Pivot steering accurate control method based on active disturbance rejection

Abstract

The invention discloses an in-situ steering accurate control method based on active disturbance rejection. The method comprises the steps that firstly, yaw velocity under different road conditions is determined; 2, setting an initial starting torque, and adjusting the starting torque by updating an adhesion coefficient to obtain a target starting torque; 3, on the basis of an active disturbance rejection control theory, an LADRC control model is constructed by using the yaw motion characteristics of the vehicle to obtain an active yaw moment; and step 4, establishing a pivot steering lower-layer torque optimal distribution control system: taking a longitudinal speed as a state quantity and setting a target value as 0, determining an optimal wheel end torque for tracking an active yaw moment, obtaining an optimal control quantity based on a control quantity target by using a model predictive control method, and performing optimal control. And the torque output of the hub motor is controlled according to the generated control quantity motor torque instruction value, so that the purpose of driving skid resistance is achieved. The stability of pivot steering of the whole vehicle can be remarkably enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle dynamics control and distributed drive systems, and particularly relates to a precise in-situ steering control method based on active disturbance rejection. Background Technique

[0002] In the field of vehicle control, the optimization of in-situ steering control strategies is extremely crucial. During in-situ steering, the coupling of the forces between the vehicle tires and the ground is complex. It is necessary to ensure steering accuracy while also taking into account factors such as yaw stability, resistance torque calculation, power distribution, and tire models. Traditional control methods for dealing with such non-linear, time-varying, and easily disturbed systems have the drawbacks of insufficient control accuracy and excessive dependence on accurate models.

[0003] Among common in-situ steering control strategies, wheel speed control realizes steering by adjusting the speed difference between the two sides of the wheels. It is easy to operate, but has high requirements for vehicle symmetry. In special road conditions such as low friction and unstable slopes, it is difficult to provide sufficient steering torque, resulting in difficult in-situ steering. In contrast, torque control is more flexible, does not depend on the ground friction coefficient and tire symmetry, can adjust the torque as needed, and adapt to different turning radii and road conditions, but there are challenges in accurately estimating the resistance torque and yaw torque, adapting to complex roads, and stability control.

[0004] With the development of distributed drive technology, some research has turned to four-wheel independently driven vehicles. By leveraging the advantages of independent torque control and the skid-steering principle of tracked vehicles, in-situ steering is achieved by applying a yaw torque. Algorithms such as integral sliding mode control and single neuron adaptive PID control are used to solve problems such as time delay in tracking vehicle speed and steering angular velocity, and suppressing the deviation of the steering center. However, there are still problems such as the difficulty in accurately calculating the total vehicle resistance torque and yaw torque during in-situ steering of the tires, weak adaptability to unknown roads, and unstable steering process. Summary of the Invention

[0005] The purpose of the present invention is to provide a precise in-situ steering control method based on active disturbance rejection, aiming to improve the tracking accuracy of the desired yaw angular velocity and the overall vehicle stability during in-situ steering, and ensure that the vehicle accurately executes the driver's steering intention under unknown road conditions. At the same time, slip ratio control is introduced to prevent excessive wheel slip, maintain the vehicle attitude and dynamic balance during steering, and reduce the risk of loss of control.

[0006] The technical solution for realizing the purpose of the present invention is as follows:

[0007] A precise in-situ steering control method based on active disturbance rejection, comprising:

[0008] Step 1. Determine the yaw rate under different road surface conditions: Based on the relationship between the vehicle's yaw angular acceleration and slip ratio under different adhesion coefficients obtained under different road surface conditions, obtain the extreme point angular acceleration; consider the actual driver's pedal input and the target yaw rate obtained by combining the extreme point angular acceleration.

[0009] Step 2. Design an in-situ steering start strategy: including setting an initial starting torque, and adjusting the starting torque through the update of the adhesion coefficient to obtain the target starting torque.

[0010] Step 3. Design a total yaw moment decision-making method based on active disturbance rejection control: Based on the active disturbance rejection control theory, utilize the yaw motion characteristics of the vehicle to construct an LADRC control model to obtain the active yaw moment M s ;

[0011] Step 4. Establish an in-situ steering lower-layer torque optimization distribution control system: Take the longitudinal speed as the state variable and set the target value to 0, determine the optimal wheel-end torque for tracking the active yaw moment M s ; use the model predictive control method to obtain the optimal control quantity u based on the control quantity u r target, and control the torque output of the in-wheel motor according to the generated control quantity motor torque command value to achieve the purpose of driving anti-skid.

[0012] Compared with the prior art, the remarkable advantages of the present invention are:

[0013] (1) High-precision and stable control: Build an in-situ steering control system based on active disturbance rejection, accurately track the desired yaw rate, greatly improve the control accuracy, and significantly enhance the stability of the vehicle's in-situ steering.

[0014] (2) Strong adaptability to complex road conditions: Integrate the expected yaw rate input by the driver, and generate an expected yaw rate curve by means of starting torque decision-making and road surface yaw rate threshold estimation, so that the vehicle can accurately respond to the driver's steering intention under unknown road conditions, and greatly improve the adaptability to complex road conditions.

[0015] (3) Omnidirectional safety guarantee: Determine the wheel slip ratio threshold according to the road surface adhesion coefficient, use active disturbance rejection control to obtain the wheel torque threshold, and suppress the excessive slip of the tire through the lower-layer model predictive control; at the same time, in terms of tire force, the method of the present invention makes the longitudinal force rise smoothly with small oscillation, the vehicle's yaw rate, resistance moment and tire side force are more stable, reduces the yaw oscillation, and comprehensively guarantees the vehicle driving safety and steering dynamic stability. Description of the Drawings

[0016] Figure 1 is the framework of the vehicle in-situ steering system based on active disturbance rejection control of the present invention.

[0017] Figure 2The theoretical variation curve of yaw angular acceleration with slip ratio under different adhesion coefficients.

[0018] Figure 3 The starting torque decision-making process for in-situ steering.

[0019] Figure 4 The two-degree-of-freedom closed-loop control system.

[0020] Figure 5 (a-b) The simulation results of yaw angular velocity and center-of-mass trajectory on high-adhesion road surface.

[0021] Figure 6 (a-b) The simulation results of yaw angular velocity and center-of-mass trajectory on low-adhesion road surface. Detailed implementation manner

[0022] The present invention will be further described in detail below with reference to the accompanying drawings. The specific embodiments described herein are merely used to explain the present invention and are not used to limit the present invention.

[0023] As Figure 1 shown, a precise in-situ steering control method based on active disturbance rejection provided in this embodiment includes the following steps:

[0024] Step 1: Determine the yaw angular velocity under different road surface conditions.

[0025] To accurately grasp the motion characteristics of the vehicle under different road surface conditions and determine the yaw angular velocity under different road surface conditions, it is necessary to accurately grasp the peak yaw angular acceleration. The yaw angular acceleration refers to the maximum yaw angular acceleration value that the vehicle can reach under different road surface adhesion conditions (such as high-adhesion road surface, low-adhesion road surface, etc.). The following is solved by torque. The in-situ steering motion of the whole vehicle can be based on the active yaw moment M s and the resistance moment M f is defined. For the yaw motion of the whole vehicle, the active yaw moment M s is generated by the longitudinal forces of each wheel, and the resistance moment M f is generated by the lateral forces of each wheel. Among them, the tire forces are obtained by the Dugoff formula:

[0026]

[0027] In the formula: F x_ij represents the longitudinal force of the ij-th wheel. F y_ij represents the lateral force of the ij-th wheel. The subscripts ij are respectively lf, rf, lr, rr: representing the left front wheel, right front wheel, left rear wheel, and right rear wheel (l represents left, the first r represents right, f represents front, and the second r represents rear). d represents the wheelbase of the coaxial wheels. l f and l r respectively represent the distances from the vehicle center of mass to the center of the front axle and the center of the rear axle.

[0028] The vehicle's yaw angular acceleration can be obtained as follows:

[0029]

[0030] When the vehicle is yawing during in-place steering, the longitudinal force and lateral force of the wheels are highly coupled. The maximum difference between the longitudinal torque M s and the lateral torque M f at different slip ratios can be calculated to determine the maximum total yaw torque and the corresponding angular acceleration. However, under different road surface conditions, the longitudinal force and the active yaw torque change. For example, Figure 2 shows the relationship between the vehicle's yaw angular acceleration and slip ratio under different adhesion coefficients: as the slip ratio increases, the longitudinal force and the active yaw torque increase, the resistance torque decreases, and the yaw angular acceleration increases accordingly and reaches an extreme point. Therefore, curves of M s and M f versus the slip ratio are plotted to find the appropriate slip ratio to generate the maximum total yaw torque ΔM z_max .

[0031] Assume the driver's accelerator pedal input, and a signal ζ representing the degree of pressing the accelerator pedal is generated. drive , and its maximum value, which is the value when the accelerator pedal is fully pressed, is ζ max . It is input that the desired vehicle reaches a steady-state yaw after 3τ s , where τ s is the basic time for control response. It is known that as the adhesion condition deteriorates, the stability optimization weight increases and the yaw response speed decreases. To make the yaw angular velocity reach a stable value after 3τ s , through calculation, the yaw angular velocity curve should be:

[0032]

[0033] where is the yaw angular acceleration required by the driver at time t. γ target is the target yaw angular velocity.

[0034] Step 2: Design of the in-place steering start strategy

[0035] To enable the vehicle to obtain estimated parameters such as speed and tire force adhesion coefficient within a short time in the early stage of in-place steering, and to avoid logical dead loops caused by no input to the observer and instability due to sudden changes in wheel torque during the starting stage, a start strategy is designed for before 2 s. This strategy provides sufficient sensor input for the observer to determine the desired yaw angular velocity, enabling the vehicle torque to smoothly transition to the yaw angular velocity tracking link. When the time t is less than 2 seconds, the vehicle starting torque decision is divided into two steps:

[0036] The first step is the initial starting torque decision. First, determine the initial adhesion coefficient as μ0, and use the vertical load F z and μ0 to calculate the initial starting torque T through the following formula r0 .

[0037] T r0 = μ0F z (4)

[0038] Then, based on the degree signal ξ of the accelerator pedal being depressed drive , its maximum value ξ max , time t, and a specific time parameter t s controls the speed at which the starting torque transitions from the initial value to the stable value. Calculate the target starting torque T using the following formula tar .

[0039]

[0040] The second step is to update the starting torque decision. Update the adhesion coefficient to μ obs , and combine it with the vertical load F z to calculate the updated starting torque T r . Then, according to the same formula above, recalculate the target starting torque T through the formula tar . This process adjusts the starting torque by updating the adhesion coefficient, enabling the vehicle to adapt to changes in road surface adhesion conditions and optimizing the starting performance. As Figure 3 visually shows the above description.

[0041] Step 3: Design of the total yaw moment decision method based on active disturbance rejection control

[0042] After obtaining the target starting torque, in order to achieve precise tracking of the vehicle's yaw angular velocity, it is necessary to further make a decision on the total yaw moment. After determining the vehicle's target yaw angular velocity in Step 1, this step is based on the active disturbance rejection control (LADRC) theory, and uses the yaw motion characteristics of the vehicle to construct a control model to obtain the precise active yaw moment. From the vehicle's yaw motion formula, it is written in the general form of LADRC:

[0043]

[0044] where is the yaw angular acceleration, ΔM z is the total yaw moment, where f(t) is the total disturbance of the vehicle's yaw motion system, f(t) = M f + w(t), M f is the resistance moment, and w(t) is the internal and external disturbance of the system.

[0045] Convert the above equation into the state - space form, and define the state variables where f is the total disturbance state variable of the wheel yaw motion:

[0046]

[0047] where A is the system matrix, B is the control input matrix, C is the output matrix, D is the direct transmission matrix, E is the disturbance input matrix, h is the dynamic change of the total disturbance, and u is the total yaw moment. Their respective values are where b0 represents the control gain 1 / I z . C = [1 0], D = 0, u = ΔM z ,

[0048] where LADRC tracks y and f by configuring the parameters of the linear extended state observer (LESO). According to the state equation, the state variables are defined as z1 is the estimated value of the true state y (yaw rate) of the system. z2 is the estimated value of the total disturbance f(t). Through the gain matrix L = [β1 β2] T the dynamics of the original system are adjusted so that the observer can quickly track the true state and disturbance. The matrices A LESO and B LESO of the LESO do not directly follow the A and B in the original system formula (7), but are generated through the following steps: A LESO = A - LC, B LESO = B, where C = [1 0] is the output matrix of formula (7). (x1 - z1) is the output error of the observer, and by adjusting L, the corrective action of the feedback observer is the estimated output of the observer, which is the output value estimated in real time by the linear extended state observer (LESO).

[0049] The state - space equation of the LESO is established as:

[0050]

[0051] where u = ΔM z is the total yaw moment. Substituting L = [β1 β2] T , the formula is as follows:

[0052]

[0053] where, are the derivative values of z1 and z2. z1 is the estimated value of the true state y (yaw rate) of the system. z2 is the estimated value of the total disturbance f(t), which are the estimated values of the yaw angular acceleration and the dynamic change of the disturbance respectively. The extended observer parameters β1 and β2 are two parameters of the LESO and can control the tracking speed.

[0054] In LADRC, linear state error feedback (LSEF) enhances the system's dynamic performance and steady-state accuracy by providing feedback on the system state, thus avoiding significant overshoot caused by disturbances or system nonlinearity. The general form of LSEF can be expressed as:

[0055] u0(t)=K(x r -z1) (10)

[0056] Where K is the gain parameter. u0(t) represents the control input, x r Represents the reference state quantity.

[0057] Substituting into the state space equation of LESO, we get

[0058]

[0059] in, is the estimated value of the yaw angular acceleration, is the estimated value of the dynamic change of the disturbance.

[0060] Based on factors such as vehicle dynamics, control objectives, and actual driving conditions, simulation experiments were conducted with yaw rate tracking accuracy and interference rejection as the targets. Yaw rate tracking accuracy focuses on the degree of agreement between the actual yaw rate and the desired yaw rate, while interference rejection focuses on the suppression of the total disturbance in the equation. The LADRC controller parameters were determined as follows: a gain parameter K of 10 ensures a fast system response and reasonable tracking error control; the first parameter β1 of the extended state observer is 20, and the second parameter β2 of the extended state observer is 100. This allows the observer to quickly and accurately estimate disturbances, providing compensation information to the controller, improving yaw rate tracking accuracy and the system's interference rejection. This parameter selection balances response speed and stability under different vehicle driving conditions.

[0061] Input target yaw rate γ target Then, the LADRC control model is established according to the above equation 11, and the tracking γ target The precise active yaw moment required to ensure M s Meet the requirement of accurately tracking the target yaw angular velocity. Then the next step is to s Further optimization of the distribution can be done to determine the torque per wheel.

[0062] Step 4: Establishment of the control system for optimizing the lower torque distribution of stationary steering

[0063] The active yaw moment M obtained based on the upper total yaw moment decision s Torque distribution for the main goal.

[0064] Based on the vehicle model established above, the torque average distribution strategy is now adopted to obtain the absolute value of the expected wheel-end torque of a single wheel:

[0065]

[0066] M s is the active yaw moment obtained from the upper-layer decision-making, and d is the wheelbase between the front and rear axles. Taking u = [T lf , T lr , T rf , T rr as the system input quantity, the control quantity target u T is: r

[0067] u r = [(T lf ) r (T lr ) r (T rf ) r (T rr ) r T = [-T r -T r T r T r T (13)

[0068] Among them, (T lf ) r (T lr ) r (T rf ) r (T rr ) r represent the reference torques of the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively.

[0069] To ensure that the position of the center of mass does not shift, it is necessary to try to keep the coordinate position of the vehicle's center of mass unchanged. Therefore, it is necessary to control the longitudinal speed of the whole vehicle. Taking the longitudinal speed as the state variable and setting the target value to 0, the optimal wheel-end torque for tracking the active yaw moment M s is determined. Using the model predictive control (MPC) method, based on the control quantity target u r the optimal control quantity u, that is, the torque T, is obtained. The torque output of the in-wheel motor is controlled according to the generated control quantity motor torque command value to achieve the purpose of driving anti-skid;

[0070] The following is a drive anti-skid control simulation example of an in-wheel motor-driven electric vehicle using the method of the present invention:

[0071] ​​​This example takes an electric vehicle driven by four-wheel hub motors as the control object. The following are some parameters of the experimental vehicle itself: the total vehicle mass m = 1412 kg, the effective rolling radius of the wheel R w = 0.325 m, the moment of inertia of the wheel I w = 0.9 kg·m2, the yaw moment of inertia I Z = 1536.7 kg·m2.

[0072] To verify the robustness, effectiveness of the control strategy proposed in this paper under different road surface conditions, as well as the response speed and accuracy of tracking the expected yaw angular velocity, a simulation experiment was designed. The control algorithm was verified by selecting high-adhesion and low-adhesion road surfaces in sequence:

[0073] Under high-adhesion road surface: Figure 5 (a) In it, except for a short lag after starting, the yaw angular velocity has no overshoot throughout the process and the tracking effect is good; Figure 5 (b) The trajectory curve of the center of mass also shows a stable change matching the yaw angular velocity accordingly. In the low-adhesion road surface: Figure 6 (a) In it, the peak value of the yaw moment breaks through the resistance moment, and the yaw angular velocity quickly responds to the driving request without overshoot; Figure 6 (b) The trajectory curve of the center of mass also reflects the effective control of the yaw motion. The results of the two types of road surfaces show that this control strategy has the ability to accurately track the yaw angular velocity under different adhesion conditions, effectively avoiding overshoot and lag problems, and achieving a stable response to the driving request.

[0074] After MPC optimization allocation, the longitudinal and lateral vehicle speeds of the whole vehicle on high-adhesion and low-adhesion road surfaces are both suppressed to fluctuate within a narrow range, and the longitudinal and lateral offsets of the center of mass are controlled within a very small range. Among them, on the high-adhesion road surface, by reasonably allocating the torque, the vehicle speed and the trajectory of the center of mass during yaw motion are ensured to be stable; on the low-adhesion road surface, by restricting the wheel torque, excessive slip is avoided, and the offset of the center of mass is further restricted. The simulation results under the two types of working conditions jointly verify the accurate constraint ability of this strategy for the trajectory of the vehicle center of mass and the stability guarantee effect under complex road surface conditions.

Claims

1. A precise in-situ steering control method based on active disturbance rejection, characterized in that, Including: Step 1, determining the yaw rate under different road surface conditions: Based on the relationship between the vehicle's yaw angular acceleration and slip ratio under different adhesion coefficients obtained under different road surface conditions, the extreme point angular acceleration is obtained; considering the actual driver's pedal input and combining with the extreme point angular acceleration to obtain the target yaw rate; Step 2, designing an in-situ steering start strategy: including setting an initial starting torque, and adjusting the starting torque through the update of the adhesion coefficient to obtain the target starting torque; Step 3. Design a total yaw moment decision-making method based on active disturbance rejection control: Based on the active disturbance rejection control theory, utilize the yaw motion characteristics of the vehicle to construct an LADRC control model to obtain the active yaw moment M s ; Step 4. Establish an in-situ steering lower-layer torque optimization distribution control system: Taking the longitudinal speed as the state variable and setting the target value to 0, determine the optimal wheel-end torque for tracking the active yaw moment M s ; Using the model predictive control method, based on the control quantity u r target, obtain the optimal control quantity u, and control the torque output of the in-wheel motor according to the generated control quantity motor torque command value to achieve the purpose of driving anti-skid.

2. The in-situ steering precise control method based on active disturbance rejection according to claim 1, wherein, The control quantity u = [T lf , T lr , T rf , T rr T As the input quantity, the control quantity target u r = [(T lf ) r (T lr ) r (T rf ) r (T rr ) r T = [-T r -T r T r T r T , where (T lf ) r (T lr ) r (T rf ) r (T rr ) r represent the reference torques of the left front wheel, right front wheel, left rear wheel, and right rear wheel respectively; T r is the absolute value of the expected wheel-end torque of a single wheel:​​​ d is the wheelbase between the front and rear axles.

3. The in-situ steering precise control method based on active disturbance rejection according to claim 1, characterized in that Constructing the LADRC control model as: Define state variables The state - space equation for establishing a linear extended state observer is as follows: Among them, is the derivative value of z1 and z2, A LESO and B LESO are the matrices of the linear extended state observer; the gain matrix L = [β1 β2] T , A LESO = A - LC, B LESO = B; the system matrix control input matrix output matrix C = [1 0], b0 represents the control gain; z1 is the estimated value of the yaw rate y, z2 is the estimated value of the total disturbance f(t), which is the estimated value of the yaw angular acceleration, is the estimated value of the dynamic change of the disturbance, β1 and β2 are the two parameters of the extended state observer, u = ΔM z is the total yaw moment, and K is the gain parameter.

4. The in-situ steering precise control method based on active disturbance rejection according to claim 1, characterized in that The initial starting torque is: where μ0 is the initial adhesion coefficient and F z is the vertical load; The target starting torque is: ξ drive 、ξ max are the degree signal of depressing the accelerator pedal and its maximum value respectively, t s is the set time parameter, t is time, T r is the starting torque calculated after updating the adhesion coefficient.

5. The in-situ steering precise control method based on active disturbance rejection according to claim 1, wherein The target yaw rate is: where ξ drive and ξ max are the degree signal of depressing the accelerator pedal and its maximum value respectively, τ s is the control response base time, is the extreme point angular acceleration, t is time, is the yaw angular acceleration required by the driver at time t.

Images