Vehicle curve adaptive cruise stability control method using differential steering

Abstract

The invention relates to a vehicle curve self-adaptive cruise stability control method utilizing differential steering, which adopts a hierarchical control architecture, and combines a differential steering control strategy based on MFAC and a fuzzy logic rule of multi-target weight coefficient self-adjustment based on MPC in upper layer control, so that the stability of the vehicle curve self-adaptive cruise stability is improved. Taking the actual vehicle speed, acceleration, position, yaw velocity and side slip angle of the vehicle as input, and outputting to obtain the expected acceleration of the vehicle; in the lower-layer control, driving and braking switching logic based on a deceleration protection range threshold value and a feedforward-feedback fusion control strategy based on an acceleration error are combined, the actual vehicle speed, the acceleration and the expected vehicle acceleration of the vehicle serve as input, and a vehicle hub motor torque signal is output and obtained; and transmitting the vehicle hub motor torque signal to the vehicle hub motor so as to correspondingly control and change the driving state of the vehicle. Compared with the prior art, the transverse stability of the vehicle in the curve vehicle following process can be improved, and the transverse displacement deviation can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle control, and in particular to a vehicle curve adaptive cruise stability control method utilizing differential steering. Background Art

[0002] With the rapid increase in the number of cars, road safety issues have become increasingly serious. As an advanced driver assistance system, the advanced ACC (Adaptive Cruise Control) system plays an important role in improving driving safety. It can adjust the speed of the vehicle in a timely and accurate manner according to the driving status of the two vehicles in front and behind and the distance between the vehicles, so as to always maintain a safe following distance. However, the current research on the ACC system still has many limitations.

[0003] First, in terms of adaptability, the system has poor adaptability to environmental changes. In the widely used control strategies, the dynamic adjustment capabilities of various performance indicators are obviously insufficient. For example, in the complex environment where the front vehicle suddenly changes speed frequently, the ACC system may not be able to make appropriate adjustments quickly, thus affecting driving safety. Secondly, the existing research lacks consideration and analysis of driving economy, which limits its application value to a certain extent. In addition, from the perspective of research conditions, the current research on ACC systems is mainly focused on several typical driving conditions such as straight roads, and there are few studies on multi-objective cruise control methods under curved conditions. In actual driving, when the front vehicle frequently changes speed in a curve, the lateral stability of the self-vehicle in the following process will be greatly affected, and the driving path is also prone to deviation. Therefore, in the design of the curve adaptive cruise control method, the coordination of multiple objectives must be fully considered, both to ensure the safe following distance of the vehicle in the curve and to take into account the lateral stability of the vehicle. The existing ACC system cannot be well applied to the complex conditions of variable curvature curves and frequent acceleration and deceleration of the front vehicle. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a vehicle curve adaptive cruise stability control method using differential steering, which can improve the lateral stability of the vehicle during curve following and reduce lateral displacement deviation.

[0005] The object of the present invention can be achieved by the following technical scheme: a vehicle curve adaptive cruise stability control method using differential steering, adopting a hierarchical control architecture including upper-level control and lower-level control, in which a differential steering control strategy based on MFAC (Model-Free Adaptive Control) and a fuzzy logic rule of multi-objective weight coefficient self-adjustment based on MPC (Model Predictive Control) are combined in the upper-level control, and the actual vehicle speed, acceleration, position, yaw angular velocity and sideslip angle of the center of mass of the vehicle are used as inputs, and the desired acceleration of the vehicle is obtained as output;

[0006] In the lower-level control, the drive-brake switching logic based on the deceleration protection range threshold and the feedforward-feedback fusion control strategy based on the acceleration error are combined. The actual vehicle speed, acceleration and expected vehicle acceleration are used as inputs, and the vehicle wheel hub motor torque signal is output.

[0007] The vehicle hub motor torque signal is transmitted to the vehicle hub motor to control and change the vehicle driving state accordingly.

[0008] Furthermore, the MFAC-based differential steering control strategy adopts a two-input-two-output nonlinear discrete-time system, wherein the inputs are the actual vehicle speed, yaw rate and sideslip angle of the center of mass of the vehicle, and the outputs are the driving torques of the left and right front wheels.

[0009] Furthermore, the MFAC-based differential steering control strategy is specifically:

[0010] p f (k+1)=f f (m f (k), m f (k-1),…,m f (kn i ), p f (k), p f (k-1),…,p f (k-n0))

[0011] Among them, m f (k) = [T fl , T fr ] T ∈R 2×1 is the control input vector of the vehicle, p f (k) = [e β (k) e ω (k ) ] T ∈R 2×1 is the vehicle stability output vector, f f(k)∈R 2×1 is the unknown nonlinear function vector, n i , n0∈Z are two integers, Z is a set of integers, representing the lengths of the input and output signal sequences of the system respectively.

[0012] Furthermore, the fuzzy logic rules with self-adjustment of multi-objective weight coefficients based on MPC are specifically based on the longitudinal kinematic model of the workshop, and the objective function, performance indicators and constraints are set under the MPC framework. The expected acceleration in the control time domain is obtained through optimization calculation, and the fuzzy logic rules are designed to update the weight coefficients of the performance indicators in real time to meet the complex and changeable driving conditions.

[0013] Furthermore, the performance indicators include safety, vehicle following ability, economy and comfort.

[0014] Furthermore, the objective function is specifically:

[0015]

[0016] in, U is a set of control variables, including each control variable and its weight coefficient. The control variables include relative vehicle speed, vehicle spacing error, vehicle acceleration, and vehicle acceleration change rate. f is the weight coefficient of vehicle following and safety, L c is the weight coefficient of vehicle comfort and economy, q Δd is the weight factor of the inter-vehicle distance error, They represent the relative speed of the two vehicles, the acceleration of the vehicle and the weight factor of the acceleration change rate, ε is the vector relaxation factor, A H is the coefficient matrix of the constraint conditions, b H is the constant term matrix of the constraints.

[0017] Furthermore, the fuzzy logic rules are specifically:

[0018] When the inter-vehicle distance error Δd < 0, the vehicle is likely to collide with the vehicle in front. At this time, the weight coefficient of the inter-vehicle distance error q Δd It should be increased rapidly, first of all to ensure the safety of vehicles;

[0019] As the inter-vehicle distance error Δd increases, the collision risk of the ego vehicle will gradually be eliminated. At this time, the weight coefficient q of the inter-vehicle distance error is reduced. Δd When the distance between vehicles increases to more than the preset threshold, the distance between vehicles is controlled to prevent the distance between vehicles from continuing to increase and vehicles in adjacent lanes from cutting in at will;

[0020] When the relative speed v relWhen <0, it means that the vehicle speed is greater than the front vehicle speed. If the inter-vehicle distance error Δd<0 at this time, the vehicle will be controlled to decelerate rapidly.

[0021] If the relative speed v rel <0 and the inter-vehicle distance error Δd>0, the vehicle is controlled to decelerate slowly to improve driving economy;

[0022] When the vehicle is in a dangerous state, the control quickly reduces the weight coefficient of the vehicle's acceleration. Ensure that the vehicle can respond quickly. Conversely, when the vehicle is in a safe state, increase the weight coefficient of the vehicle's acceleration. Control the vehicle to accelerate or decelerate slowly to improve driving economy.

[0023] Furthermore, in the driving and braking switching logic based on the deceleration protection range threshold, the deceleration protection range threshold is 0.1m / s 2 , when the desired deceleration of the vehicle is a des When ≥a+0.1, it is drive control;

[0024] When the desired deceleration of the vehicle is a des When ≤a-0.1, it is brake control;

[0025] When a-0.1 des When <a+0.1, the vehicle maintains the current state and does not switch between driving and braking;

[0026] Where a is the base deceleration.

[0027] Furthermore, the specific control process of the feedforward-feedback fusion control strategy based on acceleration error is:

[0028] The expected driving torque T obtained by the inverse dynamics model of the driving system motor_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity during modeling on the control system, the expected acceleration a is used. des and the actual acceleration a act The error between the two is fed back and regulated, and the driving torque T motor_ref As the feedback quantity, it is the final desired driving torque;

[0029] The desired brake pressure P obtained by the inverse dynamics model of the brake system m_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity during modeling on the control system, the expected deceleration a is used. des and the actual deceleration a act The error between the two is fed back and regulated, and the obtained brake pressure P m_ref The feedback amount is the final desired braking pressure. ​

[0030] Furthermore, the driving torque T motor_ref Specifically:

[0031]

[0032]

[0033] Among them, T motor_ref is the expected driving torque after adjustment, K tp and K ti are the proportional and integral parameters of the drive feedback PI controller respectively;

[0034] The brake pressure P m_ref Specifically:

[0035] P m_ref =P m_des +K pp [a des (n)-a act (n)]+K pi ∑[a des (n)-a act (n)]

[0036]

[0037] Among them, P m_ref is the expected brake pressure after adjustment, K pp and K pi are the proportional parameter and integral parameter of the brake feedback PI controller respectively.

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] The present invention is based on a hierarchical control architecture. In the upper control layer, a differential steering control method based on MFAC is proposed, and a fuzzy logic rule based on MPC multi-objective weight coefficient self-adjustment is designed to improve the lateral stability and path tracking ability of the vehicle when following a car on a curve, and optimize the expected acceleration of the vehicle in real time; in the lower control layer, a control method of driving and braking switching logic, feedforward and feedback fusion based on the deceleration protection range threshold is proposed to control the vehicle to stably complete the following car on the curve. In this way, the lateral stability of the vehicle in the process of following a car on a curve is improved and the lateral displacement deviation is reduced through differential steering.

[0040] The present invention proposes a differential steering control strategy based on MFAC, and according to the dynamic model of the differential steering system, determines the input and output variables of the designed control system, wherein the input variables are the errors between the yaw rate and the sideslip angle of the center of mass of the vehicle and their respective ideal values, and the output variables are the driving torques of the left and right front wheels (there is a differential driving torque). It is shown through verification that the tracking error steady-state value output by the differential steering control strategy based on MFAC eventually converges to 0, that is, it shows that the differential steering control strategy based on MFAC is stable, and can effectively improve the lateral stability and path tracking ability of the vehicle when following a vehicle in a curve, and meet the driver's use requirements for the ACC system in a variable curvature curve.

[0041] The present invention proposes a fuzzy logic rule with self-adjustment of multi-objective weight coefficients based on MPC, selects appropriate performance indicators (including safety, following performance, economy and comfort) and constraints under the MPC framework, and introduces relaxation vectors on the basis of system hard constraints, so that the range of variation of performance indicators is further expanded to ensure that the system will not have an unsolvable situation, obtains the expected acceleration in the control time domain through optimization calculation, and designs fuzzy logic rules to update the weight coefficients of performance indicators in real time, so as to better meet complex and changeable driving conditions, so that the vehicle has different emphases on following performance, economy, safety and comfort under different driving environments, thereby improving the environmental adaptability of the ACC system.

[0042] The invention is designed to be 0.1m / s 2 The protection range threshold is used to realize the switching control judgment of driving / braking. Taking into account the nonlinearity of the vehicle that will cause errors in model establishment, a feedforward and feedback fusion control method is designed to obtain accurate expected driving torque and expected braking pressure, which can reduce the discomfort caused by frequent switching of driving and braking, reduce the impact of acceleration error on the system, and improve the robustness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0044] Figure 2 is a schematic diagram of an application framework of an embodiment;

[0045] Figure 3 Schematic diagram of the longitudinal kinematics model of a workshop in the embodiment;

[0046] Figure 4a to Figure 4c A surface diagram showing the relationship between the weight coefficient and the performance index in the fuzzy logic rule in the embodiment;

[0047] Figure 5 It is a schematic diagram of the structure of the driving and braking switching strategy in the embodiment;

[0048] Figure 6a to Figure 6cand Figure 7a to Figure 7c It is the simulation result under FHWA variable curvature curve condition in the embodiment. DETAILED DESCRIPTION

[0049] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0050] Example

[0051] This scheme takes into account that for electric vehicles driven by in-wheel motors, the steering function of the vehicle can be achieved by precisely controlling the differential torque of the wheels on both sides, which is the so-called differential steering. In view of the fact that current passenger cars are generally equipped with traditional steering systems, many related studies are currently using differential steering as a fault-tolerant control method to ensure the trajectory tracking and yaw stability of the vehicle after the original steering system fails. Considering that differential steering can flexibly adjust the body posture and front wheel steering angle, it can be combined with other safety assistance driving systems to ensure the safety and stability of the vehicle during driving. Therefore, a vehicle curve adaptive cruise stability control method using differential steering is designed, such as Figure 1 As shown in the figure, a hierarchical control architecture including upper control and lower control is adopted. In the upper control, the differential steering control strategy based on MFAC and the fuzzy logic rule of multi-objective weight coefficient self-adjustment based on MPC are combined. The actual vehicle speed, acceleration, position, yaw rate and sideslip angle of the center of mass of the vehicle are used as inputs, and the desired acceleration of the vehicle is obtained as output.

[0052] In the lower-level control, the drive-brake switching logic based on the deceleration protection range threshold and the feedforward-feedback fusion control strategy based on the acceleration error are combined. The actual vehicle speed, acceleration and expected vehicle acceleration are used as inputs, and the vehicle wheel hub motor torque signal is output.

[0053] The vehicle hub motor torque signal is transmitted to the vehicle hub motor to control and change the vehicle driving state accordingly.

[0054] This embodiment applies the above solution to build Figure 2 The application framework shown has the following main contents:

[0055] Step 1: In the upper-level control of the curve adaptive cruise control, the center of mass side slip angle error e β and the yaw rate error e ω As a feedback control variable, a differential steering control strategy based on MFAC is proposed by adjusting the differential driving torque of the left and right front wheels to improve the lateral stability and path tracking ability of the vehicle when following a car on a curve.

[0056] Step 2: Based on the longitudinal kinematic model of the workshop, a fuzzy logic rule with multi-objective weight coefficient self-adjustment based on MPC is proposed to optimize the expected acceleration of the ego vehicle in real time;

[0057] Step 3: In the lower-level control of the curve adaptive cruise control, a control method based on the driving and braking switching logic, feedforward and feedback fusion based on the deceleration protection range threshold is proposed to control the vehicle to stably complete the curve cruise following driving.

[0058] In step 1, in order to improve the lateral stability and path tracking ability of the vehicle when following a car on a curve, a differential steering control strategy based on MFAC is proposed. The control law is extreme valued by introducing a criterion function, and the estimation algorithm of the pseudo-Jacobian matrix is ​​designed by using the minimum parameter estimation method. At the same time, a reset mechanism is added to ensure the rationality of the variation range of the pseudo-Jacobian matrix.

[0059] According to the dynamic model of the differential steering system, the input and output variables of the designed control system are determined. The input variables are the errors between the vehicle's yaw rate and the sideslip angle of the center of mass and their respective ideal values, and the output variables are the driving torques of the left and right front wheels (there is a differential driving torque). At this time, the differential steering control system based on MFAC can be designed as a two-input-two-output nonlinear discrete time system:

[0060]

[0061] In the formula, m f (k) = [T fl , T fr ] T ∈R 2×1 is the control input vector of the vehicle, p f (k) = [e β (k) e ω (k)] T ∈R 2×1 is the vehicle stability output vector, f f (k)∈R 2×1 is the unknown nonlinear function vector, n i , n0∈Z (Z is an integer set) are two integers, representing the length of the input and output signal sequences of the system respectively.

[0062] Make the following two assumptions:

[0063] Assumption 1: Function f f (·) Regarding the input variable m f The partial derivative of (k) is continuous;

[0064] Assumption 2: The designed control system satisfies the Lipschitz condition, that is, for any k and ||Δmf (k)||≠0, and there exists||Δp f (k+1)||≤ζ||Δm f (k)||. In the formula, Δp f (k+1)=p f (k+1)-p f (k), Δm f (k) = m f (k)-m f (k-1), ζ is a constant greater than 0, and ||·|| is the 2-norm.

[0065] For a nonlinear discrete-time system satisfying Assumptions 1 and 2, any k and ||Δm f (k)||≠0, time-varying parameters of the pseudo-Jacobian matrix will exist, which can transform the nonlinear discrete time system into a CFDL mathematical model. The time-varying parameters of the pseudo-Jacobian matrix It reflects the relationship between two inputs and two outputs in the control system. The value in the matrix represents the closeness of the relationship between the input and output. The weight assumption is introduced into the element values ​​of .

[0066] In order to express the stability of the differential steering controller effect during the vehicle steering process, the criterion function is introduced:

[0067] J m (k)=||p d (k+1)-p f (k+1)|| 2 +λ||m f (k)-m f (k-1)|| 2

[0068] In the formula, the weight factor λ>0, indicating the degree of restriction on the range of input variation. Considering the minimization problem of the criterion function, for m f (k) By performing extreme value calculation, we can obtain:

[0069]

[0070] Where ρ∈(0, 1] is the step size factor.

[0071] The estimation algorithm of the pseudo-Jacobian matrix is ​​designed by the minimum parameter estimation method. First, the estimation criterion function is defined:

[0072]

[0073] In the formula, ζ>0 is the weight factor. By performing extreme value operation on formula (4-23), we can obtain The estimation algorithm is:

[0074]

[0075] In the formula, η∈(0,2] is the step size factor. According to the estimation, the expression of MFAC designed in this embodiment is:

[0076]

[0077] In order to ensure the rationality of the variation range of the pseudo-Jacobian matrix, a reset mechanism is added, that is, when or When Where ε is a very small positive constant.

[0078] The error variable of the control system is defined as:

[0079] e f = pd(k)-p f (k)

[0080] After derivation and proof, we have:

[0081]

[0082] It is proved that the differential steering control system designed based on MFAC in this scheme is stable, and the steady-state value of the tracking error output by the control system eventually converges to 0.

[0083] In step 2, according to Figure 3 The longitudinal kinematic relationship of the workshop shown in the figure gives the discrete time domain relationship of the ACC system:

[0084]

[0085] In the formula, all variables need to satisfy the following relationship:

[0086]

[0087] Where d(k) is the actual distance between the vehicle and the front vehicle, v rel (k) is the relative speed of the two vehicles in front, v f (k) is the vehicle speed, a f (k) is the acceleration of the vehicle, jerk(k) is the rate of change of acceleration of the vehicle, v p (k) is the speed of the preceding vehicle, is the system disturbance, i.e. the acceleration of the preceding vehicle, u a (k) is the expected acceleration of the vehicle, T s is the system sampling period.

[0088] Define x(k)=[d(k),v f (k), v rel (k), a f (k), jerk (k)] d , with x(k) as the state variable of the system, a f (k) is the system disturbance, and further considering the acceleration change rate of the vehicle, the discrete state equation model is obtained:

[0089]

[0090] In the formula,

[0091] This scheme selects appropriate performance indicators (taking safety as the most important performance indicator, taking into account following performance, economy and ride comfort) and constraints under the MPC framework, obtains the expected acceleration in the control time domain through optimization calculation, and designs fuzzy logic rules to update the weight coefficients of performance indicators in real time to meet complex and changeable driving conditions.

[0092] 2.1 Car-following prediction model based on feedback correction

[0093] The error between the actual driving state and the predicted driving state of the ACC system at time k is defined as:

[0094] e x (k) = x(k) - x p (k|k-1)

[0095] Where x(k) is the actual driving state of the ACC system at time k, x p (k|k-1) is the predicted driving state at time k-1. The predicted driving state error can be weighted to correct the model prediction value of the ACC system for the future time, and the result is:

[0096]

[0097] In the formula, x(k-1) is the state variable of the control system at time k-1, u a (k-1) is the input of the control system at time k-1, is the disturbance of the control system at time k-1.

[0098] Through the prediction error e x (k) The future state variables x of the longitudinal kinematic model of the ACC system workshop p (k+1|k) is corrected to improve the accuracy of the prediction model. Therefore, the system prediction state variable at time k is:

[0099]

[0100] Wherein, W=diag(w1, w2, w3, w4, w5) is the system correction matrix, and the value range of each element in W is (0, 1).

[0101] The MPC cost function designed in this embodiment is:

[0102]

[0103] In the formula, Q and R u is the weight coefficient.

[0104] q Δd is the weight factor of the vehicle distance error, q Δd The larger the value, the smaller the error between the actual vehicle distance and the expected value. The larger the value is, the smaller the relative speed of the two vehicles, the acceleration of the vehicle and the rate of change of acceleration are. In order to ensure that the ACC system can adapt to different driving conditions, the objective function selects different weights, and

[0105]

[0106] Where, L f is the weight coefficient of vehicle following and safety, L c is the weight coefficient of vehicle comfort and economy, and there is a contradictory relationship between the two.

[0107] 2.2 Introduction of relaxation factor and multi-objective optimization solution

[0108] This solution introduces a relaxation vector based on the hard constraints of the system, which further expands the range of variation of the performance indicators and ensures that the system will not have no solution.

[0109] According to the designed objective function, performance indicators and related constraints, MPC is used to transform the multi-objective optimization control of the ACC system into a quadratic programming problem, namely:

[0110]

[0111] In the formula, U is the set of control variables, ε is the vector relaxation factor, A H is the coefficient matrix of the constraint conditions, b H is the constant term matrix of the constraints.

[0112] Find the optimal solution for performance indicators in each control cycle That is, a series of control quantities including relaxation factors are:

[0113]

[0114] The first component u of the control system input a * (k) is used as the actual input of the system at the next moment and is used for rolling optimization and prediction. The MPC principle is used to solve the expected acceleration of the vehicle.

[0115] This scheme defines the input quantities Δd, v rel and output q Δd , The fuzzy logic linguistic variables are:

[0116]

[0117] In the formula, NB, NM, NS, ZO, PS, PM, and PB represent negative large, negative medium, negative small, zero, positive small, positive medium, and positive large respectively; VS, S, M, B, and VB represent negative small, small, medium, large, and negative large respectively.

[0118] The seven fuzzy logic rules designed are as follows: when Δd<0, the vehicle is likely to collide with the vehicle in front. At this time, the weight coefficient of Δd is q Δd It should be increased rapidly to ensure the safety of the vehicle first. As Δd increases, the collision risk of the vehicle will gradually be eliminated. At this time, q can be appropriately reduced. Δd However, when the distance between vehicles is too large, it is also necessary to control the distance between vehicles to prevent the distance between vehicles from continuing to increase and causing vehicles in adjacent lanes to cut in at will. rel When Δd<0, it means that the vehicle speed is greater than the front vehicle speed. If Δd<0 at this time, the vehicle needs to be controlled to decelerate sharply. rel If Δd>0, the vehicle can decelerate slowly to improve driving economy. If the vehicle is in a dangerous state, it should be quickly reduced. Ensure that the vehicle can respond quickly; on the contrary, when the vehicle is in a safe state, it should increase Control the vehicle to accelerate or decelerate slowly to improve driving economy.

[0119] When designing fuzzy logic rules, it is necessary to defuzzify the output result set to obtain a relatively accurate structure set. This embodiment uses the area centroid method to perform defuzzification calculation, thereby obtaining the weight coefficient and the vehicle distance error Δd, relative vehicle speed v rel and the vehicle acceleration a f The relationship surface diagram between Figures 4a to 4c shown.

[0120] In step 3, according to Figure 5 The driving and braking control switching acceleration curve shown in the figure is used to judge the acceleration and deceleration of the vehicle. This solution adds 0.1m / s around the relationship curve between vehicle speed and deceleration 2The protection range threshold, when the vehicle's expected deceleration is a des ≥a+0.1, it is drive control; when the desired deceleration of the vehicle is a des ≤a-0.1, it is brake control; when a-0.1 des When <a+0.1, the vehicle maintains the current state and does not switch between driving and braking, so as to reduce the discomfort caused by frequent switching between driving and braking, where a is the reference deceleration.

[0121] The relative speed of the two vehicles and the distance error between the two vehicles are combined to generate the expected driving torque / brake pressure required for the acceleration / deceleration of the vehicle. Considering that the nonlinearity of the vehicle will cause errors when the model is established, a feedforward and feedback fusion control method is designed. The expected driving torque T obtained by the inverse dynamics model of the drive system is converted into motor_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity on the control system during modeling, the expected acceleration a is used. des and the actual acceleration a act The driving torque T is obtained by feedback adjustment of the error between motor_ref As feedback quantity. The final desired driving torque T motor_ref for:

[0122] T motor_ref =T motor_des +K tp [a des (n)-a act (n)]+K ti ∑[a des (n)-a act (n)]

[0123]

[0124] Where, T motor_ref is the expected driving torque after adjustment, K tp and K ti are the proportional parameter and integral parameter of the drive feedback PI controller respectively.

[0125] The desired brake pressure P obtained by the inverse dynamics model of the brake system m_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity during modeling on the control system, the expected deceleration a is used. des and the actual deceleration a act The brake pressure P is obtained by feedback adjustment of the error between m_ref As feedback quantity. The final desired braking pressure P m_ref for:

[0126] P m_ref =P m_aes +K​pp [a des (n)-a act (n)]+K pi ∑[a des (n)-a act (n)]

[0127]

[0128] Where P m_ref is the expected brake pressure after adjustment, K pp and K pi are the proportional parameter and integral parameter of the brake feedback PI controller respectively.

[0129] In order to prove the effectiveness of the control method for vehicle curve adaptive cruise stability using differential steering proposed in this scheme, this embodiment performs simulation verification on a FHWA variable curvature curve when the leading vehicle frequently changes speed.

[0130] Figure 6a-6c It mainly reflects the vehicle's curve following performance under the control method of this scheme. The vehicle distance error under the fuzzy MPC upper control is 0.25m, which is 20.24% less than that of the traditional MPC, and the driving safety is improved. At the same time, the acceleration of the vehicle can track the expected value well, and its variation range is [-1.71,1.65]m / s 2 , ensuring riding comfort.

[0131] Figures 7a to 7c The main effect of the lateral stability control based on differential steering designed in this scheme is reflected. The yaw rate and sideslip angle of the vehicle without stability control produce obvious oscillations, while the vehicle with differential steering control can track the expected values ​​well, and the driving stability of the vehicle is improved.

[0132] In summary, this solution comprehensively considers the multi-objective optimization control of the vehicle in the process of following the vehicle and the curved driving conditions, so that the vehicle has different emphasis on following performance, economy, safety and comfort in different driving environments. It can dynamically coordinate the vehicle's safety, following performance, comfort and economy under complex conditions such as variable curvature curves and frequent acceleration and deceleration of the leading vehicle, thereby improving the environmental adaptability of the ACC system; and uses differential steering to improve the vehicle's lateral stability and path tracking ability in the process of following the vehicle on curves, meeting the driver's demand for the use of the ACC system in variable curvature curves and expanding the use scenarios of the ACC system.

Claims

1. A vehicle curve adaptive cruise stability control method using differential steering, characterized in that: A hierarchical control architecture including upper-level control and lower-level control is adopted. In the upper-level control, the differential steering control strategy based on MFAC and the fuzzy logic rule with multi-objective weight coefficient self-adjustment based on MPC are combined. The actual vehicle speed, acceleration, position, yaw rate and sideslip angle of the center of mass of the vehicle are used as inputs, and the desired acceleration of the vehicle is output. In the lower-level control, the drive-brake switching logic based on the deceleration protection range threshold and the feedforward-feedback fusion control strategy based on the acceleration error are combined. The actual vehicle speed, acceleration and expected vehicle acceleration are used as inputs, and the vehicle wheel hub motor torque signal is output. The vehicle hub motor torque signal is transmitted to the vehicle hub motor to control and change the vehicle driving state accordingly.

2. The vehicle curve adaptive cruise stability control method using differential steering according to claim 1, characterized in that: The MFAC-based differential steering control strategy adopts a two-input-two-output nonlinear discrete time system, wherein the inputs are the actual vehicle speed, yaw rate and sideslip angle of the center of mass of the vehicle, and the outputs are the driving torques of the left and right front wheels.

3. The vehicle curve adaptive cruise stability control method using differential steering according to claim 2, characterized in that: The MFAC-based differential steering control strategy is specifically: p f (k+1)=f f (m f (k),m f (k-1),…,m f (k-n i ),p f (k),p f (k-1),…,p f (k-n0)) Among them, m f (k) = [T fl ,T fr ] T ∈R 2×1 is the control input vector of the vehicle, p f (k) = [e β (k) e ω (k)] T ∈R 2×1 is the vehicle stability output vector, f f (k)∈R 2×1 is the unknown nonlinear function vector, n i ,n0∈Z are two integers, Z is a set of integers, representing the lengths of the input and output signal sequences of the system respectively.

4. The vehicle curve adaptive cruise stability control method using differential steering according to claim 1, characterized in that: The MPC-based multi-objective weight coefficient self-adjusting fuzzy logic rule is specifically based on the workshop longitudinal kinematics model, sets the objective function, performance indicators and constraints under the MPC framework, obtains the expected acceleration in the control time domain through optimization calculation, and designs fuzzy logic rules to update the weight coefficients of performance indicators in real time to meet complex and changeable driving conditions.

5. The vehicle curve adaptive cruise stability control method using differential steering according to claim 4, characterized in that: The performance indicators include safety, following ability, economy and comfort.

6. The vehicle curve adaptive cruise stability control method using differential steering according to claim 5, characterized in that: The objective function is specifically: in, U is a set of control variables, including each control variable and its weight coefficient. The control variables include relative vehicle speed, vehicle spacing error, vehicle acceleration, and vehicle acceleration change rate. f is the weight coefficient of vehicle following and safety, L c is the weight coefficient of vehicle comfort and economy, q Δd is the weight factor of the inter-vehicle distance error, q jerk They represent the relative speed of the two vehicles, the acceleration of the vehicle and the weight factor of the acceleration change rate, ε is the vector relaxation factor, A h is the coefficient matrix of the constraint conditions, b H is the constant term matrix of the constraints.

7. The vehicle curve adaptive cruise stability control method using differential steering according to claim 6, characterized in that: The fuzzy logic rules are specifically: When the inter-vehicle distance error Δd<0, the vehicle is likely to collide with the vehicle in front. At this time, the weight coefficient of the inter-vehicle distance error q Δd It should be increased rapidly, first of all to ensure the safety of vehicles; As the inter-vehicle distance error Δd increases, the collision risk of the ego vehicle will gradually be eliminated. At this time, the weight coefficient q of the inter-vehicle distance error is reduced. Δd When the distance between vehicles increases to more than the preset threshold, the distance between vehicles is controlled to prevent the distance between vehicles from continuing to increase and vehicles in adjacent lanes from cutting in at will; When the relative speed v rel <0, it means that the vehicle speed is greater than the front vehicle speed. If the inter-vehicle distance error Δd<0 at this time, the vehicle is controlled to decelerate rapidly. If the relative speed v rel <0 and the inter-vehicle distance error Δd>0, the vehicle is controlled to decelerate slowly to improve driving economy; When the vehicle is in a dangerous state, the control quickly reduces the weight coefficient of the vehicle's acceleration. Ensure that the vehicle can respond quickly. Conversely, when the vehicle is in a safe state, increase the weight coefficient of the vehicle's acceleration. Control the vehicle to accelerate or decelerate slowly to improve driving economy.

8. The vehicle curve adaptive cruise stability control method using differential steering according to claim 1, characterized in that: In the driving and braking switching logic based on the deceleration protection range threshold, the deceleration protection range threshold is 0.1m / s 2 , when the desired deceleration of the vehicle is a des When ≥a+0.1, it is drive control; When the desired deceleration of the vehicle is a des When ≤a-0.1, it is brake control; When a - 0.1 < a des < When a + 0.1, the vehicle maintains its current state and does not switch between driving and braking; Where a is the base deceleration.

9. The vehicle curve adaptive cruise stability control method using differential steering according to claim 8, characterized in that: The specific control process of the feedforward-feedback fusion control strategy based on acceleration error is as follows: The expected driving torque T obtained by the inverse dynamics model of the driving system motor_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity during modeling on the control system, the expected acceleration a is used. des and the actual acceleration a act The error between the two is fed back and regulated, and the driving torque T motor_ref As the feedback quantity, it is the final desired driving torque; The desired brake pressure P obtained by the inverse dynamics model of the brake system m_des As the feedforward quantity, in order to eliminate the influence of the modeling error generated by the feedforward quantity during modeling on the control system, the expected deceleration a is used. des and the actual deceleration a act The error between the two is fed back and regulated, and the obtained brake pressure P m_ref The feedback amount is the final desired braking pressure.

10. The vehicle curve adaptive cruise stability control method using differential steering according to claim 9, characterized in that: The driving torque T motor_ref Specifically: Among them, T motor_ref is the expected driving torque after adjustment, K tp and K ti are the proportional and integral parameters of the drive feedback PI controller respectively; The brake pressure P m_ref Specifically: Among them, P m_ref is the expected brake pressure after adjustment, K pp and K pi are the proportional parameter and integral parameter of the brake feedback PI controller respectively.

Images