Angle control method for smooth angle response and rapid intervention

Abstract

The invention relates to the technical field of steering systems, in particular to an angle control method for smooth angle response and rapid intervention. An angle control method for smooth angle response and rapid intervention comprises an angular velocity feed-forward PID algorithm, and the angular velocity feed-forward PID algorithm comprises the steps that the difference between a request angle and an actual angle is obtained to obtain an angle difference, the angle difference is subjected to the PID algorithm to obtain target motor torque, the target motor torque is converted into actually-executed motor torque through gradient output, and a steering wheel is controlled to rotate. The method is characterized in that the angle smooth response comprises a low-frequency compensation strategy and a high-frequency compensation strategy. Compared with the prior art, the angle control method for smooth angle response and rapid intervention is provided, the torque of the motor is calculated by adopting an angular velocity feed-forward PID algorithm, a low-frequency and high-frequency compensation strategy and a rapid intervention strategy are introduced, and the purpose of controlling a steering wheel to smoothly and rapidly reach a target angle can be achieved.

Description

Technical Field

[0001] The invention relates to the technical field of steering systems, and in particular to an angle control method with smooth angle response and rapid intervention. Background Art

[0002] At present, autonomous driving technology has been widely used in the automotive field. The lateral control of autonomous driving is mainly steering. The steering end often designs an angle control system based on the electronic power steering system as a platform, and completes the steering by responding to the angle request of the autonomous driving end. From the perspective of stability, the autonomous driving end does not want the angle control function to be easily interrupted by the outside world and exit. The vehicle end often requires that the angle control function will exit only when the manual torque value and duration reach the set threshold at the same time, and requires the electric power steering system end not to respond to the driver's input when the manual torque value is small so that manual intervention cannot easily affect the path planning of the vehicle end. However, when the autonomous driving system cannot cope with complex or sudden situations, there are risks such as the driver's inability to move the steering wheel when taking over the steering wheel quickly, and the angle control function exits unevenly, which can easily cause driving panic and the inability to avoid obstacles in time when problems occur in the autonomous driving planning.

[0003] The traditional angle control system uses the angular velocity feedforward PID algorithm as the main algorithm to calculate the motor torque. One part is to make a difference between the requested angle and the actual angle to obtain the angle difference, and the angle difference is used to obtain the target motor torque through the PID algorithm. The other part is to make a difference between the actual angular velocity and 0, and multiply it by the corresponding coefficient to obtain a part of the motor torque. The two are added together to form the basic torque for angle control. When the vehicle is in place or driving at low speed, there is a large resistance torque due to the large resistance between the tire and the ground, the cancellation structure of the vehicle and the large longitudinal force return torque. When the vehicle end requests a low-frequency and small angle, the motor torque calculated by the PID algorithm in a short time is small, causing the motor end to be unable to overcome the resistance, resulting in the steering wheel being unable to turn or turning slowly. Since the angle difference always exists, the integral term continues to accumulate, and after a period of time, the motor torque increases again, driving the steering wheel to run quickly. When the target angle is reached, the above actions are repeated, resulting in the problem of the steering wheel running slowly first and then fast. However, when the vehicle requests a high frequency and large angle, the starting angle difference is large, the calculated motor torque is large, and the actual angular velocity is large, resulting in a faster steering wheel rotation speed and a large acceleration at the start time. However, when the target angle is about to be reached, the calculated target torque begins to decrease. The increase in the angle causes the steering system's return force and mechanical friction to increase, resulting in a decrease in the steering wheel rotation speed and acceleration in a short period of time. This repeated process causes the steering wheel to get stuck, first fast and then slow down.

[0004] The above reasons can easily cause the steering wheel to run unevenly, affecting the driving experience. Based on this, this paper proposes an angle smoothness response and driver rapid intervention angle control system to solve the above problems and improve the driving experience. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides an angle control method with smooth angle response and rapid intervention, which adopts an angular velocity feedforward PID algorithm to calculate the motor torque, and introduces low-frequency and high-frequency compensation strategies, as well as a rapid intervention strategy, so as to control the steering wheel to smoothly and quickly reach the target angle.

[0006] In order to achieve the above purpose, a method for angle control with smooth angle response and rapid intervention is designed, including an angular velocity feedforward PID algorithm. The angular velocity feedforward PID algorithm is to obtain an angle difference by subtracting the requested angle from the actual angle. The angle difference is obtained by the PID algorithm to obtain the target motor torque. The target motor torque is converted into the actual motor torque through the gradient output to control the steering wheel rotation. The method is characterized in that the angle smooth response includes a low-frequency compensation strategy and a high-frequency compensation strategy.

[0007] The low frequency compensation strategy is as follows:

[0008] S11, setting the vehicle speed range to 0~5Km / h;

[0009] S12, calculating the angle difference between the requested angle and the actual angle, and then calculating the angle following coefficient σ1 according to the angle difference; S13, calculating the viscosity compensation coefficient σ2 according to the angular velocity;

[0010] S14, calculating the low-frequency compensation initial torque value LowFreqMotTorq according to the vehicle speed table;

[0011] S15, calculating the low-frequency compensation torque LowMotTorq and outputting it to the motor end with a gradient;

[0012] The high frequency compensation strategy is as follows:

[0013] S21, calculating the maximum limit of the virtual hand torque according to the vehicle speed table;

[0014] S22, calculate the virtual hand force gain coefficient according to the angular velocity by looking up the table

[0015] S23, calculate the virtual hand force gain coefficient according to the vehicle speed by looking up the table

[0016] S24, reading the actual hand torque value TBT measured by the electric power steering system;

[0017] The rapid intervention strategy is as follows:

[0018] S31, calculating the change rate of the real hand torque value TBT at adjacent n moments;

[0019] S32, determining whether the angular velocity is less than a boundary;

[0020] S33, calculating whether the accumulated value of the real hand torque value TBT within n moments is greater than a limit;

[0021] S34, when the conditions of step S32 and step S33 are met, the quick intervention hand torque value QuickGetTBT is calculated and sent to the basic power assist module after gradient processing to complete the steering operation.

[0022] The calculation formula of the target motor torque in the angular velocity feedforward PID algorithm is:

[0023] Among them, TarMotTorq is the target motor torque, TarAng is the requested angle, CurAng is the actual steering wheel angle, CurAngSpd is the actual steering wheel angular velocity, Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, and Kv is the coefficient of angular velocity participating in torque calculation.

[0024] The formula of the low-frequency compensation torque is LowMotTorq=LowFreqMotTorq*σ1*σ2; wherein, LowMotTorq is the low-frequency compensation torque, LowFreqMotTorq is the initial torque value of the low-frequency compensation, σ1 is the angle following coefficient, and σ2 is the viscosity compensation coefficient.

[0025] The calculation formula of the virtual hand torque value is: Among them, HighMotTorq is the high-frequency compensation torque, TBT is the actual hand torque value, is the virtual hand force gain coefficient with angular velocity, is the virtual hand force gain coefficient with vehicle speed.

[0026] The actual hand torque value TBT during the high-frequency compensation process is less than 1 Nm.

[0027] In order to ensure that the high-frequency compensation strategy does not conflict with the low-frequency compensation strategy, the virtual hand force gain coefficient involved in the high-frequency compensation strategy is When the angular velocity is small, the calibration is small, and the virtual hand force gain coefficient involved varies with the vehicle speed. The calibration is small at low speed; since the actual hand torque value TBT is relatively small during low-frequency motion, the target motor torque for the final angle following is Among them, TarMotTorq is the target motor torque, TarAng is the requested angle, CurAng is the actual steering wheel angle, CurAngSpd is the actual steering wheel angular velocity, Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, Kv is the coefficient of angular velocity participating in torque calculation, LowMotTorq is the low-frequency compensation torque, and HighMotTorq is the high-frequency compensation torque.

[0028] The calculation formula for the quick intervention hand torque value is QuickGetTBT=(abs(TBT)-δ)*ω*Sign(TBT); wherein QuickGetTBT is the quick intervention hand torque value, abs(TBT) is the absolute value of the hand force, δ is the set TBT value, ω is the coefficient obtained by looking up the table with the vehicle speed, and Sign(TBT) is the actual hand force direction.

[0029] In the step S34, a portion of the compensation virtual hand force value GetOutTBT is first calculated and given to the basic power assist module, and then the driver is continued to be supported in controlling the steering wheel. The specific process is as follows:

[0030] (1) Calculate the virtual hand force value VirtualTBT according to the vehicle speed table;

[0031] (2) Calculate the virtual compensation coefficient μ of the hand force by looking up the table according to the TBT value, and multiply the two to obtain GteOutTBT, that is, GetOutTBT = VirtualTBT*μ;

[0032] (3) GteOutTBT drops to 0 after a certain gradient. During the process of GteOutTBT gradient decreasing, the basic assistance is increasing.

[0033] Compared with the prior art, the present invention provides an angle control method with smooth angle response and rapid intervention, adopts an angular velocity feedforward PID algorithm to calculate the motor torque, and introduces low-frequency and high-frequency compensation strategies, as well as a rapid intervention strategy, which can control the steering wheel to smoothly and quickly reach the target angle. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of the present invention.

[0035] Figure 2 Flowchart of rapid intervention strategy.

[0036] Figure 3 This is a curve chart of the actual vehicle performance during rapid intervention.

[0037] Figure 4 This is the vehicle angle response performance curve. DETAILED DESCRIPTION

[0038] The present invention will be further described below with reference to the accompanying drawings.

[0039] like Figure 1 As shown, the control method of the present invention includes an angular velocity feedforward PID algorithm. The angular velocity feedforward PID algorithm is to obtain an angle difference by subtracting the requested angle from the actual angle. The angle difference is obtained by the PID algorithm to obtain a target motor torque. The target motor torque is converted into the actual motor torque through gradient output to control the steering wheel rotation.

[0040] In the control, the steering wheel is expected to reach the target angle. After reaching the target angle, the actual angular velocity is 0, so the actual angular velocity is introduced to participate in the torque calculation. The parameters involved in the PID algorithm and the angular velocity feedforward part are calculated according to the vehicle speed, and the P, I, and D items are all set with motor torque limits at different vehicle speeds to prevent the risk of vehicle instability caused by rapid rotation of the steering wheel due to oversaturation of motor torque at high speeds.

[0041] The calculation formula for the target motor torque in angle control is:

[0042] Among them, TarMotTorq is the target motor torque, TarAng is the requested angle, CurAng is the actual steering wheel angle, CurAngSpd is the actual steering wheel angular velocity, Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, and Kv is the coefficient of angular velocity participating in torque calculation.

[0043] The existing electric power steering system calculates the power-assisting torque based on the vehicle speed, hand torque value TBT, etc., which is called the basic power-assisting mode. The basic power-assisting mode achieves the steering requirement by controlling the motor.

[0044] The angle control function involved in the present invention can realize the decoupling of the basic power-assist mode and the angle control mode. When the angle control function is activated, if the hand torque size and the short-time integral of the hand torque do not meet the requirements, the steering is not changed in response to the hand torque value, thereby reducing the influence of the driver on the automatic driving.

[0045] The angle smoothness response processing strategy involved in the present invention is divided into two parts: low-frequency compensation strategy and high-frequency compensation strategy. The purpose of the low-frequency compensation strategy is to overcome the resistance caused by the friction between the tire and the ground, the vehicle kingpin structure and the longitudinal force return torque, the friction between the machinery itself, etc., so as to speed up the electric power steering system end to respond to the target angle requested by the vehicle end. The specific implementation is as follows: 1) Set the vehicle speed range Vspd: 0~5Km / h to ensure that the low-frequency compensation torque will be output only within the vehicle speed range, and the vehicle speed range can support calibration; 2) Calculate the difference between the requested angle and the actual angle, and calculate the coefficient σ1 according to the angle difference; 3) Calculate the coefficient σ2 according to the angular velocity; 4) Calculate the low-frequency compensation initial torque value LowFreqMotTorq according to the vehicle speed table (the table is the X-axis for the vehicle speed and the Y-axis for the motor execution torque); 5) Calculate the low-frequency compensation torque and output it to the motor end in a gradient.

[0046] The calculation formula of the low-frequency compensation torque is LowMotTorq=LowFreqMotTorq*σ1*σ2.

[0047] The purpose of the high-frequency compensation strategy is to provide an initial reaction force to improve the problem of high angular velocity and speed when the steering wheel starts to run in response to a large angle request. Since the angular velocity is large during a period of time in response to a large angle request, it will drag the TAS sensor in the opposite direction, causing the real hand torque value TBT to be large. Therefore, the real hand torque value TBT and angular velocity can be introduced into the compensation logic during high-frequency compensation. The specific implementation is as follows: 1) Calculate the maximum limit of the virtual hand torque by looking up the table with the vehicle speed; 2) Calculate the gain coefficient by looking up the table with the angular velocity (the table has the angular velocity on the X axis and the gain coefficient on the Y axis) 3) Calculate the gain coefficient by looking up the table according to the vehicle speed 4) Read the actual hand torque value TBT measured at the electric power steering system end; 5) In order to prevent the driver from accidentally triggering the steering wheel, the actual hand torque value TBT is required to be less than 1Nm during the high-frequency compensation process.

[0048] During the starting operation stage when the angle control system receives high-frequency angle requests, the actual hand torque value TBT is large, but the hand torque value collected by the sensor during the starting stage of automatic driving is in the opposite direction to the value collected by the driver operating the steering wheel. The virtual hand torque value HighMotTorq is calculated, and then the virtual hand torque is passed to the basic power assist module after being limited, and the real hand torque value TBT is used to look up the table to obtain the virtual power assist, so as to improve the steering wheel sticking problem caused by excessive torque and sudden acceleration at starting during the high-frequency angle request stage.

[0049] The calculation formula of the virtual hand torque value is:

[0050] To ensure that high-frequency logic does not conflict with low-frequency logic, the gain coefficient involved in the high-frequency logic part When the angular velocity is small, the calibration is small, and the gain coefficient involved is The calibration is small at low speeds, and because TBT is relatively small during low-frequency motion, according to the calculation formula for the target motor torque in angle control, the influence of the high-frequency logic part on the low-frequency logic is greatly reduced.

[0051] The formula for calculating the target motor torque required for the final angle following is:

[0052] like Figure 2 As shown in the figure, the quick intervention part calculates the virtual hand torque value QuickGetTBT according to the hand force value and vehicle speed and gives it to the basic power assist module to obtain power assist, helping the driver to change direction in the angle control mode. The specific implementation is as follows: 1) Calculate the change rate of the real hand torque value TBT at adjacent n moments; 2) Determine whether the angular velocity is less than the boundary to prevent the problem of the real hand torque value TBT being too large due to the excessive angular velocity when responding to the angle, and reduce the probability of false triggering; 3) Calculate whether the cumulative value of the real hand torque value TBT within n moments is greater than the boundary to prevent the problem of false triggering caused by the human hand accidentally touching the steering wheel. When all conditions are met, the quick intervention hand torque is calculated and given to the basic power assist module after gradient processing to complete the steering operation.

[0053] The calculation formula for the quick intervention hand torque value is QuickGetTBT = (abs(TBT)-δ)*ω*Sign(TBT). Among them, abs(TBT) is the absolute value of the hand force, δ is the set TBT value, ω is the coefficient obtained by looking up the table with the vehicle speed, and Sign(TBT) is the actual hand force direction. In traditional steering, the TAS sensor can collect a value, which is the hand force, abbreviated as TBT in English. abs(TBT) is the absolute value of TBT, that is, the absolute value of the hand force.

[0054] like Figure 3 The figure below shows the performance curve of the real vehicle during quick intervention. After the quick intervention flag is set, QuickGetTBT obtains the response value according to the calculation formula of the quick intervention hand torque value. When the vehicle angle request remains unchanged, the actual steering wheel angle can still be changed to help the driver cope with working conditions that the automatic driving cannot handle.

[0055] Functional safety in the angle control system requires that the angular velocity is less than the safety boundary value, and when the rapid intervention logic is in effect, the high-frequency compensation part stops working. When the rapid intervention part is working, since the system starts to provide basic assistance, the angular velocity at this time is likely to fail to meet the functional safety requirements, causing the rapid intervention part to stop working. If the angle control function exits at this time, the present invention can still continue to calculate a part of the compensation virtual hand force value GetOutTBT according to the calculation formula of the compensation virtual hand force value and give it to the basic assistance module to continue to support the driver to control the steering wheel, so as to prevent the rapid intervention part from immediately exiting and causing the hand force to increase sharply due to the slow increase in assistance and causing the two-stage hand feel problem. The specific implementation is as follows: 1) Calculate the virtual hand force value VirtualTBT according to the vehicle speed table; 2) Calculate the coefficient μ according to the TBT value table, and multiply the two to obtain GteOutTBT, that is, GetOutTBT=VirtualTBT*μ; 3) GteOutTBT drops to 0 after a certain gradient. During the process of the gradient decrease of GteOutTBT, the basic assistance is rising.

[0056] When the fast intervention logic is running, the table will be looked up according to the torque calculated by the current PID to obtain a larger attenuation coefficient α1 and multiply it with the torque calculated by the PID itself and transmit it to the motor execution end, weakening the PID torque's resistance to the human hand changing the steering wheel. When the fast intervention part does not cause the function to be exited, the human hand will change the steering wheel movement, resulting in a large angle difference. Because the integral term of the PID algorithm is not hit or accumulated, a larger torque value will be calculated at this moment, causing the steering wheel to run quickly, especially at high speeds, which will bring greater risks and hand-beating problems.

[0057] When the present invention determines that the rapid intervention condition does not cause the angle control function to exit, the rapid intervention part uses the motor torque calculated by PID to look up the table to calculate the attenuation coefficient α2, and then multiplies it with the torque calculated by PID to attenuate the torque calculated by PID, and the attenuation coefficient α1 is less than the attenuation coefficient α2. When it is detected that the difference between the requested angle and the actual execution angle is small, the motor torque attenuation strategy is exited to prevent the steering wheel from returning to the target angle position quickly due to the large angle difference, thereby reducing driving risks and improving the steering wheel hand-beating problem.

[0058] like Figure 4 As shown, this is a curve diagram of the vehicle angle response performance. According to the above theory, during the test, the actual response angle of the steering wheel of the vehicle is smooth and close to the target angle curve requested by the vehicle, meeting the requirements.

Claims

1. An angle control method with smooth angle response and rapid intervention, including an angular velocity feedforward PID algorithm, wherein the angular velocity feedforward PID algorithm is to obtain an angle difference by subtracting a requested angle from an actual angle, and obtain a target motor torque through a PID algorithm, and the target motor torque is converted into an actual motor torque through a gradient output to control the steering wheel rotation, characterized in that: The angle smooth response includes a low-frequency compensation strategy and a high-frequency compensation strategy. The low frequency compensation strategy is as follows: S11, setting the vehicle speed range to 0~5Km / h; S12, calculating the angle difference between the requested angle and the actual angle, and then calculating the angle following coefficient σ1 according to the angle difference; S13, calculating the viscosity compensation coefficient σ2 according to the angular velocity; S14, calculating the low-frequency compensation initial torque value LowFreqMotTorq according to the vehicle speed table; S15, calculating the low-frequency compensation torque LowMotTorq and outputting it to the motor end with a gradient; The high frequency compensation strategy is as follows: S21, calculating the maximum limit of the virtual hand torque according to the vehicle speed table; S22, calculate the virtual hand force gain coefficient according to the angular velocity by looking up the table S23, calculate the virtual hand force gain coefficient according to the vehicle speed by looking up the table S24, reading the actual hand torque value TBT measured by the electric power steering system; The rapid intervention strategy is as follows: S31, calculating the change rate of the real hand torque value TBT at adjacent n moments; S32, determining whether the angular velocity is less than a boundary; S33, calculating whether the accumulated value of the real hand torque value TBT within n moments is greater than a limit; S34, when the conditions of step S32 and step S33 are met, the quick intervention hand torque value QuickGetTBT is calculated and sent to the basic power assist module after gradient processing to complete the steering operation.

2. The angle control method with smooth angle response and rapid intervention according to claim 1, characterized in that: The calculation formula of the target motor torque in the angular velocity feedforward PID algorithm is: Among them, TarMotTorq is the target motor torque, TarAng is the requested angle, CurAng is the actual steering wheel angle, CurAngSpd is the actual steering wheel angular velocity, Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, and Kv is the coefficient of angular velocity participating in torque calculation.

3. The angle control method with smooth angle response and rapid intervention according to claim 1, characterized in that: The formula of the low-frequency compensation torque is LowMotTorq=LowFreqMotTorq*σ1*σ2; wherein, LowMotTorq is the low-frequency compensation torque, LowFreqMotTorq is the initial torque value of the low-frequency compensation, σ1 is the angle following coefficient, and σ2 is the viscosity compensation coefficient.

4. The angle control method with smooth angle response and rapid intervention according to claim 1, characterized in that: The calculation formula of the virtual hand torque value is: Among them, HighMotTorq is the high-frequency compensation torque, TBT is the actual hand torque value, is the virtual hand force gain coefficient with angular velocity, is the virtual hand force gain coefficient with vehicle speed.

5. The angle control method with smooth angle response and rapid intervention according to claim 1 or 4, characterized in that: The actual hand torque value TBT during the high-frequency compensation process is less than 1 Nm.

6. The angle control method with smooth angle response and rapid intervention according to claim 1 or 4, characterized in that: In order to ensure that the high-frequency compensation strategy does not conflict with the low-frequency compensation strategy, the virtual hand force gain coefficient involved in the high-frequency compensation strategy is When the angular velocity is small, the calibration is small, and the virtual hand force gain coefficient involved varies with the vehicle speed. The calibration is small at low speed; since the actual hand torque value TBT is relatively small during low-frequency motion, the target motor torque for the final angle following is Among them, TarMotTorq is the target motor torque, TarAng is the requested angle, CurAng is the actual steering wheel angle, CurAngSpd is the actual steering wheel angular velocity, Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, Kv is the coefficient of angular velocity participating in torque calculation, LowMotTorq is the low-frequency compensation torque, and HighMotTorq is the high-frequency compensation torque.

7. The angle control method with smooth angle response and rapid intervention according to claim 1, characterized in that: The calculation formula for the quick intervention hand torque value is QuickGetTBT=(abs(TBT)-δ)*ω*Sign(TBT); wherein QuickGetTBT is the quick intervention hand torque value, abs(TBT) is the absolute value of the hand force, δ is the set TBT value, ω is the coefficient obtained by looking up the table with the vehicle speed, and Sign(TBT) is the actual hand force direction.

8. The angle control method with smooth angle response and rapid intervention according to claim 1, characterized in that: In the step S34, a portion of the compensation virtual hand force value GetOutTBT is first calculated and given to the basic power assist module, and then the driver is continued to be supported in controlling the steering wheel. The specific process is as follows: (1) Calculate the virtual hand force value VirtualTBT according to the vehicle speed table; (2) Calculate the virtual compensation coefficient μ of the hand force by looking up the table according to the TBT value, and multiply the two to obtain GteOutTBT, that is, GetOutTBT = VirtualTBT*μ; (3) GteOutTBT drops to 0 after a certain gradient. During the process of GteOutTBT gradient decreasing, the basic assistance is increasing.

Images