Man-machine cooperative steering control method considering driving characteristics in rainy and foggy weather

Abstract

The invention belongs to the technical field of intelligent driving control, and relates to a man-machine cooperative steering control method considering rain and fog weather characteristics, which comprises the following steps: establishing a driver model and a man-machine co-driving vehicle model considering environmental visibility and road adhesion coefficient change; the driver model dynamically adjusts a long-distance preview point and a steering wheel torque of a driver through a preview weight coefficient determined by a fuzzy control method; a man-machine co-driving vehicle model is utilized, a man-machine cooperative steering control strategy is designed through a model prediction control method, and the man-machine cooperative steering control strategy is used for dynamically adjusting the steering control quantity according to the vehicle state and the environment condition; according to the driving authority distribution strategy, a fuzzy control method is adopted to determine a man-machine cooperation coefficient under a rainy and foggy weather condition, and the man-machine cooperation coefficient is used for dynamically adjusting the man-machine cooperation degree; executing man-machine cooperative driving according to the man-machine cooperative steering control strategy and the driving authority distribution strategy; the stability and safety of the vehicle are improved in rainy and foggy weather.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent driving control, and particularly relates to a human-machine collaborative steering control method considering driving characteristics in rainy and foggy weather. Background Art

[0002] Existing driving assistance and autonomous driving systems are mainly designed for ordinary driving scenarios, without fully considering complex driving conditions such as reduced visibility and decreased road surface adhesion in rainy and foggy weather. These special conditions can lead to a mismatch between the driver's operation behavior and the vehicle's automatic control system, increasing driving risks. Improving the vehicle-road-human model and dynamically adjusting the control strategy is an effective way to solve this problem. Therefore, there is an urgent need for a steering control method that can integrate the characteristics of rainy and foggy weather and optimize human-machine collaborative control. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a human-machine collaborative steering control method considering driving characteristics in rainy and foggy weather.

[0004] To achieve the purpose of the present invention, the following technical solutions are adopted for implementation.

[0005] A human-machine collaborative steering control method considering driving characteristics in rainy and foggy weather includes the following steps:

[0006] S1. Based on the ground coordinate system and the vehicle body coordinate system, establish a driver model and a human-machine co-driving vehicle model in a rainy and foggy environment; where:

[0007] The driver model, considering the changes in environmental visibility and road surface adhesion coefficient, dynamically adjusts the long-distance preview point distance and steering wheel torque through the preview weight coefficient determined by the fuzzy control method to improve the driver's perception ability of the path, and is used to describe the visual behavior and steering control behavior of the driver when driving a vehicle in a rainy and foggy environment for path tracking.

[0008] The human-machine co-driving vehicle model describes the lateral and yaw motions of the vehicle through a two-degree-of-freedom vehicle dynamics model, and uses the magic tire model to simulate the tire force changes in a rainy and foggy environment.

[0009] S2. Use the human-machine co-driving vehicle model established in step S1 to design a human-machine collaborative steering control strategy through the model predictive control method, which is used to dynamically adjust the steering control amount according to the vehicle state and environmental conditions.

[0010] S3. Design a driving authority allocation strategy based on the fuzzy control method; the driving authority allocation strategy uses the fuzzy control method to determine the human-machine collaboration coefficient under rainy and foggy weather conditions, which is used to dynamically adjust the human-machine collaboration degree to reasonably allocate the driving authorities of humans and machines.

[0011] S4. Execute cooperative human-machine driving based on the cooperative human-machine steering control strategy and driving authority allocation strategy, in combination with the control quantities of humans and machines.

[0012] As a preferred embodiment of the present invention, the ground coordinate system is constructed with the origin O fixed at the position of the vehicle's center of mass O at the current moment, the X-axis pointing directly in front of the vehicle at the current moment, and the direction obtained by rotating the X-axis 90 degrees counterclockwise being the positive direction of the Y-axis.

[0013] As a preferred embodiment of the present invention, the vehicle body coordinate system is constructed with the origin coinciding with the vehicle's center of mass O, the X-axis pointing directly in front of the vehicle body, the direction obtained by rotating the X-axis 90 degrees counterclockwise being the positive direction of the Y-axis, and the Z-axis pointing directly above the vehicle body and perpendicular to the X-axis and Y-axis.

[0014] As a preferred embodiment of the present invention, the distance L of the long-distance preview point eff is expressed as:

[0015] L eff = κL f +(1 - κ)L n ,

[0016] wherein, L n is the distance of the near preview point N, i.e., the distance from the vehicle's center of mass O to the near point N; L f is the distance of the far preview F, i.e., the distance from the vehicle's center of mass O to the fixed far point F; L eff is the distance of the dynamically adjusted long-distance preview point, i.e., the distance from the vehicle's center of mass O to the dynamic long-distance preview point F eff ; κ is the preview weight coefficient designed based on the fuzzy control method, considering the environmental visibility and road surface adhesion coefficient.

[0017] As a preferred embodiment of the present invention, the fuzzy control rules of the preview weight coefficient κ are shown in the following table:

[0018]

[0019] As a preferred embodiment of the present invention, the basis for formulating the fuzzy control rules is: when the environmental visibility V is high and the road surface adhesion coefficient μ is high, it is considered at this time that the preview weight of the far point is large and the weight coefficient is large; when the environmental visibility V is small and the road surface adhesion coefficient μ is small, it is considered that the preview weight of the near point is large and the weight coefficient is small.

[0020] As a preferred embodiment of the present invention, the establishment process of the human-machine co-driving vehicle model includes the following steps:

[0021] S71. Establish a vehicle dynamics model:

[0022] According to the torque and torque balance equation, obtain the vehicle lateral velocity vy and the expression of vehicle yaw rate r:

[0023]

[0024] where m is the mass of the vehicle, v y is the lateral velocity of the vehicle in the vehicle coordinate system, v x is the longitudinal velocity of the vehicle in the vehicle coordinate system, r is the vehicle yaw rate, F yf is the lateral force of the vehicle's front wheels, F yr is the lateral force of the vehicle's rear wheels, I z is the moment of inertia of the vehicle about the z-axis, and a and b are the distances from the front and rear axles to the center of mass respectively;

[0025] The normal load of the tire is expressed as:

[0026]

[0027] The vehicle front wheel slip angle and the vehicle rear wheel slip angle are approximately expressed as:

[0028]

[0029] The linear tire model cannot accurately reflect the actual change trend of tire force in rain and fog environments. This method ignores the influence of tire longitudinal force and uses the Magic Tire formula under pure lateral slip conditions to calculate the lateral force of the tire:

[0030]

[0031] where μ is the road surface adhesion coefficient, F zf and F zr are the tire normal loads respectively, B f , C f and D f are the Magic Formula empirical parameters of the vehicle's front wheels, B r , C r and D r are the Magic Formula empirical parameters of the vehicle's rear wheels;

[0032] Although the Magic Formula can already represent the non-linear characteristics of vehicle tires in rain and fog environments more accurately, due to its relatively complex expression form, substituting it into the vehicle dynamics equation and integrating the vehicle model will cause the calculation task of the non-linear model predictive control algorithm to be very heavy and difficult to solve; therefore, in this method, at each sampling moment, the tire model is continuously locally linearized to obtain the linearized tire lateral force equation:

[0033]

[0034] where, represents the front wheel cornering angle of the vehicle at the current sampling moment, and represents the rear wheel cornering angle of the vehicle at the current sampling moment; represents the nominal cornering stiffness of the front wheel at the current sampling moment, and represents the nominal cornering stiffness of the rear wheel at the current sampling moment; represents the residual cornering force of the front wheel at the current sampling moment, and represents the residual cornering force of the rear wheel at the current sampling moment;

[0035] At each sampling moment, after updating the vehicle state information, the vertical load F of the front wheel of the vehicle is calculated through the above formula zf and the vertical load F of the rear wheel zr as well as the front wheel cornering angle and the rear wheel cornering angle; Substituting into the formula can calculate the front wheel cornering force at the current sampling moment and the rear wheel cornering force

[0036] The nominal cornering stiffness of the front wheel at the current sampling moment The nominal cornering stiffness of the rear wheel The residual cornering force of the front wheel and the residual cornering force of the rear wheel are calculated through the following formula

[0037]

[0038] After sorting, the vehicle dynamic equation at each sampling moment is obtained:

[0039]

[0040] In the formula, m is the mass of the vehicle, v x is the longitudinal speed of the vehicle in the vehicle coordinate system, r is the yaw angular velocity of the vehicle, F yf is the lateral force of the front wheel of the vehicle, F yr is the lateral force of the rear wheel of the vehicle, I z is the moment of inertia of the vehicle about the z-axis, a and b are the distances from the front and rear axles to the center of mass respectively is the nominal cornering stiffness of the front wheel at the current sampling moment, is the nominal cornering stiffness of the rear wheel at the current sampling moment, is the residual cornering force of the front wheel, is the residual cornering force of the rear wheel, δ f is the steering angle of the front wheel of the vehicle.

[0041] S72. Establish a vehicle kinematic model:

[0042] Under the assumption of the angle deviation ψ eIn the case of a smaller value, the lateral displacement and angular deviation can be expressed by the following equations:

[0043]

[0044] In the formula, the vertical distance between the near preview point N and the road in the vehicle's forward direction is defined as the lateral displacement deviation y e = y - y ref , where y is the lateral displacement and y ref is the reference value along the road center line; v x is the vehicle's longitudinal speed in the vehicle coordinate system; β is the vehicle's center-of-mass sideslip angle; r is the vehicle's yaw rate; L n is the distance of the near preview point N, that is, the distance from the vehicle's center of mass O to the near point N; the reference curvature ρ ref = 1 / R ref is the curvature of the inner lane of the lane, where R ref is the radius of curvature; the angular deviation ψ e is the angle between the tangent direction of the road center line in the vehicle's forward direction, and the angular deviation ψ e = ψ ref - ψ, where ψ is the vehicle's yaw angle and ψ ref is the reference along the tangent direction of the road center line of ψ.

[0045] S73. Establish a vehicle steering system model

[0046] The relationship between the vehicle's steering wheel angle δ s and the vehicle's front wheel angle δ f is as follows:

[0047] δ s = g s ·δ f ,

[0048] where δ s is the vehicle's steering wheel angle; g s is the vehicle steering system transmission ratio coefficient;

[0049] According to the vehicle dynamics principle, the expression of the wheel return torque of the steering system is as follows:

[0050]

[0051] where K s is the return torque coefficient and the sideslip angle α f in the linear region model is expressed as follows:

[0052] K s = -K p C f η t ,

[0053]

[0054] Among them, η t is the sum of the tire drag distance and the tilt moment arm, and K p is the steering system coefficient;

[0055] The total externally applied torque T tot has the following expression:

[0056] T tot = λT auto +(1 - λ)T dr ,

[0057] Among them, T auto is the torque output by the automation system, T dr is the torque output by the driver model, and λ is the cooperation coefficient;

[0058] The torque balance equation of the vehicle steering system is as follows:

[0059]

[0060] Among them, b s and J s are the friction coefficient and moment of inertia of the steering column respectively, and ω s is the angular velocity of the steering wheel.

[0061] S74. Establish a human-machine co-driving vehicle model

[0062] Select the angular velocity ω s of the vehicle's steering wheel, the steering wheel angle δ s , the sideslip angle β of the center of mass, the yaw rate r, the lateral displacement deviation y e and the yaw angle deviation ψ e as the system states, the total torque T tot of the vehicle's steering wheel as the system input, and the lateral displacement deviation y e and the yaw angle deviation ψ e of the vehicle as the system output;

[0063] After organizing the above vehicle model, the vehicle model in rainy and foggy environments can be written in state-space form as follows:

[0064]

[0065] Among them, w = ρ ref is the external input, the state vector and the output vector are x = [ω s , δ s , β, r, y e , ψ e and y = [y e, ψ e and the control input u = T tot , the system matrix is:

[0066]

[0067] D d = [0000 - v x L n v x Τ ,

[0068]

[0069] For the convenience of controller design, the above state - space model is discretized by Euler method to obtain the discretized vehicle model:

[0070]

[0071] where, C d = C0, T s is the sampling time.

[0072] As a preferred solution of the present invention, the design process of the human - machine collaborative steering control strategy includes the following steps:

[0073] S81. Design the human - machine torque collaborative automation controller

[0074] Define the control quantity sequence U(k) as:

[0075]

[0076] Assume that the prediction horizon is P steps, the control horizon is N steps, and N ≤ P. At the same time, assume that the control quantities outside the control horizon remain unchanged, that is, u(k + N) = u(k + N + 1) = … = u(k + P - 1), and derive the prediction equations for the next P steps:

[0077]

[0078] where x(k + i) is the system state quantity at time k + i, i = 0, 1, …, P; u(k + i) is the optimized quantity at time k + i, i = 0, 1, …, P - 1; w(k + i) is the road curvature at time k + i, i = 0, 1, …, P - 1;

[0079] The output prediction equations within the prediction horizon of P steps are as follows:

[0080]

[0081] where y(k + i) is the system output quantity at time k + i, i = 0, 1, …, P;​

[0082] S82. Describe the trajectory tracking problem as the following constrained optimization problem:

[0083]

[0084] where R(K + 1)=[r(k + 1), r(k + 2), …, r(k + p)] Τ 2p×1 is the reference vector; the control increment vector ΔU(k)=[Δu(k), Δu(k + 1), …, Δu(k + m - 1)] Τ m×1 is the independent variable of the constrained optimization problem; the output Y(k + 1|k)=[y(k + 1|k), y(k + 2|k),..., y(k + p|k)] Τ 2p×1 of the prediction horizon P is predicted by the system model at time k; according to the actuator saturation of the steering system, introduce the control input constraint u max (k)= -u min (k); the state constraint Hx(k) ≤ G is defined by the stable handling envelope to ensure vehicle stability, which is related to the limitations of vehicle roll angle and yaw rate; these limitations show the maximum capacity of the given tire and are based on the steady-state assumption; the vehicle yaw rate limit is:

[0085]

[0086] where g is the acceleration due to gravity and μ is the road surface adhesion coefficient. The limitation condition of the vehicle center of mass sideslip angle is:

[0087]

[0088] where is the slip angle related to the maximum lateral force, and the state constraint matrices H and G are respectively:

[0089]

[0090] Solve the above constrained optimization problem to obtain the optimal solution u(k) at time k.

[0091] As a preferred solution of the present invention, the fuzzy control rules of the human - machine cooperation coefficient are shown in the following table:

[0092]

[0093] As a preferred solution of the present invention, the basis for formulating the fuzzy control rules: when the lateral displacement deviation |y e | is large and the angle deviation |ψ eWhen it is larger, it is considered that the driver is greatly affected by the rain and fog environment, and the cooperation coefficient is high or the vehicle's lateral displacement is completely controlled by the auxiliary system; when the lateral displacement deviation |y e | is small and the angle deviation |ψ e | is small, it is considered that the driver has a good ability to adapt to the rain and fog environment, and the cooperation coefficient should be small or small.

[0094] Beneficial effects

[0095] 1. By considering the visibility and road surface adhesion coefficient in the rain and fog environment, the present invention establishes a two-point preview model and dynamically adjusts the far-point preview distance combined with the dynamic preview weight coefficient adjustment method based on fuzzy control, enabling the vehicle to better cope with complex weather conditions and continuously optimize its driving trajectory in a changing environment, which can effectively improve the system stability adaptability and driving safety of the vehicle under low visibility conditions;

[0096] 2. Through effective human-machine cooperative control, the present invention reduces the driver's pressure on path tracking and improves the driver's sense of security and driving comfort under weather conditions such as rain and fog with low visibility and low road surface adhesion coefficient;

[0097] 3. The present invention optimizes the control strategy through the model predictive control method and combines fuzzy logic for driving authority allocation to ensure the effective cooperation between the automated system and the driver, achieving real-time rolling optimization control. This control method can effectively avoid emergencies and make the system respond more efficiently to different driving demands. Brief description of the drawings

[0098] Figure 1 is a flow diagram of a human-machine cooperative steering control method considering the characteristics of rain and fog weather according to the present invention;

[0099] Figure 2 is a control framework diagram of a human-machine cooperative steering control method considering the characteristics of rain and fog weather according to the present invention;

[0100] Figure 3 is a schematic diagram of a two-point preview driver model in the present invention;

[0101] Figure 4 is a schematic diagram of a path tracking model in the present invention;

[0102] Figure 5 is a schematic diagram of the environmental visibility membership function in this method;

[0103] Figure 6 is a schematic diagram of the road surface adhesion coefficient membership function in this method;

[0104] Figure 7 is a schematic diagram of the weight coefficient membership function in this method;

[0105] Figure 8 Schematic diagram of the weight coefficient fuzzy rule surface in this method;

[0106] Figure 9 Schematic diagram of the membership function of the lateral displacement deviation in this method;

[0107] Figure 10 Schematic diagram of the membership function of the yaw angle deviation in this method;

[0108] Figure 11 Schematic diagram of the membership function of the cooperation coefficient in this method;

[0109] Figure 12 Schematic diagram of the cooperation coefficient fuzzy rule surface in this method. Specific implementation manner

[0110] The present invention will be further described in conjunction with the embodiments and the accompanying drawings.

[0111] As an embodiment of the present invention, as Figure 1 shown, a human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather includes the following steps:

[0112] Step 1. Establish a driver model and a vehicle model in a rainy and foggy environment

[0113] Establish a ground coordinate system, with the origin O fixed at the position of the vehicle's center of mass O at the current moment, the X-axis pointing directly in front of the vehicle at the current moment, and the positive direction of the Y-axis being the direction obtained by rotating the X-axis counterclockwise by 90 degrees;

[0114] Establish a vehicle body coordinate system, with the origin coinciding with the vehicle's center of mass O, the X-axis pointing directly in front of the vehicle body, and the X-axis

[0115] rotated counterclockwise by 90 degrees to be the positive direction of the Y-axis, and the Z-axis pointing directly above the vehicle body and perpendicular to the X-axis and the Y-axis;

[0116] (1) Establish a driver model

[0117] The driver preview model describes the visual behavior and steering control behavior of the driver in path tracking. By dynamically adjusting the weights, the influence of environmental variables on driving behavior is incorporated into the model to form a more robust driver model. Provide a driver preview model with adjustable preview weight coefficients, and optimize the preview weight coefficients through a fuzzy controller to achieve real-time adjustment of the preview behavior, thereby improving the driver's perception ability of the path and enhancing driving safety and adaptability.

[0118] As Figure 4As shown, the two-point preview driver model combines the road information of two regions of the far and near roads. The far point F is used as the prediction point, reflecting the general direction to be reached in the future. The near point N is used as the compensation point to gradually adjust the driver to the desired trajectory. This method dynamically adjusts the far point preview distance L considering the environmental visibility and road surface adhesion coefficient eff , the dynamically adjusted long-distance preview point distance L eff can improve the driving stability, comfort, response speed and safety of the vehicle in rainy and foggy environments, help enhance the adaptability of the system, and ensure that the system can make optimized decisions according to real-time environmental conditions. The following formula is the expression of the long-distance preview point distance L eff :

[0119] L eff =κL f +(1-κ)L n

[0120] Among them, L n is the distance of the near preview point N, that is, the distance from the vehicle center of mass O to the near point N; L f is the distance of the far preview F, that is, the distance from the vehicle center of mass O to the fixed far point F; L eff is the dynamically adjusted long-distance preview point distance, that is, the distance from the vehicle center of mass O to the dynamic long-distance preview point F eff ; κ is the preview weight coefficient designed based on the fuzzy control method, considering the environmental visibility and road surface adhesion coefficient.

[0121] The long-distance preview angle θ f is the angle between the direction from the vehicle center of mass O to the dynamic far point F eff and the front of the vehicle body; the near preview angle θ n is the angle between the direction from the vehicle center of mass O to the near point N and the front of the vehicle body; the near angle θ n and the far angle θ f previewed by the driver in the driving direction can be approximated according to geometric relations and kinematic principles as:

[0122]

[0123] Among them, the lateral displacement deviation y e is the vertical distance from the near point N to the vehicle's forward direction, R ref is the radius of curvature of the vehicle center of mass O's trajectory, and the yaw angle deviation ψ e is the angle between the tangent direction of the road center line in the vehicle's forward direction.

[0124] The feedforward control and compensation control can be expressed as follows:

[0125]

[0126] Among them, K ais the proportional gain of the long-distance preview angle θ f and reflects the driver's perception of the long-distance preview angle θ f ; among them, K c is the proportional gain of the near preview angle θ n and reflects the driver's perception of the near preview angle θ n ; T L is the lead time constant of the driver model; T I is the lag time constant of the driver model.

[0127] The driver response lag and the muscle response lag can be expressed as follows:

[0128]

[0129] Among them, τ p is the time constant of the delay link, and T N is the time constant of the driver's arm dynamics model.

[0130] The driver's response to the feedback torque on the steering wheel and the compensation for the resistance torque of the steering column can be expressed as follows:

[0131]

[0132] Among them, K d is the proportional gain of the sensing link, K G is the proportional gain of the action link, T1 is the time constant of the sensing link, and T K1 is the lead time constant of the action link, and T K2 is the lag time constant of the action link.

[0133] As Figure 3 shown, through the collaborative action of the multi-level control module, the driver model finally outputs the steering wheel torque T dr to achieve the human-machine collaborative steering control. In the figure, the driver model uses the long-distance preview angle θ f and the near preview angle θ n as inputs, and after being processed by a series of transfer functions, generates the steering wheel torque T dr .

[0134] Since the relationship between the environmental visibility V, the road surface adhesion coefficient μ, and the weight coefficient κ cannot be accurately expressed by a mathematical formula, a fuzzy control method is adopted to determine the dynamic weight coefficient. The range of the environmental visibility V is [0, 500], the range of the road surface adhesion coefficient μ is [0, 1], and the basic domain of the weight coefficient κ is [0, 1]. The fuzzy subsets of the environmental visibility V and the road surface adhesion coefficient μ are {L, ML, M, MH, H}, representing the five states of low, relatively low, moderate, relatively high, and high of the environmental visibility V and the road surface adhesion coefficient μ respectively. The fuzzy subset of the weight coefficient κ is {L, ML, M, MH, H}, representing the five states of low, relatively low, moderate, relatively high, and high of the cooperation coefficient respectively. Among them, the function images of the environmental visibility V and the road surface adhesion coefficient μ are respectively as Figure 5 and Figure 6 shown, and the weight coefficient κ is as Figure 7 shown. Since the basic domains of the environmental visibility V and the road surface adhesion coefficient μ are 5 respectively, a total of 5 * 5 = 25 rules need to be defined. The basis for formulating the fuzzy rules is as follows: when the environmental visibility V is relatively high and the road surface adhesion coefficient μ is relatively high, it is considered that the preview weight at the far point is large, and the weight coefficient is large; when the environmental visibility V is small and the road surface adhesion coefficient μ is small, it is considered that the preview weight at the near point is large, and the weight coefficient is small. The specific rules are shown in Table 1, and the fuzzy rule surface of the weight coefficient is as Figure 8 shown.

[0135] Table 1 Fuzzy control rules for the dynamic weight coefficient κ

[0136]

[0137] (2) Establishment of the vehicle dynamics model

[0138] In this method, the vehicle dynamics model uses a two-degree-of-freedom vehicle model to represent the lateral and yaw motions of the vehicle. According to the moment and torque balance equations, the expressions for the vehicle lateral velocity v y and the vehicle yaw rate r are as follows:

[0139]

[0140] Among them, m is the mass of the vehicle, v y is the vehicle lateral velocity in the vehicle coordinate system, v x is the vehicle longitudinal velocity in the vehicle coordinate system, r is the vehicle yaw rate, F yf is the lateral force of the front wheels of the vehicle, F yr is the lateral force of the rear wheels of the vehicle, I z is the moment of inertia of the vehicle about the z-axis, and a and b are the distances from the front and rear axles to the center of mass respectively.

[0141] The normal load of the tire is as follows:

[0142]

[0143] The sideslip angles of the front wheels and rear wheels of the vehicle are approximated as follows:

[0144]

[0145] The linear tire model cannot accurately reflect the actual change trend of tire forces in rainy and foggy environments. This method ignores the influence of tire longitudinal forces and uses the Magic Tire formula under the pure lateral slip condition to calculate the lateral force of the tire, as shown in the following formula:

[0146]

[0147] where μ is the road surface adhesion coefficient, F zf and F zr are the normal loads of the tire respectively, B f , C f and D f are the Magic Formula empirical parameters of the front wheels of the vehicle, and B r , C r and D r are the Magic Formula empirical parameters of the rear wheels of the vehicle.

[0148] Although the Magic Formula can already represent the non-linear characteristics of vehicle tires in rainy and foggy environments relatively accurately, due to its relatively complex expression form, substituting it into the vehicle dynamics equation and integrating the vehicle model will cause the calculation task of the non-linear model predictive control algorithm to be very heavy and difficult to solve; therefore, in this method, at each sampling moment, the tire model is continuously locally linearized to obtain the linearized tire lateral force equation, as follows:

[0149]

[0150] where represents the sideslip angle of the front wheels of the vehicle at the current sampling moment, represents the sideslip angle of the rear wheels of the vehicle at the current sampling moment; represents the nominal sideslip stiffness of the front wheels at the current sampling moment, represents the nominal sideslip stiffness of the rear wheels at the current sampling moment; represents the residual sideslip force of the front wheels at the current sampling moment, represents the residual sideslip force of the rear wheels at the current sampling moment;

[0151] At each sampling moment, after updating the vehicle state information, the vertical load F zf of the front wheels and the vertical load F zr of the rear wheels of the vehicle, as well as the sideslip angles of the front wheels and rear wheels, are calculated through the above formula; substituting into the formula can calculate the lateral force of the front wheels at the current sampling moment and the rear wheel cornering force

[0152] The nominal cornering stiffness of the front wheels at the current sampling moment The nominal cornering stiffness of the rear wheels The residual cornering force of the front wheels and the residual cornering force of the rear wheels are calculated by the following formula

[0153]

[0154] After arrangement, the vehicle dynamic equation at each sampling moment is as follows:

[0155]

[0156] (3) Vehicle kinematic model

[0157] The angle deviation ψ e is the angle between the tangent direction of the center line of the road in the vehicle's forward direction, where ψ e = ψ ref - ψ, ψ is the yaw angle of the vehicle, ψ ref is the reference along the tangent direction of the center line of the road. Assuming ψ e is small, the lateral displacement and angle deviation can be expressed by the following formula:

[0158]

[0159] (4) Vehicle steering system model

[0160] The steering wheel angle δ of the vehicle s and the front wheel angle δ of the vehicle f are related as follows:

[0161] δ s = g s ·δ f

[0162] where δ s is the steering wheel angle of the vehicle; g s is the transmission ratio coefficient of the vehicle steering system;

[0163] According to the vehicle dynamics principle, the expression of the wheel return moment of the steering system is as follows:

[0164]

[0165] where K s is the return moment coefficient and the side slip angle α f in the linear region model is expressed as follows:

[0166] K s = -K p C f η t

[0167]

[0168] where η t is the sum of the tire drag distance and the tilt moment arm, and K p is the steering system coefficient.

[0169] The expression of the total externally applied torque T tot is as follows:

[0170] T tot = λT auto + (1 - λ)T dr

[0171] where T auto is the torque output by the automation system, T dr is the torque output by the driver model, and λ is the cooperation coefficient.

[0172] The torque balance equation of the vehicle steering system is as follows:

[0173]

[0174] where b s and J s are the friction coefficient and moment of inertia of the steering column respectively, and ω s is the angular velocity of the steering wheel.

[0175] (5) Establish a co-driving vehicle model

[0176] Select the angular velocity ω s of the vehicle's steering wheel, the steering wheel angle δ s , the sideslip angle β of the center of mass, the yaw rate r, the lateral displacement deviation y e and the yaw angle deviation ψ e as the system states, and the total torque T tot of the vehicle's steering wheel as the system input, and the lateral displacement deviation y e and the yaw angle deviation ψ e of the vehicle as the system outputs.

[0177] After organizing the above vehicle model, the co-driving vehicle model in the rain and fog environment can be written in the state space form as follows:

[0178]

[0179] where w = ρ ref is the external input, and the state vector and output vector are x = [ωs , δ s , β, r, y e , ψ e and y = [y e , ψ e and the control input u = T tot , the system matrix is:

[0180]

[0181] (2) For the convenience of controller design, the above state - space model is discretized by Euler method, and the discrete vehicle model is obtained as follows:

[0182]

[0183] Where, C d = C0, T s is the sampling time.

[0184] Step 2: Using the vehicle model in Step 1, the model predictive control method is adopted to design the human - machine torque collaborative automation controller

[0185] (1) Design of the human - machine torque collaborative automation controller

[0186] Define the control quantity sequence U(k) as:

[0187]

[0188] Assume that the prediction horizon is P steps, the control horizon is N steps, and N ≤ P. At the same time, assume that the control quantities outside the control horizon remain unchanged, that is, u(k + N) = u(k + N + 1) =... = u(k + P - 1). The following prediction equations for the next P steps can be deduced:

[0189]

[0190] Where, x(k + i) is the system state quantity at time k + i, i = 0, 1,..., P; u(k + i) is the optimized quantity at time k + i, i = 0, 1,..., P - 1; w(k + i) is the road curvature at time k + i, i = 0, 1,..., P - 1k + i;

[0191] The output prediction equations within the prediction horizon of P steps are as follows:

[0192]

[0193]

[0194] Where y(k + i) is the system output quantity at time k + i, i = 0, 1,..., P;

[0195] (2) The trajectory tracking problem can be described as the following optimization problem:

[0196]

[0197] s.t. x(k + 1) = A d x(k) + B d u(k) + D d w(k)

[0198] Hx(k) ≤ G

[0199] Δu(k) = u(k) - u(k - 1)

[0200] u min (k + i) ≤ u(k + i) ≤ u max (k + i), i = 0, 1, …, m - 1

[0201] u(k + i) = 0, i = m, m + 1, …, p - 1

[0202] where R(K + 1) = [r(k + 1), r(k + 2), …, r(k + p)] Τ 2p×1 is the reference vector. The control increment vector ΔU(k) = [Δu(k), Δu(k + 1), …, Δu(k + m - 1)] Τ m×1 is the independent variable of the constrained optimization problem. The output of the prediction horizon P, Y(k + 1|k) = [y(k + 1|k), y(k + 2|k),..., y(k + p|k)] Τ 2p×1 is predicted by the system model at time k. According to the actuator saturation of the steering system, the control input constraint u max (k) = -u min (k). The state constraint Hx(k) ≤ G is defined by the stable handling envelope to ensure vehicle stability, which is related to the limits of vehicle roll angle and yaw rate. These limits show the maximum capacity of the given tires and are based on the steady-state assumption. The vehicle yaw rate limit is

[0203]

[0204] where g is the gravitational acceleration and μ is the road surface adhesion coefficient. The limit condition of the vehicle sideslip angle of the center of mass is

[0205]

[0206] where is the slip angle related to the maximum lateral force, and the state constraint matrices H and G are respectively

[0207]

[0208] Solve the above constrained optimization problem to obtain the optimal solution u(k) at time k;

[0209] Step 3: Design a driving authority allocation strategy based on the fuzzy logic method

[0210] Use fuzzy control to determine the cooperation coefficient under rainy and foggy weather conditions. Since the lateral displacement deviation y e and the angular deviation ψ e are the input components of the preview driver model and the automation system, and they reflect the accuracy of the vehicle in terms of tracking accuracy. Since the signs of the variables y e and ψ e only represent opposite directions, the absolute values |y e | and |ψ e | are used instead.

[0211] The basic domain of the lateral displacement deviation y e is [0, 1.4], the basic domain of the angular deviation |ψ e | is [0, 0.25], and the basic domain of the cooperation coefficient λ is [0, 1]. The fuzzy subsets of the vehicle lateral displacement deviation |y e | and the vehicle angular deviation |ψ e are {S, MS, M, MB, B}, representing the five states of small, relatively small, medium, relatively large, and large of the vehicle lateral displacement deviation |y e | and the vehicle angular deviation |ψ e | respectively. The fuzzy subset of the cooperation coefficient λ is {S, MS, M, MB, B}, representing the five states of small, relatively small, medium, relatively large, and large of the cooperation coefficient respectively. Among them, the function images of the vehicle lateral displacement deviation |y e | and the yaw angle deviation |ψ e | are shown in Figure 9 and Figure 10 respectively, and the cooperation coefficient λ is shown in Figure 11 respectively. Since the basic domains of the vehicle lateral displacement deviation |y e | and the vehicle angular deviation |ψ e | are both 5, a total of 5 * 5 = 25 rules need to be defined. The basis for formulating the fuzzy rules is: when the lateral displacement deviation |y e | is relatively large and the angular deviation |ψ e | is relatively large, it is considered that the driver is greatly affected by the rainy and foggy environment at this time, and the cooperation coefficient is high or the vehicle lateral displacement is completely controlled by the auxiliary system; when the lateral displacement deviation |y e | is relatively small and the angular deviation |ψ eWhen it is relatively small, it is considered that the driver has a good ability to adapt to the rain and fog environment, so the cooperation coefficient should be relatively small or small. The specific rules are shown in Table 2, and the fuzzy rule surface of the cooperation coefficient is as Figure 12 shown.

[0212] Table 2 Fuzzy control rules for the cooperation coefficient λ

[0213]

[0214] Step 4: Execute the human-vehicle cooperative steering control:

[0215] As Figure 2 shown in the control framework diagram, based on the input of the road surface adhesion coefficient and the environmental visibility, the driver torque is generated through the preview model and the driver model. Combining the system torque output by the automation system, the driving authority is allocated through the cooperation coefficient to form the total torque acting on the vehicle steering system, driving the vehicle dynamics to update the state, and acting on the driver and the automation system through the state feedback; at the next moment, according to the updated vehicle state and environmental conditions, the control quantity is recalculated, the constraint optimization problem of the human-machine torque cooperative steering control is solved, and a new one is selected to act on the steering system. This process is repeated to achieve the rolling optimization control of the human-machine cooperative steering in rainy and foggy weather. The control quantity is applied to the vehicle steering system, and the corresponding feedback effect is generated according to the vehicle state and acts on the driver and the automation system. At the next moment, the constraint optimization problem of the human-machine torque cooperative steering control method is re-solved according to the current vehicle state, the control quantity is selected to act on the vehicle steering system, and the corresponding feedback effect is generated according to the vehicle state and acts on the driver and the automation system. This process is repeated to achieve the rolling optimization control.

[0216] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the rights of the embodiments of the present application. Any modification, equivalent replacement, and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the rights of the embodiments of the present application.

Claims

1. A human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather, characterized by: The steps include: S1. Based on the ground coordinate system and the vehicle body coordinate system, a driver model and a human-machine co-driving vehicle model in a rainy and foggy environment are established; wherein: The driver model dynamically adjusts the long-distance preview point distance and steering wheel torque by using the preview weight coefficient determined by the fuzzy control method, taking into account the changes in environmental visibility and road adhesion coefficient, so as to improve the driver's perception of the path. It is used to describe the driver's visual behavior and steering control behavior when driving a vehicle in rainy and foggy environments while performing path tracking. The human-machine co-driving vehicle model uses a two-degree-of-freedom vehicle dynamics model to describe the lateral and yaw motion of the vehicle, and uses a magic tire model to simulate tire force changes in rainy and foggy environments; S2, using the human-machine co-driving vehicle model established in step S1, designing a human-machine collaborative steering control strategy through a model predictive control method, which is used to dynamically adjust the steering control amount according to the vehicle state and environmental conditions; S3. Designing a driving authority allocation strategy based on a fuzzy control method; the driving authority allocation strategy adopts a fuzzy control method to determine the human-machine coordination coefficient under rainy and foggy weather conditions, which is used to dynamically adjust the degree of human-machine coordination to reasonably allocate driving authority between humans and machines; S4. According to the human-machine collaborative steering control strategy and driving authority allocation strategy, combined with the human-machine control quantity, perform human-machine collaborative driving.

2. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1 is characterized in that: The ground coordinate system is constructed with the origin O fixed to the position of the vehicle mass center O at the current moment, the X-axis points to the front of the vehicle body at the current moment, and the direction of the X-axis rotated 90 degrees counterclockwise is the positive direction of the Y-axis.

3. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1 is characterized in that: The vehicle body coordinate system is constructed with the origin coinciding with the vehicle mass center O, the X-axis pointing to the front of the vehicle body, the X-axis rotated 90 degrees counterclockwise to the positive direction of the Y-axis, and the Z-axis pointing to the top of the vehicle body and perpendicular to the X-axis and the Y-axis.

4. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1 is characterized in that: The long-distance preview point distance is expressed as: L eff =κL f +(1-k)L n , Where, L n is the distance from the near preview point N, that is, the distance from the vehicle's center of mass O to the near point N; L f is the distance of the far preview F, that is, the distance from the vehicle's center of mass O to the fixed far point F; L eff is the distance of the dynamically adjusted long-distance preview point, that is, the distance from the vehicle's center of mass O to the dynamic long-distance preview point F eff distance; κ is the preview weight coefficient designed based on the fuzzy control method, taking into account the environmental visibility and road adhesion coefficient.

5. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 4 is characterized in that: The fuzzy control rules of the preview weight coefficient κ are shown in the following table:

6. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 5 is characterized in that: The basis for formulating the fuzzy control rules is: when the environmental visibility V is high and the road adhesion coefficient μ is high, it is considered that the far point preview weight is large and the weight coefficient is large; when the environmental visibility V is small and the road adhesion coefficient μ is small, it is considered that the near point preview weight is large and the weight coefficient is small.

7. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1 is characterized in that: The process of establishing the human-machine co-driving vehicle model includes the following steps: S71. Establishing vehicle dynamics model: According to the moment and torque balance equation, the vehicle lateral velocity v is obtained y And the vehicle yaw rate r expression: Where m is the mass of the vehicle, v y is the lateral velocity of the vehicle in the vehicle coordinate system, v x is the longitudinal velocity of the vehicle in the vehicle coordinate system, r is the yaw angular velocity of the vehicle, F yf is the lateral force of the front wheel of the vehicle, F yr is the lateral force of the rear wheel of the vehicle, I z is the moment of inertia of the vehicle around the z-axis, a and b are the distances from the front and rear wheel axles to the center of mass, respectively; The normal load on the tire is expressed as: The vehicle front wheel slip angle and the vehicle rear wheel slip angle are approximately expressed as: The linear tire model cannot accurately reflect the actual change trend of tire force in rainy and foggy environments. This method ignores the influence of tire longitudinal force and uses the magic tire formula under pure lateral slip conditions to calculate the tire lateral force: Where μ is the road adhesion coefficient, F zf and F zr are tire normal load, B f , C f and D f is the MagicFormula empirical parameter of the front wheel of the vehicle, B r , C r and D r Magic Formula empirical parameters for the rear wheels of the vehicle; Although the magic formula can accurately represent the nonlinear characteristics of vehicle tires in rainy and foggy environments, its expression is relatively complex. Substituting it into the vehicle dynamics equation and integrating the vehicle model will result in a very heavy calculation task for the nonlinear model predictive control algorithm and make it difficult to solve. Therefore, this method performs continuous local linearization on the tire model at each sampling moment to obtain the linearized tire lateral force equation: in, Indicates the front wheel slip angle of the vehicle at the current sampling moment, Indicates the rear wheel slip angle of the vehicle at the current sampling moment; represents the nominal cornering stiffness of the front wheel at the current sampling moment, Indicates the nominal cornering stiffness of the rear wheel at the current sampling moment; Represents the residual cornering force of the front wheel at the current sampling moment, Indicates the residual cornering force of the rear wheel at the current sampling moment; At each sampling moment, after updating the vehicle status information, the vehicle's front wheel vertical load F is calculated by the above formula zf and rear wheel vertical load F zr As well as the front wheel slip angle and the rear wheel slip angle; Substituting into the formula, the front wheel slip force at the current sampling moment can be calculated and rear wheel cornering force Nominal cornering stiffness of the front wheel at the current sampling moment Nominal cornering stiffness of rear wheel Residual cornering force on the front wheel and the residual cornering force of the rear wheel By calculating the following formula, The vehicle dynamic equation at each sampling moment is obtained by sorting: Where m is the mass of the vehicle, β is the side slip angle v of the center of mass of the vehicle x is the longitudinal velocity of the vehicle in the vehicle coordinate system, r is the yaw angular velocity of the vehicle, F yf is the lateral force of the front wheel of the vehicle, F yr is the lateral force of the rear wheel of the vehicle, I z is the moment of inertia of the vehicle around the z-axis, a and b are the distances from the front and rear wheel axles to the center of mass, respectively. is the nominal cornering stiffness of the front wheel at the current sampling moment, is the nominal cornering stiffness of the rear wheel at the current sampling moment, is the residual cornering force of the front wheel, is the residual cornering force of the rear wheel, δ f is the front wheel turning angle of the vehicle; S72. Establishing vehicle kinematics model: Assuming an angular deviation ψ e In smaller cases, the lateral displacement and angle deviations are expressed as follows: In the formula, the vertical distance between the near-preview point N and the road in the vehicle's forward direction is defined as the lateral displacement deviation y e =yy ref , where y is the lateral displacement, y ref is the reference value along the center line of the road; v x is the longitudinal velocity of the vehicle in the vehicle coordinate system; β is the sideslip angle of the vehicle's center of mass; r is the vehicle's yaw rate; L n is the distance from the near-preview point N, that is, the distance from the vehicle's center of mass O to the near point N; the reference curvature ρ ref =1 / R ref is the curvature of the lane line, where R ref is the radius of curvature; the angular deviation ψ e is the angle between the centerline of the road and the tangent of the vehicle’s forward direction. The angle deviation ψ e =ψ ref -ψ, where ψ is the yaw angle of the vehicle, ψ ref is the reference of ψ along the tangent direction of the road centerline; S73. Establishing vehicle steering system model Vehicle steering wheel angle δ s and the vehicle front wheel steering angle δ f The relationship is as follows: d s =g s ·d f , Among them, δ s is the vehicle steering wheel angle; g s is the transmission ratio coefficient of the vehicle steering system; According to the principle of vehicle dynamics, the expression of the wheel aligning torque of the steering system is as follows: Among them, K s is the righting moment coefficient and the sideslip angle α f The expression of the model in the linear region is as follows: K s =-K p C f or t , Among them, η t is the sum of the tire drag and the tilt moment arm, K p is the steering system coefficient; Externally applied total torque T tot The expression is as follows: T tot =λT auto +(1-λ)T dr , Among them, T auto is the torque output by the automation system, T dr is the torque output by the driver model, λ is the synergy coefficient; The torque balance equation of the vehicle steering system is as follows: Among them, b s and J s are the friction coefficient and moment of inertia of the steering column, ω s is the steering wheel angular velocity. S74. Establish a human-machine co-driving vehicle model Select the vehicle's steering wheel angular velocity ω s , steering wheel angle δ s , center of mass sideslip angle β, yaw rate r, lateral displacement deviation y e and the yaw angle deviation ψ e As the system state, the vehicle's total steering wheel torque T tot As the system input, the lateral displacement deviation y of the vehicle e and the yaw angle deviation ψ e As system output; After arranging the vehicle dynamics model, vehicle kinematics model and vehicle steering system model, the vehicle model in rainy and foggy environment can be written in state space form as follows: Where w = ρ ref is the external input, the state vector and output vector are x=[ω s ,δ s ,β,r,y e ,ψ e ] and y=[y e ,ψ e ] and control input u=T tot , the system matrix is: D d =[0 0 0 0 -v x L n v x ] Τ , In order to facilitate the controller design, the above state space model is discretized by Euler to obtain the discretized vehicle model: in, C d =C0,T s is the sampling time.

8. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1 is characterized in that: The design process of the human-machine collaborative steering control strategy includes the following steps: S81. Design of human-machine torque collaborative automation controller Define the control quantity sequence U(k) as: Assuming that the prediction time domain is P steps, the control time domain is N steps, and N≤P, and assuming that the control quantity outside the control time domain remains unchanged, that is, u(k+N)=u(k+N+1)=…=u(k+P-1), the prediction equation for the next P steps is derived: Among them, x(k+i) is the system state at time k+i, i=0,1,…,P; u(k+i) is the optimization quantity at time k+i, i=0,1,…,P-1; w(k+i) is the road curvature at time k+i, i=0,1,…,P-1k+i; The output prediction equation within the prediction time domain P step is as follows: Where y(k+i) is the system output at time k+i, i=0,1,…,P; S82. The trajectory tracking problem is described as the following constrained optimization problem: Where R(K+1)=[r(k+1),r(k+2),…,r(k+p)] Τ 2p×1 is the reference vector; the control increment vector ΔU(k)=[Δu(k),Δu(k+1),...,Δu(k+m-1)] Τ m×1 is the independent variable of the constrained optimization problem; the output of the prediction domain P is Y(k+1|k)=[y(k+1)|k,y(k+2)|k,…,y(k+p)|k] Τ 2p×1 is predicted by the system model at time k; according to the actuator saturation of the steering system, the control input constraint u is introduced max (k) = -u min (k); the state constraint Hx(k)≤G is defined by the stable handling envelope to ensure vehicle stability, which is related to the limits of the vehicle roll angle and yaw rate; these limits show the maximum capacity of the given tires and are based on the steady-state assumption; the vehicle yaw rate limit is: Where g is the acceleration of gravity and μ is the road adhesion coefficient. The restriction condition of the vehicle's center of mass sideslip angle is: in, is the slip angle associated with the maximum lateral force, and the state constraint matrices H and G are: Solve the above constrained optimization problem and obtain the optimal solution u(k) at time k.

9. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 1, characterized in that: The fuzzy control rules of the human-machine collaboration coefficient are shown in the following table:

10. The human-machine cooperative steering control method considering driving characteristics in rainy and foggy weather according to claim 9, characterized in that: The basis for formulating the fuzzy control rules is: when the lateral displacement deviation |y e |large and angular deviation|ψ e | is large, it is considered that the driver is greatly affected by the rain and fog environment, the coordination coefficient is high or the lateral displacement of the vehicle is completely controlled by the auxiliary system; when the lateral displacement deviation |y e | is small and the angle deviation |ψ e |When it is smaller, it is considered that the driver has better ability to adapt to the rainy and foggy environment, and the synergy coefficient should be smaller or small.

Images