under test when the vehicle is under a preset safe operating condition, and to acquire the response signal of each wheel end angle module under test.

[0223] The online identification module 82 is used to identify the dynamic response characteristic parameters of each wheel end angle module under test based on the perturbation excitation signal and the response signal of the wheel end angle module under test. The dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules.

[0224] The first calculation module 83 is used to calculate the feedforward compensation coefficient corresponding to the wheel end angle module based on the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested.

[0225] The injection module 84 is used to inject the feedforward compensation coefficient corresponding to each wheel end angle module under test into the control target of the corresponding wheel end angle module under test, so as to compensate for the dynamic response differences between the wheel end angle modules.

[0226] Furthermore, since the dynamic response characteristic parameters include equivalent dynamic gain and equivalent delay, the online identification module 82 is specifically used for:

[0227] For each wheel end angle module to be tested, the response signal of the wheel end angle module to be tested is extracted to obtain the response feature quantity;

[0228] Based on the response feature quantities, the wheel end angle module under test is modeled and the dynamic characteristics of the wheel end are fitted. The equivalent dynamic gain and equivalent delay corresponding to the wheel end angle module under test are calculated.

[0229] Furthermore, the first calculation module 83 specifically includes:

[0230] A multi-round synchronization target is constructed based on the dynamic response characteristic parameters corresponding to each wheel end angle module under test. The multi-round synchronization target includes reference gain and reference delay.

[0231] For each wheel end angle module under test, the feedforward compensation coefficient corresponding to the wheel end angle module under test is calculated based on the deviation between the dynamic response characteristic parameters of the wheel end angle module under test and the multi-wheel synchronization target.

[0232] Furthermore, the cooperative control device 80 of the distributed steering system also includes:

[0233] The verification module 85 is used to verify the reliability of the dynamic response characteristic parameters corresponding to each wheel end angle module under test, and to store the parameters that pass the verification into the valid parameter set corresponding to the wheel end angle module under test.

[0234] Accordingly, the first calculation module 83 specifically includes:

[0235] Based on the effective parameter set corresponding to each wheel end angle module under test, the corresponding feedforward compensation coefficient is calculated for each wheel end angle module under test.

[0236] Furthermore, the verification module 85 specifically includes:

[0237] For each wheel end angle module to be tested, the fitting residual during the online identification process of the wheel end angle module to be tested is obtained, and it is determined whether the fitting residual exceeds the preset residual threshold.

[0238] Determine whether the change amplitude of the dynamic response characteristic parameters corresponding to the wheel end angle module under test exceeds the preset amplitude threshold compared with the dynamic response characteristic parameters obtained from the previous online identification of the wheel end angle module under test.

[0239] Determine whether the vehicle remains within the preset safe operating conditions during the online identification process of the wheel end angle module under test;

[0240] If all the above checks pass, the credibility check is considered successful.

[0241] Furthermore, the cooperative control device 80 of the distributed steering system also includes:

[0242] The correction module 86 is used to correct and standardize the parameters for each wheel end angle module under test, based on the historical dynamic response characteristic parameters corresponding to the wheel end angle module under test and the operating environment information, so as to obtain the processed parameters.

[0243] The second calculation module 87 is used to calculate the average offset, short-term fluctuation and long-term drift trend of the processed parameters.

[0244] Processing module 88 is used to perform anomaly count, continuous anomaly length extraction and anomaly duration encoding based on the anomaly event records corresponding to the test wheel end angle module acquired in real time, to obtain anomaly degradation items;

[0245] The fusion module 89 is used to fuse the average offset, short-term fluctuation, long-term drift trend and abnormal degradation items to form a health index.

[0246] The classification module 810 is used to classify the wheel end status according to the health index and the duration of the health index continuously being in any of the preset multiple health levels, and output status records, maintenance prompts, function restriction instructions, fault alarm signals or degradation control requests respectively.

[0247] Furthermore, the cooperative control device 80 of the distributed steering system also includes:

[0248] The receiving module 811 is used to receive real-time status signals of the vehicle;

[0249] The determination module 812 is used to determine whether the vehicle is in a preset safe operating condition based on real-time status signals and preset multi-dimensional safety thresholds.

[0250] Furthermore, the real-time status signals include vehicle speed, vehicle steering wheel angle, steering wheel angular velocity, steering wheel torque, yaw rate, and lateral acceleration. Therefore, module 812 specifically includes:

[0251] The vehicle is considered to be in a preset safe operating condition when all of the following conditions are met:

[0252] The absolute value of the vehicle steering wheel angle is less than the first threshold, the absolute value of the steering wheel angular velocity is less than the second threshold, the absolute value of the steering wheel torque is less than the third threshold, the absolute value of the vehicle lateral acceleration is less than the fourth threshold, the absolute value of the yaw rate is less than the fifth threshold, the vehicle speed is between the preset minimum speed and the maximum speed, and all of the above conditions are maintained continuously for a preset duration threshold.

[0253] Furthermore, the perturbation excitation signal is one or more combinations of positive and negative symmetrical double pulse signals, single pulse signals, pseudo-random sequence signals, or small-amplitude sweep frequency signals.

[0254] Figure 9 The automotive structure diagram provided for this solution is as follows: Figure 9 As shown, the automobile includes a vehicle body 90, a storage unit 91 disposed in the vehicle body, and an electronic control unit 92;

[0255] Optionally, the vehicle also includes a communication component 93. The electronic control unit 92 and the communication component 93 are connected via a bus 94.

[0256] In the specific implementation process, at least one electronic control unit 92 executes the computer execution instructions stored in the storage unit 91, so that at least one electronic control unit 92 performs the above-described method.

[0257] The specific implementation process of the electronic control unit 92 can be found in the above method embodiment, and its implementation principle and technical effect are similar. It will not be repeated here.

[0258] In the above embodiments, it should be understood that the electronic control unit can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0259] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0260] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0261] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0262] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0263] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0264] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0265] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0266] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0267] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0268] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0269] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0270] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

Claims

1. A cooperative control method for a distributed steering system, characterized in that, include: When the vehicle is in a preset safe operating condition, a limited perturbation excitation signal is sequentially applied to multiple wheel end angle modules to be tested, and the response signal of each wheel end angle module to be tested is collected. For each wheel end angle module under test, based on the perturbation excitation signal and the response signal of the wheel end angle module under test, the dynamic response characteristic parameters corresponding to the wheel end angle module under test are identified online. The dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules. The feedforward compensation coefficient corresponding to the wheel end angle module under test is calculated based on the dynamic response characteristic parameters corresponding to the wheel end angle module under test. The feedforward compensation coefficient corresponding to each wheel end angle module under test is injected into the control target of the corresponding wheel end angle module under test in order to compensate for the dynamic response differences between the wheel end angle modules.

2. The method according to claim 1, characterized in that, The dynamic response characteristic parameters include equivalent dynamic gain and equivalent delay. Therefore, the step of identifying the dynamic response characteristic parameters corresponding to the wheel end angle module under test online based on the perturbation excitation signal and the response signal of the module under test includes: For each wheel end angle module to be tested, feature extraction is performed on the response signal of the wheel end angle module to obtain response feature quantities; The test wheel end angle module is modeled and fitted with the dynamic characteristics of the wheel end based on the response feature quantity, and the equivalent dynamic gain and equivalent delay corresponding to the test wheel end angle module are calculated.

3. The method according to claim 1, characterized in that, The calculation of the feedforward compensation coefficient corresponding to the wheel end angle module under test based on the dynamic response characteristic parameters of the module under test includes: A multi-round synchronization target is constructed based on the dynamic response characteristic parameters corresponding to each wheel end angle module under test. The multi-round synchronization target includes reference gain and reference delay. For each wheel end angle module to be tested, the feedforward compensation coefficient corresponding to the wheel end angle module to be tested is calculated based on the deviation between the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested and the multi-wheel synchronization target.

4. The method according to claim 1, characterized in that, The method further includes: For each wheel end angle module to be tested, the reliability of the dynamic response characteristic parameters corresponding to the wheel end angle module to be tested is verified, and the parameters that pass the verification are stored in the valid parameter set corresponding to the wheel end angle module to be tested. Accordingly, the step of calculating the corresponding feedforward compensation coefficient for each wheel end angle module based on the dynamic response characteristic parameters of each module under test includes: Based on the effective parameter set corresponding to each wheel end angle module under test, the corresponding feedforward compensation coefficient is calculated for each wheel end angle module under test.

5. The method according to claim 4, characterized in that, The credibility verification of the dynamic response characteristic parameters corresponding to the wheel end angle module under test includes: For each wheel end angle module to be tested, the fitting residual during the online identification process of the wheel end angle module to be tested is obtained, and it is determined whether the fitting residual exceeds a preset residual threshold. Determine whether the change range of the dynamic response feature parameter corresponding to the wheel end angle module under test and the dynamic response feature parameter obtained from the previous online identification of the wheel end angle module under test exceeds a preset amplitude threshold. Determine whether the vehicle is always within a preset safe operating condition during the online identification process of the wheel end angle module under test; If all the above checks pass, the credibility check is considered successful.

6. The method according to claim 1, characterized in that, The method further includes: For each wheel end angle module to be tested, the parameters are corrected and standardized based on the historical dynamic response characteristic parameters corresponding to the wheel end angle module and the operating environment information, so as to obtain the processed parameters. Calculate the average offset, short-term fluctuation, and long-term drift trend corresponding to the processed parameters; Based on the abnormal event records corresponding to the wheel end angle module under test acquired in real time, the number of abnormal events is counted, the length of continuous abnormal events is extracted, and the duration of abnormal events is encoded to obtain abnormal degradation items. A health index is formed by fusing the average offset, the short-term fluctuation, the long-term drift trend, and the abnormal degradation items. Based on the health index and the duration of the health index continuously falling within any of the preset multiple health levels, the wheel end status is classified into levels, and status records, maintenance prompts, function restriction instructions, fault alarm signals, or degradation control requests are output respectively.

7. The method according to claim 1, characterized in that, The method further includes: Receive the real-time status signal of the vehicle; The vehicle is determined to be in a preset safe operating condition based on the real-time status signal and the preset multi-dimensional safety threshold.

8. The method according to claim 7, characterized in that, The real-time status signals include vehicle speed, vehicle steering wheel angle, steering wheel angular velocity, steering wheel torque, yaw rate, and lateral acceleration. Determining whether the vehicle is in a preset safe operating condition based on the real-time status signals and a pre-set multi-dimensional safety threshold includes: The vehicle is determined to be in a preset safe operating condition when all of the following conditions are met: The absolute value of the vehicle steering wheel angle is less than a first threshold, the absolute value of the steering wheel angular velocity is less than a second threshold, the absolute value of the steering wheel torque is less than a third threshold, the absolute value of the vehicle lateral acceleration is less than a fourth threshold, the absolute value of the yaw rate is less than a fifth threshold, the vehicle speed is between a preset minimum speed and a maximum speed, and all of the above conditions are maintained continuously for a preset duration threshold.

9. The method according to claim 1, characterized in that, The perturbation excitation signal is one or more combinations of positive and negative symmetrical double pulse signal, single pulse signal, pseudo-random sequence signal or small-amplitude sweep frequency signal.

10. A cooperative control device for a distributed steering system, characterized in that, include: An acquisition module is used to apply restricted perturbation excitation signals to multiple wheel end angle modules under test when the vehicle is in a preset safe operating condition, and to acquire the response signal of each wheel end angle module under test. The online identification module is used to identify the dynamic response characteristic parameters of each wheel end angle module under test based on the perturbation excitation signal and the response signal of the wheel end angle module under test. The dynamic response characteristic parameters are used to characterize the differences in response speed and response intensity between different wheel end angle modules. The calculation module is used to calculate the feedforward compensation coefficient corresponding to the wheel end angle module under test based on the dynamic response characteristic parameters corresponding to the wheel end angle module under test; The injection module is used to inject the feedforward compensation coefficient corresponding to each wheel end angle module under test into the control target of the corresponding wheel end angle module under test, so as to compensate for the dynamic response differences between the wheel end angle modules.

11. A car, characterized in that, include: The vehicle body, including storage units and electronic control units located within it; The storage unit stores computer-executed instructions; The electronic control unit executes the computer execution instructions stored in the storage unit to implement the method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 9.

Images