Steering collaborative control method in steering wheel silent mode

Abstract

The present application belongs to the technical field of steer-by-wire closed-loop control, and relates to a steering cooperative control method in a steering wheel mute mode in steer-by-wire closed-loop control, which comprises the following steps: obtaining a target trajectory signal in a mute tracking working condition, and adjusting a controlled angle variable by using a position closed-loop circuit; multiplying an input torque parameter and a feedback angle deviation change rate parameter to generate a characteristic power scalar; calculating a dynamic critical energy boundary operator of a counteracting transmission loss according to transmission intrinsic damping and inertia parameters; comparing the characteristic power scalar with the dynamic critical energy boundary operator, and triggering a control right state migration to an external command following state when a condition is met, wherein the present application changes state judgment dimensions, distinguishes forced vibration of a transmission chain and active takeover control, blocks parasitic dynamics disturbance, and guarantees the adjustment stability of the position closed-loop control circuit.

Description

Technical Field

[0001] This invention belongs to the field of steer-by-wire closed-loop control technology, and relates to a steering coordination control method in steering wheel silent mode. Background Technology

[0002] With the development of intelligent driving technology, combining data from multiple sensors with the motion state of chassis actuators for active vehicle safety and cooperative control has become a common technical approach. However, existing technologies still have certain limitations in utilizing external driving environment risk information, processing the dynamic state of the actuators themselves, and coordinating the control boundaries between the driver and automatic control. First, existing technologies include schemes for assisting vehicle stability control based on environmental risk information. For example, Chinese invention patent application CN116853232A discloses a driving safety assistance method, device, and intelligent terminal based on vehicle positioning. This scheme obtains external environmental information such as rainfall and temperature through the Global Positioning System. After identifying risk events, it calculates the similarity between real-time data and classified historical vehicle driving information to match similar historical scenarios. Based on this, it adjusts the vehicle stability control threshold and implements mandatory speed limits to assist driving safety. This scheme mainly targets stability intervention and amplitude limitation at the vehicle level. It does not establish a feedforward connection between external risk information and the real-time control of specific actuators. In extreme emergency obstacle avoidance or complex human-machine co-driving conditions, the adjustment capabilities of specific actuators are difficult to fully utilize, and the system decision-making and control response may not be fully covered.

[0003] Secondly, existing technologies also include schemes that estimate the input torque of steer-by-wire based on internal state-space models. For example, Chinese invention patent CN120117034B discloses a vehicle control method and device, electronic equipment, and storage medium. This scheme obtains the operating condition information of the steering wheel actuator and the system dynamic parameters related to the friction model when the driver operates the steering wheel. This information is then input into a pre-established state-space model, and the corrected target input torque is obtained through Kalman filtering to control the vehicle steering. The state estimation of this scheme mainly relies on the internal dynamic characteristics and operating state data of the steering wheel actuator. State transition and gain correction are also completed inside the actuator. Due to the lack of external risk information from the vehicle's active safety system as feedforward input, this scheme is difficult to adapt to the uncertain interference caused by extreme and sudden road conditions in a timely manner. When the vehicle is close to losing control, the estimated target input torque may not be consistent with the actual physical safety boundary of the vehicle.

[0004] Third, existing technologies also include schemes that use Kalman filter state observers to estimate the rack force of steer-by-wire and reproduce road feel. For example, Chinese invention patent application CN116902066A discloses a resistance feedback steering actuator and a steer-by-wire system control method. This scheme uses a torque angle sensor to collect the rotation angle and torque information of the pinion shaft during steering, estimates the rack force through a Kalman filter state observer, and calculates the steering resistance torque in combination with a friction model. Then, it drives a road feel feedback motor to generate corresponding road feel feedback. This scheme mainly feeds back the state of the actuator to the driver, which is a one-way passive interaction method that transmits road feel from the actuator to the driver. It does not use the estimated real-time rack force and road feel resistance information to actively avoid risks or correct extreme trajectories at the actuator. In addition, when the viscous characteristics of mechanical components or the friction characteristics of the ground suddenly change, causing the noise covariance to be inconsistent with the actual working conditions, the state estimation is prone to deviation due to the lack of clear constraints on the overall vehicle safety risk information, which in turn causes the road feel feedback to be too large or distorted.

[0005] Therefore, the technical problem to be solved by this invention is how to establish an integrated predictive control architecture that can utilize various environmental risk information in advance under different working conditions and coordinate with the dynamic state of the chassis actuators themselves, so that the vehicle chassis intervention and road feel feedback are kept within the vehicle's true physical safety boundary throughout the entire control process. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art, solve the technical problems of control control transfer being falsely triggered by mechanical parasitic noise and position feedback jumps in the control loop during online switching and reconfiguration, and provide a steering coordination control method in the silent mode of the steering wheel.

[0007] To achieve the above-mentioned objective, the present invention provides a steering coordination control method in a steering wheel silent mode, comprising the following steps: Step S101, Closed-loop adjustment of the controlled angle variable to follow the target trajectory: Under the silent tracking condition of input target trajectory signal, the vehicle target trajectory signal at the angle input end is acquired, and the controlled angle variable is adjusted by the position closed-loop loop to follow the vehicle target trajectory signal. Step S102, Multiplying torque and feedback deviation rate to generate a power scalar: Collect the input torque parameter and feedback angle deviation change rate parameter during the controlled angle variable follow-up adjustment process, multiply the input torque parameter and feedback angle deviation change rate parameter to generate a characteristic power scalar that characterizes the external control energy input rate; Step S103, accumulate various parameters to solve the dynamic energy boundary operator: Based on the transmission intrinsic damping parameter and transmission intrinsic inertia parameter corresponding to the position closed loop, accumulate the product of the absolute value of the reference dissipation power constant, the absolute value of the feedback angle deviation change rate parameter and the damping adaptive compensation coefficient, and the absolute value of the steering angle acceleration parameter and the inertia counterbalancing correction coefficient stored in the non-volatile storage medium to generate a dynamic critical energy boundary operator for counterbalancing the intrinsic loss of mechanical transmission. The steering angle acceleration parameter is generated by performing a first-order discrete difference operation on the feedback angle deviation change rate parameter. Step S104, compare the scalar and operator to trigger the state transition of control: compare the characteristic power scalar with the dynamic critical energy boundary operator. When the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator within the range of continuous discrete sampling periods, the control is triggered to transition from the target trajectory signal tracking state to the external command following state.

[0008] The step S103 of this invention, which generates a dynamic critical energy boundary operator for offsetting intrinsic losses in mechanical transmission, includes the following sub-steps: Step S1031, performing a first-order discrete difference operation on the feedback angle deviation rate of change parameter to generate a steering angle acceleration parameter; Step S1032, multiplying the absolute value of the feedback angle deviation rate of change parameter by a damping adaptive compensation coefficient to generate a damping compensation term; Step S1033, multiplying the absolute value of the steering angle acceleration parameter by an inertial offset correction coefficient to generate an inertial correction term; Step S1034, adding the reference dissipation power constant, the damping compensation term, and the inertial correction term to output the dynamic critical energy boundary operator.

[0009] The damping adaptive compensation coefficient and inertial offset correction coefficient described in this invention are obtained through the following steps: Step S1035, in offline state, a frequency domain sweep signal under standard disconnection condition is input to the steering transmission structure to measure the intrinsic energy loss gradient of the steering transmission structure under different angular velocities and steering angular accelerations; Step S1036, based on the intrinsic energy loss gradient of the transmission, a mapping model about angular velocity and steering angular acceleration is established, the corresponding damping adaptive compensation coefficient and inertial offset correction coefficient are calculated and stored in a non-volatile storage medium.

[0010] Step S104 of the present invention includes the following sub-steps: Step S1041, within the continuous discrete sampling period, the statistical characteristic power scalar exceeds the number of discrete samplings of the dynamic critical energy boundary operator; Step S1042, when the number of discrete samplings reaches a predetermined continuous sliding window threshold, the control right is triggered to migrate from the target trajectory signal tracking state to the external command following state.

[0011] The present invention, after triggering the transfer of control from the target trajectory signal tracking state to the external command following state, includes the following sub-steps: Step S1043, calculating the power overflow difference of the characteristic power scalar exceeding the dynamic critical energy boundary operator; Step S1044, using a nonlinear increasing function about the power overflow difference to solve the takeover weighting factor; Step S1045, based on the takeover weighting factor, adjusting the control weight ratio in the target trajectory signal tracking state during the switching cycle and simultaneously adjusting the control weight ratio in the external command following state to complete the switching of control states.

[0012] The present invention includes the following steps after the control weight ratio has been smoothly switched and is in the external command follow state: Step S105, continuously monitor the amplitude change trend of the characteristic power scalar within a predetermined reset time window; Step S106, when the characteristic power scalar is always lower than the reference dissipation power constant within the reset time window, trigger the system to exit the external command follow state.

[0013] The step S106 of the present invention includes the following sub-steps: Step S1061, collecting the current real-time physical position parameters of the steering; Step S1062, synchronizing the real-time physical position parameters of the steering as the initial state variables of the position closed loop, so that the controlled angle variables reconverge to the vehicle target trajectory signal within the predetermined tracking deviation range.

[0014] In step S102 of the present invention, when facing external periodic high-frequency disturbance conditions, the input torque parameter and the feedback angle deviation change rate parameter exhibit alternating positive and negative signs in the time domain. The cumulative value of their product after being added and averaged in the discrete time domain approaches 0 and is restricted within the dynamic critical energy boundary operator.

[0015] In step S101 of the present invention, the calculation process of position closed-loop loop following adjustment includes the following sub-steps: step S1011, acquiring the vehicle target trajectory signal with a fixed discrete control cycle; step S1012, calculating the target angle deviation value between the current steering angle feedback parameter and the vehicle target trajectory signal; step S1013, calculating the target angle deviation value using a preset proportional-integral-derivative control algorithm, and generating a control output quantity for driving and adjusting steering non-electrical variables.

[0016] Compared with the prior art, the present invention has at least the following beneficial effects: 1. By generating a characteristic power scalar through input torque parameters and feedback angle deviation change rate parameters, it is possible to distinguish between periodic disturbances to the steering transmission structure and continuous driver intervention operations from the energy input state. The characteristic power scalar generated by external periodic high-frequency disturbances such as road vibrations alternates between positive and negative values, making it less likely to form a continuous exceedance condition. When the driver actively takes over, a continuous external energy input is generated. Therefore, it is possible to reduce the false transfer of control rights caused by mechanical vibration, and at the same time, it is not necessary to avoid disturbances by expanding the static judgment interval, which is conducive to maintaining the adjustment sensitivity and stability of the position closed loop.

[0017] 2. The dynamic critical energy boundary operator can reflect the damping, friction, and inertia effects of the steering transmission structure itself based on the feedback angle deviation rate of change parameter and steering angle acceleration parameter as the steering state changes. It can also reflect the damping, friction, and inertia effects of the steering transmission structure itself through the reference dissipation power constant, damping adaptive compensation coefficient, and inertia offset correction coefficient. This can avoid mistaking the intrinsic losses of mechanical transmission as external control input, improve the accuracy of control transfer judgment, enable the control system to maintain a relatively consistent judgment standard under different steering speeds and accelerations, and reduce the lag and interference caused by mechanical transmission characteristics to the control switching process.

[0018] 3. By judging whether the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator within the continuous discrete sampling period, the control state switching triggered by single sampling noise or short-term disturbance can be avoided. After confirming that the takeover condition is met, the control weight ratio of the target trajectory signal tracking state and the external command following state is gradually adjusted according to the takeover weighting factor. This can reduce the sudden change in steering output caused by direct switching of control power and avoid frequent switching between the two control states, thereby improving the stability of the steering non-electric variable closed-loop regulation and the continuity of control state switching. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the control status transition determination process of the present invention; Figure 2 This is a schematic diagram comparing the characteristic power scalar and the dynamic critical energy boundary operator of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings.

[0021] Example 1: This embodiment discloses a steering coordination control method in a silent steering wheel mode, including the following steps: Step S101, Closed-loop adjustment of the controlled angle variable to follow the target trajectory: Under the silent tracking condition of input target trajectory signal, the vehicle target trajectory signal at the angle input end is acquired, and the controlled angle variable is adjusted by the position closed-loop loop to follow the vehicle target trajectory signal. Step S102, Multiplying torque and feedback deviation rate to generate a power scalar: Collect the input torque parameter and feedback angle deviation change rate parameter during the controlled angle variable follow-up adjustment process, multiply the input torque parameter and feedback angle deviation change rate parameter to generate a characteristic power scalar that characterizes the external control energy input rate; Step S103, accumulate various parameters to solve the dynamic energy boundary operator: Based on the transmission intrinsic damping parameter and transmission intrinsic inertia parameter corresponding to the position closed loop, accumulate the product of the absolute value of the reference dissipation power constant, the absolute value of the feedback angle deviation change rate parameter and the damping adaptive compensation coefficient, and the absolute value of the steering angle acceleration parameter and the inertia counterbalancing correction coefficient stored in the non-volatile storage medium to generate a dynamic critical energy boundary operator for counterbalancing the intrinsic loss of mechanical transmission. The steering angle acceleration parameter is generated by performing a first-order discrete difference operation on the feedback angle deviation change rate parameter. Step S104, compare the scalar and operator to trigger the state transition of control: compare the characteristic power scalar with the dynamic critical energy boundary operator. When the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator within the range of continuous discrete sampling periods, the control is triggered to transition from the target trajectory signal tracking state to the external command following state.

[0022] The step S103 in this embodiment, which generates the dynamic critical energy boundary operator for offsetting intrinsic losses of mechanical transmission, includes the following sub-steps: Step S1031, performing a first-order discrete difference operation on the feedback angle deviation rate of change parameter to generate the steering angle acceleration parameter; Step S1032, multiplying the absolute value of the feedback angle deviation rate of change parameter by the damping adaptive compensation coefficient to generate the damping compensation term; Step S1033, multiplying the absolute value of the steering angle acceleration parameter by the inertial offset correction coefficient to generate the inertial correction term; Step S1034, adding the reference dissipation power constant, the damping compensation term, and the inertial correction term to output the dynamic critical energy boundary operator.

[0023] The damping adaptive compensation coefficient and inertial offset correction coefficient described in this embodiment are obtained through the following steps: Step S1035, in offline state, a frequency domain sweep signal under standard disconnection condition is input to the steering transmission structure to measure the intrinsic energy loss gradient of the steering transmission structure under different angular velocities and steering angular accelerations; Step S1036, based on the intrinsic energy loss gradient of the transmission, a mapping model about angular velocity and steering angular acceleration is established, the corresponding damping adaptive compensation coefficient and inertial offset correction coefficient are calculated and stored in a non-volatile storage medium.

[0024] Step S104 in this embodiment includes the following sub-steps: Step S1041, within the continuous discrete sampling period, the statistical characteristic power scalar exceeds the number of discrete samplings of the dynamic critical energy boundary operator; Step S1042, when the number of discrete samplings reaches a predetermined continuous sliding window threshold, the control is triggered to migrate from the target trajectory signal tracking state to the external command following state.

[0025] The following sub-steps are included in this embodiment after the control authority is transferred from the target trajectory signal tracking state to the external command following state: Step S1043, calculate the power overflow difference of the characteristic power scalar exceeding the dynamic critical energy boundary operator; Step S1044, use a nonlinear increasing function about the power overflow difference to solve the takeover weighting factor; Step S1045, based on the takeover weighting factor, reduce the control weight ratio in the target trajectory signal tracking state during the switching cycle, and simultaneously increase the control weight ratio in the external command following state to complete the switching of the control state.

[0026] The embodiment describes the following steps after the control weight ratio has been smoothly switched and is in the external command follow state: Step S105, continuously monitor the amplitude change trend of the characteristic power scalar within a predetermined reset time window; Step S106, when the characteristic power scalar is always lower than the reference dissipation power constant within the reset time window, trigger the system to exit the external command follow state.

[0027] Step S106 in this embodiment includes the following sub-steps: Step S1061, collecting the current real-time physical position parameters of the steering; Step S1062, synchronizing the real-time physical position parameters of the steering as the initial state variables of the position closed loop, so that the controlled angle variables reconverge to the vehicle target trajectory signal within the predetermined tracking deviation range.

[0028] In step S102 of this embodiment, when facing external periodic high-frequency disturbance conditions, the input torque parameter and the feedback angle deviation change rate parameter exhibit alternating positive and negative signs in the time domain. The cumulative value of their product after being added and averaged in the discrete time domain approaches 0 and is restricted within the dynamic critical energy boundary operator.

[0029] In step S101 of this embodiment, the calculation process of position closed-loop loop following adjustment includes the following sub-steps: Step S1011, acquiring the vehicle target trajectory signal with a fixed discrete control cycle; Step S1012, calculating the target angle deviation value between the current steering angle feedback parameter and the vehicle target trajectory signal; Step S1013, calculating the target angle deviation value using a preset proportional-integral-derivative control algorithm, and generating a control output quantity for driving and adjusting steering non-electrical variables.

[0030] Example 2: In this embodiment, when the autonomous vehicle traverses bumpy roads or performs continuous high-frequency obstacle avoidance on complex urban roads, the vehicle's steering transmission system operates in silent tracking mode. The position closed-loop circuit independently reads the vehicle's target trajectory signal, generates a steering drive signal in real time, and adjusts the controlled angle variable. The control output generated by the position closed-loop circuit is a pulse-width modulated current signal. This current signal drives the steering motor to output electromagnetic torque, which is then converted into linear displacement of the rack by the reduction gearbox and steering rack mechanism. The controlled angle variable is adjusted by changing the mechanical and physical position of the steering wheel push-pull rod. When the chassis drive chain experiences transient forced vibration due to uneven road surfaces or is subjected to external periodic high-frequency disturbances, the sensor circuit generates a torque disturbance that does not correspond to the intention of manual intervention. The system's underlying arbitration logic reads the input torque parameters in real time. The feedback angle deviation change rate parameter of the position closed loop And calculate the characteristic power scalar representing the external control energy input rate according to the discrete control cycle. Its expression is as follows: ;in, Input torque parameters to the steering column collected by the sensor; The feedback angle deviation change rate parameter is the controlled angle variable. The unit is W; under the condition of high-frequency periodic excitation of the road surface, the mechanical elastic hysteresis and damping characteristics inside the transmission chain cause a phase difference of about 90° between the torque of the forced vibration of the steering column and the rate of change of the feedback angle deviation. The product of the two changes alternately between positive and negative within a single vibration cycle; the arbitration logic adds up the characteristic power scalars continuously collected within the discrete time window and averages them according to the number of samplings. The resulting cumulative average value approaches 0 and remains within the dynamic critical energy boundary operator; the duration of the positive half-wave of the high-frequency oscillation is usually less than 5ms, which cannot reach the threshold of the continuous sliding window for 10 consecutive sampling cycles, and therefore will not trigger the transfer of control.

[0031] The dynamic boundary calculation unit, based on the intrinsic damping parameters and intrinsic inertia parameters of the transmission corresponding to the position closed-loop circuit, calls upon the reference dissipation power constant, damping adaptive compensation coefficient, and inertia offset correction coefficient stored in a non-volatile storage medium to generate a dynamic critical energy boundary operator in real time for offsetting intrinsic losses of the mechanical transmission. Its expression is as follows: ;in, This is the reference power dissipation constant stored in a non-volatile storage medium; For the parameter of the rate of change of feedback angle deviation The steering angle and acceleration parameters are obtained by performing a first-order discrete difference operation. The damping adaptive compensation coefficient; The inertial offset correction coefficient is used; therefore, both the damping compensation term and the inertial correction term are power quantities and can be added to the reference dissipation power constant. During offline calibration, the steering transmission structure is placed in a standard disconnected operating condition, the position closed-loop circuit stops applying drive output to the steering transmission structure, and only a frequency sweep signal is input to the steering transmission structure. By measuring the transmission intrinsic energy loss gradient under different angular velocities and steering angular accelerations, a mapping model of angular velocity and steering angular acceleration is established, and the corresponding damping adaptive compensation coefficient is calculated. With inertial hedging correction factor Both are then permanently stored in a non-volatile storage medium.

[0032] When the characteristic power scalar The operator continuously exceeds the dynamic critical energy boundary within a preset continuous discrete sampling period window. When the state machine trigger determines that the external manual takeover intention is valid, it triggers the transfer of control from the target trajectory signal tracking state to the external command following state; during this period, the control weight allocation function is based on... Exceed The power overflow difference is used to update the control weight ratio in a non-linear incremental manner, achieving a smooth switch between two control states. This non-linear incremental function adopts a smooth exponential curve response mechanism, and the control weighting factor ranges from 0 to 1. When the power overflow difference is between 0 and 5W, the control weighting factor increases exponentially with the power overflow difference. When the power overflow difference exceeds 5W, the control weighting factor smoothly converges to the upper limit of 1, enabling the control weight to complete a smooth switch within 10ms. In the external command follow state, the state observation unit continuously monitors the characteristic power scalar. ;when The power dissipation constant remains below the reference dissipation constant throughout the reset time window. At this time, the system collects the current real-time physical position parameters of the steering and uses them as the initial state variables of the position closed loop for synchronous alignment, so that the controlled angle variable reconverges to the vehicle target trajectory signal within the predetermined tracking deviation range. Then, it exits the external command following state and completes the closed loop reconstruction under the silent tracking condition. The reset time window duration is fixed at 100ms. In order to prevent the characteristic power scalar from fluctuating slightly near the critical value and causing high-frequency oscillation of the control state, the exit judgment adopts a hysteresis interval of 0.05W. Only when the characteristic power scalar is continuously lower than the difference between the reference dissipation power constant and the hysteresis interval is the exit condition considered to be met.

[0033] Example 3: This embodiment uses an experimental vehicle equipped with a steer-by-wire system. The test site is a standardized verification track containing random road surface excitations to simulate the steering conditions of the vehicle during continuous high-frequency obstacle avoidance in complex urban road environments. The experimental group uses the steering cooperative control method of this application, while the control group uses a fixed torque threshold switching scheme on the same physical platform. The experimental group independently reads the vehicle target trajectory signal through the position closed loop, runs dynamic arbitration logic, and sets three road surface excitation conditions: light, medium, and heavy. Input torque parameters of different amplitudes are formed by random road surface inputs. Under the light disturbance condition, the peak fluctuation range of the input torque parameter is 1.5 N·m to 2.0 N·m. The dynamic arbitration logic calculates the characteristic power scalar based on the input torque parameter and the feedback angle deviation change rate parameter, and compares it with the dynamic critical energy boundary operator for each sampling period. Under this condition, the characteristic power scalar is always lower than the dynamic critical energy boundary operator and does not reach the continuous sliding window threshold, so the system maintains the target trajectory signal tracking state. During the experiment, the average fluctuation of the feedback angle deviation change rate parameter was measured to be 0.05 rad / s, and no control transfer occurred.

[0034] Under moderate disturbance conditions, the peak fluctuation range of the input torque parameter is 3.5 N·m to 4.5 N·m. The dynamic boundary calculation unit calls the damping adaptive compensation coefficient and inertial offset correction coefficient, which are calculated offline and stored in a fixed manner. Based on the real-time collected feedback angle deviation change rate parameter and the steering angle acceleration parameter obtained from its first-order discrete difference, the damping compensation term and the inertial correction term are updated, so that the dynamic critical energy boundary operator increases with the current transmission state. After continuous comparison of this boundary, the periodic parasitic disturbance does not meet the continuous exceedance condition, and the average fluctuation of the feedback angle deviation change rate parameter in the experimental group remains within 0.12 rad / s. The control group uses a fixed torque threshold, which is frequently exceeded by the disturbance torque under the same conditions, resulting in 4 unexpected switching of the control state and obvious [significant changes]. Steering feedback position step; under severe disturbance conditions, the peak value of the input torque parameter exceeds 8.0 N·m; the characteristic power scalar obtained by multiplying the input torque parameter and the feedback angle deviation change rate parameter continuously exceeds the dynamic critical energy boundary operator within the continuous discrete sampling period. After reaching the continuous sliding window threshold, the state machine trigger determines that the external manual takeover intention is valid and triggers the control right to migrate from the target trajectory signal tracking state to the external command following state; the control weight allocation function calculates the takeover weighting factor based on the power overflow difference, gradually reduces the control weight ratio of the target trajectory signal tracking state within the switching period, and simultaneously increases the control weight ratio of the external command following state. As can be seen from the steering angle output curve, the test group maintains continuous change during the control state switching process without any step change.

[0035] After the input torque parameters return to normal, the state observation unit continuously monitors the characteristic power scalar. When this scalar remains below the reference dissipation power constant within the reset time window, the system acquires the current real-time steering physical position parameters and uses them as the initial state variables for synchronous alignment in the position closed-loop circuit. Then, the controlled angle variable is made to converge again to the vehicle target trajectory signal. The maximum transient value of the feedback angle deviation measured in the test group when the feedback loop is reclosed is 0.08°. The control group, which did not perform real-time steering physical position parameter synchronization, achieved a maximum transient value of 0.45° for the feedback angle deviation, which took 50.2ms. The control response lag is observed. Continuous test data from three disturbance conditions show that under mild and moderate disturbances, continuous comparison between the characteristic power scalar and the dynamic critical energy boundary operator can avoid periodic road excitation triggering control transfer. When the input torque parameter continuously increases and the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator, the control can smoothly transfer to the external command following state. After the external input is restored, by synchronizing the real-time physical position parameters of the steering, the feedback angle deviation is reduced from 0.45° in the control group to 0.08°, while avoiding the 50.2ms control response lag observed in the control group.

[0036] Example 4: In this embodiment, when the vehicle travels at high speed through continuous curves, the steering system operates in silent tracking mode. The position closed-loop circuit independently processes the vehicle target trajectory signal and drives the steering actuator to adjust the controlled angle variable. To address sensor sampling noise and mechanical transmission lag under strong torque disturbances, the system's underlying arbitration logic employs an implicit negative event recording injection mechanism to collaboratively correct the feedback angle deviation change rate parameter. The arbitration logic collects the input torque parameter and the feedback angle deviation change rate parameter in real time and calculates the characteristic power scalar. And identify periodic high-frequency disturbances by their changing states; when the characteristic power scalar within a sampling period A negative offset occurs, and the offset magnitude exceeds the preset power fluctuation threshold. At that time, the system writes the fluctuation amplitude of the input torque parameter, the rate of change of the feedback angle deviation parameter, and the instantaneous current vector state at that time node as a set of implicit negative event data into a circular queue; a preset power fluctuation threshold is set. Based on the offline measurement of the static damping dissipation power of the steering column under no external force intervention, this embodiment takes 0.2W; when the absolute value of the negative power offset is greater than 0.2W, data recording is started.

[0037] During subsequent control cycles, the system divides the input torque parameter fluctuation amplitude, feedback angle deviation change rate parameter, and instantaneous current vector state in the queue by their respective preset nominal rated reference values ​​to form a dimensionless data matrix, and calculates the independent statistical mean and statistical variance of each type of data. Simultaneously, the arbitration logic re-obtains the corresponding characteristic power scalar based on the input torque parameter and feedback angle deviation change rate parameter at each time point, extracts the negative power offset amplitude sequence, and uses the statistical mean of this sequence... and statistical variance It represents the concentration and volatility of negative disturbances within the current time window; as the historical records in the circular queue increase, the system updates the weights of each statistic according to the sliding time window, so that data from more recent sampling periods remain valid in the current correction process.

[0038] When performing deviation correction, the system calculates the statistical mean. With the current characteristic power scalar The difference between the negative offset amplitudes is calculated, and the absolute value of this difference is multiplied by the weighted compensation coefficient. The torque correction amount is obtained. Weighted compensation coefficient The torque correction is calculated by converting each 1W power difference to a torque of 0.05 N·m. Maintain N·m dimensions; then adjust the torque correction. The feedforward control branch of the position closed loop is injected to compensate for the hysteresis offset caused by mechanical transmission backlash and transmission inertia. After the collaborative correction of the feedback angle deviation change rate parameter is completed, the control transfer is still determined by the continuous comparison result between the characteristic power scalar and the dynamic critical energy boundary operator. Only when the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator within the continuous discrete sampling period will the control transfer be triggered from the target trajectory signal tracking state to the external command following state. Under the test condition of high-speed cornering accompanied by road surface micro-vibration, without the collaborative correction algorithm, the transient angle deviation response delay time of the steering actuator after receiving the interference signal is 45.2ms. After the above processing, the response delay time is shortened to 12.8ms, and the average step amplitude of the control command output of the steering actuator is reduced by 15.6%. The implicit negative event recording injection mechanism converts discrete disturbance data into torque correction amount, and compensates for the mechanical transmission hysteresis in the position closed loop without changing the control transfer determination path.

[0039] Example 5: This embodiment combines Figures 1 to 2 This paper describes a steering coordination control method in a silent steering wheel mode, such as... Figure 1As shown, in step S101, the controlled angle variable is first adjusted in a closed loop to follow the vehicle target trajectory signal; then, in step S102, the collected input torque parameter is multiplied by the feedback angle deviation change rate parameter to generate a characteristic power scalar; next, in step S103, various adjustment terms are accumulated based on the transmission intrinsic damping and inertia parameters to calculate the dynamic critical energy boundary operator; finally, in step S104, the characteristic power scalar is compared with the dynamic critical energy boundary operator. If the characteristic power scalar continuously exceeds the operator, the control is triggered to migrate to the external command following state.

[0040] like Figure 2 As shown, the horizontal axis represents time, with scale values ​​of 0, 20, 40, 60, 80, and 100. The vertical axis represents characteristic power, with scale values ​​of -4, -2, 0, 2, 4, and 6 from bottom to top. The dynamic critical energy boundary operator, indicated by dashed lines, lies between 3W and 4W across the entire time axis. The characteristic power scalar, indicated by solid lines, exhibits alternating positive and negative signs in the time domain, with peaks below 2W and troughs above -2W. The curve of the characteristic power scalar is completely confined within the curve of the dynamic critical energy boundary operator throughout the entire 100ms time span. This reflects that when facing external periodic high-frequency disturbances, the input torque parameter and the feedback angle deviation rate parameter exhibit alternating positive and negative oscillation characteristics in the time domain. The cumulative value of their product after being averaged in the discrete time domain approaches 0 and is confined within the dynamic critical energy boundary operator, thereby avoiding triggering the transfer of control to the external command follow state.

[0041] Example 6: In this embodiment, when the autonomous vehicle performs dynamic obstacle avoidance on an urban ring road, the steering system operates in a silent tracking mode. The position closed-loop circuit independently processes the vehicle target trajectory signal and adjusts the controlled angle variable to follow the signal. To address complex electromagnetic interference and sensor data drift, this embodiment determines the initial state of the system through offline calibration and performs frequency sweep calibration on the parameters used by the dynamic critical energy boundary operator. When the vehicle is stationary and the steering mechanism is not subjected to external forces, the zero-position torque offset calibration of the steering column is performed first. The test platform monitors the minute fluctuations of the controlled angle variable using a high-precision displacement sensor with a sampling rate of 2kHz. The input torque parameters are continuously recorded over 60 seconds. The zero-point torque offset reference value is determined based on the stable values ​​of the input torque parameters during the sampling period. Simultaneously, a first-order discrete difference operation is performed on the input torque parameters to obtain the torque change slope index. Both calibration results are stored in non-volatile storage. When the torque change slope index exceeds its reference range in the zero-point state for three consecutive sampling periods, and the controlled angle variable exhibits synchronous fluctuations, the arbitration logic determines that the input torque parameters have zero-point drift and initiates offset compensation. During offset compensation, the difference between the current input torque parameter and the zero-point torque offset reference value is calculated. The difference is then injected into the position closed loop as negative feedback to counteract the zero-position offset of the input torque parameter.

[0042] During the parameter calibration of the dynamic critical energy boundary operator, the steering transmission structure is placed under standard disconnection conditions. Sinusoidal excitation torque signals with frequencies ranging from 0.5Hz to 10Hz are input to the steering column, and the corresponding angular velocity, steering angular acceleration, and energy dissipation power are collected. Based on the collected results, the amplitude-frequency characteristic curve of the steering transmission structure is fitted to obtain the intrinsic energy loss gradient of the transmission under different angular velocities and steering angular accelerations. A mapping model of angular velocity and steering angular acceleration is then established. In the mapping model, the local slope at each inflection point of the energy dissipation power fitting curve with respect to angular velocity is extracted and used as the damping adaptive compensation coefficient. Its dimensions are equivalent to W / (rad / s); the local slope at each inflection point of the fitted curve of energy dissipation power with respect to steering angle acceleration is extracted and used as the inertial heave correction coefficient. Its dimensions are equivalent to W / (rad / s²).

[0043] The dynamic critical energy boundary operator is formed by the coefficients corresponding to each inflection point. The correction curve changes with the steering motion state, and the damping adaptive compensation coefficient is included. With inertial hedging correction factor The data is stored in a non-volatile storage medium. During operation, the dynamic boundary calculation unit calls the corresponding coefficients and combines them with the feedback angle deviation change rate parameter and steering angle acceleration parameter to generate a dynamic critical energy boundary operator, making it match the intrinsic mechanical transmission loss of the current steering transmission structure. Test records show that after completing the initial state calibration and offline calibration, the steering actuator has an accuracy rate of over 99% in distinguishing characteristic power scalars when facing high-frequency road disturbances. After bias compensation, the false triggering rate caused by sensor zero-point drift is reduced by 4 percentage points compared with the uncalibrated state. In the closed-loop reconstruction stage, the overshoot during the steering angle return process is controlled within 0.03° by using the damping adaptive compensation coefficient and inertial offset correction coefficient calibrated under various working conditions.

[0044] Example 7: In this embodiment, after the control weight ratio completes a smooth switch and enters the external command follow state, the state observation unit continuously collects characteristic power scalars according to the discrete control cycle. The amplitude changes are recorded in the sampling order within the reset time window; the reset time window is set to 100ms, and a hysteresis interval of 0.05W is used for exit judgment; when the characteristic power scalar of each sampling period within the window is lower than the reference dissipation power constant... When the difference with the hysteresis interval is reached, it is confirmed that the external control energy input has dissipated, triggering the system to exit the external command following state. Before exiting, the system collects the current real-time physical position parameters of the steering mechanism and writes these parameters into the position closed loop as the initial state variable when restarting the target trajectory signal tracking, so that the internal state of the position closed loop is synchronized with the current physical position of the steering mechanism. After the synchronization is completed, the position closed loop uses the target angle deviation value between the current steering angle feedback parameter and the vehicle target trajectory signal as the adjustment basis to generate a control output, so that the controlled angle variable reconverges to the vehicle target trajectory signal within the predetermined tracking deviation range. This process avoids directly using the historical state before exiting when the position closed loop recovers, thereby reducing the sudden change in steering angle when the feedback loop is closed again.

Claims

1. A steering coordination control method in a silent steering wheel mode, characterized in that, Includes the following steps: Step S101: Under the silent tracking condition with the input target trajectory signal, acquire the vehicle target trajectory signal at the angle input end, and use the position closed loop to adjust the controlled angle variable to follow the vehicle target trajectory signal. Step S102: Collect the input torque parameter and feedback angle deviation change rate parameter during the controlled angle variable follow-up adjustment process, multiply the input torque parameter and the feedback angle deviation change rate parameter to generate a characteristic power scalar that characterizes the external control energy input rate; Step S103: Based on the transmission intrinsic damping parameters and transmission intrinsic inertia parameters corresponding to the position closed loop, accumulate the product of the absolute value of the reference dissipation power constant, the absolute value of the feedback angle deviation change rate parameter and the damping adaptive compensation coefficient, and the absolute value of the steering angle acceleration parameter and the inertia counterbalancing correction coefficient stored in the non-volatile storage medium to generate a dynamic critical energy boundary operator for counterbalancing the intrinsic losses of mechanical transmission. The steering angle acceleration parameter is generated by performing a first-order discrete difference operation on the feedback angle deviation change rate parameter. Step S104: Compare the characteristic power scalar with the dynamic critical energy boundary operator. When the characteristic power scalar continuously exceeds the dynamic critical energy boundary operator within the range of continuous discrete sampling periods, the control is triggered to shift from the target trajectory signal tracking state to the external command following state.

2. The steering coordination control method in a steering wheel silent mode according to claim 1, characterized in that, The step S103, which generates the dynamic critical energy boundary operator for offsetting intrinsic losses in mechanical transmission, includes the following sub-steps: Step S1031, performing a first-order discrete difference operation on the feedback angle deviation rate of change parameter to generate the steering angle acceleration parameter; Step S1032, multiplying the absolute value of the feedback angle deviation rate of change parameter by the damping adaptive compensation coefficient to generate the damping compensation term; Step S1033, multiplying the absolute value of the steering angle acceleration parameter by the inertial offset correction coefficient to generate the inertial correction term; Step S1034, adding the reference dissipation power constant, the damping compensation term, and the inertial correction term to output the dynamic critical energy boundary operator.

3. The steering coordination control method in a steering wheel silent mode according to claim 2, characterized in that, The damping adaptive compensation coefficient and the inertial offset correction coefficient are obtained through the following steps: Step S1035, in offline state, input the frequency domain sweep signal under the standard disconnection condition to the steering transmission structure, and measure the transmission intrinsic energy loss gradient of the steering transmission structure under different angular velocities and steering angular accelerations; Step S1036, based on the transmission intrinsic energy loss gradient, establish a mapping model of angular velocity and steering angular acceleration, calculate the corresponding damping adaptive compensation coefficient and inertial offset correction coefficient, and store them in a non-volatile storage medium.

4. The steering coordination control method in a steering wheel silent mode according to claim 1, characterized in that, Step S104 includes the following sub-steps: Step S1041, within the continuous discrete sampling period, the statistical characteristic power scalar exceeds the number of discrete samples of the dynamic critical energy boundary operator; Step S1042, when the number of discrete samples reaches a predetermined continuous sliding window threshold, the control is triggered to shift from the target trajectory signal tracking state to the external command following state.

5. The steering coordination control method in a steering wheel silent mode according to claim 1, characterized in that, After triggering the transfer of control from the target trajectory signal tracking state to the external command following state, the following sub-steps are included: Step S1043, calculate the power overflow difference of the characteristic power scalar exceeding the dynamic critical energy boundary operator; Step S1044, use a nonlinear increasing function about the power overflow difference to solve the takeover weighting factor; Step S1045, based on the takeover weighting factor, reduce the control weight ratio in the target trajectory signal tracking state during the switching cycle, and simultaneously increase the control weight ratio in the external command following state to complete the switch of control state.

6. The steering coordination control method in a steering wheel silent mode according to claim 5, characterized in that, After the control weight ratio completes a smooth switch and enters the external command follow state, the following steps are included: Step S105, within a predetermined reset time window, continuously monitor the amplitude change trend of the characteristic power scalar; Step S106, when the characteristic power scalar is always lower than the reference dissipation power constant within the reset time window, trigger the system to exit the external command follow state.

7. The steering coordination control method in a steering wheel silent mode according to claim 6, characterized in that, Step S106 includes the following sub-steps: Step S1061, collect the current real-time physical position parameters of the steering; Step S1062, use the real-time physical position parameters of the steering as the initial state variable of the position closed loop for synchronous alignment, so that the controlled angle variable reconverges to the vehicle target trajectory signal within the predetermined tracking deviation range.

8. The steering coordination control method in a steering wheel silent mode according to claim 1, characterized in that, In step S102, when facing external periodic high-frequency disturbance conditions, the input torque parameter and the feedback angle deviation change rate parameter exhibit alternating positive and negative signs in the time domain. The cumulative value of their product after being added and averaged in the discrete time domain approaches 0 and is restricted within the dynamic critical energy boundary operator.

9. The steering coordination control method in a steering wheel silent mode according to claim 1, characterized in that, In step S101, the calculation process of the position closed-loop following adjustment includes the following sub-steps: Step S1011, acquire the vehicle target trajectory signal with a fixed discrete control cycle; Step S1012, calculate the target angle deviation value between the current steering angle feedback parameter and the vehicle target trajectory signal; Step S1013, calculate the target angle deviation value using a preset proportional-integral-derivative control algorithm, and generate a control output quantity for driving and adjusting the steering non-electrical variables.

Images