A three-axle commercial vehicle double-stage drift control method and safety auxiliary system for coping with extreme working conditions

Abstract

The application discloses a three-axle commercial vehicle double-stage drift control method and a safety auxiliary system for coping with extreme working conditions, and builds a double-track three-axle vehicle path tracking model containing a nonlinear tire. According to the characteristics of a curve, a steady-state drift state variable is calculated, and stability analysis is performed on a state variable root locus; based on a time-varying model prediction algorithm, an expected steady-state drift balance state is tracked, auxiliary drift control is realized, pre-look road information is obtained, a lateral error e and a heading error ΔΨ are calculated, and control modes are switched according to the two errors; when the curve is close to the end, a heading stability control integrated with active front steering and additional yaw moment is switched, so that the vehicle body posture is restored to be stable. The above obtained control instructions are transmitted to a wheel drive assembly and a steer-by-wire mechanism for real-time execution. The application realizes the fusion of curve drift and straight path, improves the deficiency that the existing method can only perform drift control in a steady-state circular working condition, and meets the drift driving demand of an unmanned commercial vehicle in a complex working condition.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of commercial vehicle dynamics control, in particular to a two-stage auxiliary drift control method and safety auxiliary system for three-axle commercial vehicles in extreme working conditions. BACKGROUND

[0002] The automatic driving technology shows explosive growth and more vigorous development in the automobile industry, aiming to solve the increasingly prominent energy shortage and road safety problems. Many auxiliary systems are installed on vehicles, from the early advanced driver assistance system (ADAS) to the latest combined active front steering (AFS) and direct yaw moment control (DYC) system, which have been successfully deployed in intelligent vehicles to reduce traffic accidents caused by instability. When the vehicle is running at high speed or making a sharp turn, the rear axle is prone to reach the adhesion limit, and under the interference of lateral force, the vehicle body will slide sideways, which is an unstable or uncontrollable extreme condition, and the vehicle mass center side slip angle will change dramatically, and the traditional stability control system cannot play a role. Vehicle instability poses a significant threat to vehicle safety, resulting in a large number of instability accidents. In this case, the non-traditional turning method, i.e. drift, represents another possible balance condition for the vehicle when turning in an unstable manner, with the characteristics of high rear wheel side slip and reverse steering of the front wheels. This unstable but controllable turning method provides a control scheme for the vehicle body with large mass center side slip angle. By exploring the drift control strategy of professional drivers in the rally, the vehicle characteristics on the road can be more comprehensively understood, and further development of automatic driving technology with professional driving ability can be developed to expand the maneuvering stability domain range of the automatic driving vehicle.

[0003] Existing methods mainly focus on bicycle-vehicle models or two-axle vehicles with two-wheel drive, track circular trajectories for steady-state circular drift, and make overly simple assumptions that the vehicle body's center of mass side slip angle needs to be reduced as soon as possible at the end of the drift through the curve. Path following can affect vehicle handling stability due to tire saturation and large center of mass side slip angle. At the same time, few methods focus on autonomous drift control of three-axle commercial vehicles with distributed drive systems. Compared with centralized drive cars, distributed drive electric vehicles can more easily adjust the longitudinal torque of each wheel, have better grip and larger side slip angle, and can provide more drift possibilities. In the drift state, the distributed drive electric vehicle behaves as a nonlinear overdrive system, with more control inputs than control states. Autonomous driving of commercial vehicles has already landed in scenarios such as logistics transportation, mines, and ports, and has realized commercial operation. Limited by technology, funds, regulations, scenarios, safety, and many other factors, the commercialization of autonomous driving in passenger vehicles has been slow. By the end of 2021, SAIC, Baidu, Little Horse, Wen Yuan, Yuan Rong, and Didi have launched autonomous commercial vehicle research, and some technologies use three-axle vehicles as carriers. The above autonomous drift controllers cannot be used for three-axle distributed drive commercial vehicles because the redundancy and coupling of the overdrive system input vector hinder the direct use of a general drift controller. Three-axle commercial vehicles have more complex dynamic models and load transfer models. A two-stage variable-curvature assisted drift scheme is proposed to promote the scenario application of the assisted drift system. It has important theoretical significance and engineering value for autonomous driving technology that pursues efficiency and limits. SUMMARY

[0004] The purpose of the present application is to propose a two-stage assisted drift control method and a safety assistance system to deal with extreme working conditions that occur during the driving of three-axle commercial vehicles. To achieve the above purpose, the method mainly includes the following steps:

[0005] S1: A double-track three-axle vehicle path tracking model containing nonlinear tires is built according to the overall architecture of a three-axle commercial vehicle.

[0006] S2: Calculate the steady-state drift state variables according to the characteristics of the curve, analyze the stability of the state variable root trajectory, and track the expected steady-state drift equilibrium state based on the time-varying model prediction algorithm to achieve assisted drift control.

[0007] S3: Obtain the pre-look road information through relevant vehicle-mounted vision sensors, calculate the lateral error e and the heading error ΔΨ, and switch the control mode accordingly to achieve assisted drift through the curve in advance with a large center of mass side slip angle and stable heading out of the curve in advance.

[0008] S4: At the end of the curve, stop the assisted drift control, and at the same time, switch to the heading stability control integrated with active front steering / additional yaw moment to quickly restore the vehicle body attitude to be smooth.

[0009] S5: transmitting the control instructions obtained in S2, S3 and S4 to the wheel drive assembly and the steer-by-wire mechanism through a communication bus for real-time execution.

[0010] Preferably, a double-track three-axle vehicle path tracking model containing a nonlinear tire is built according to the whole frame architecture of a three-axle commercial vehicle. The vehicle is equipped with intelligent sensors, computing devices and execution mechanisms related to core technologies such as environment perception, path planning and tracking, and vehicle dynamics control. A three-axle vehicle double-track model is used, containing a magic nonlinear tire model, to further approximate the linear stiffness and consider the lateral and longitudinal load transfer. To reduce the calculation load of the control algorithm, the lateral force curve F yij (α ij ) is fitted using a simplified form of the magic formula.

[0011]

[0012] where μ is the road adhesion coefficient of the tire, F zij is the vertical load of the tire, B, C, D, E are Magic tire model parameters, and α ij is the tire side slip angle.

[0013] In the drift steady-state solving stage, a three-axle vehicle double-track model is used to further accurately express the vehicle state under large mass center side slip angle, with vehicle speed V, mass center side slip angle β, and yaw rate γ as the model state variables, and front wheel steering angle and middle and rear axle driving force as the control variables for vehicle body posture adjustment.

[0014]

[0015]

[0016]

[0017] where δ f represents the equivalent front wheel steering angle, F xi (i=f, m, r) are the equivalent longitudinal tire forces of the front, middle and rear axles, respectively, F yi (i=f, m, r) are the equivalent lateral tire forces of the front, middle and rear axles, respectively, I z represents the moment of inertia of the vehicle, m is the total mass of the vehicle, a, b, c represent the shortest distances from the front axle, middle axle and rear axle to the mass center G.G, respectively. i (i=f, m, r) represent the wheel track of the front axle, middle axle and rear axle, respectively.

[0018] A path tracking model is established including a geodetic coordinate system XOY, a vehicle body coordinate system xoy and a curvilinear coordinate system defined relative to a desired reference path position, to obtain a vehicle mass center position state equation and a path error kinematics equation

[0019]

[0020]

[0021]

[0022] wherein X, Y represent longitudinal and lateral coordinates of the vehicle mass center in the absolute coordinate system XOY. The curvilinear coordinate system describes the position of the vehicle relative to the reference path, and the lateral error e is the distance from the closest point on the path to the vehicle mass center. κ road is the road curvature, s is the distance traveled along the path, ψ is the vehicle heading angle, ψ ref is the road heading angle, and Δψ is the angle between the vehicle heading and the tangent to the path at the closest point.

[0023] Preferably, the steady-state drift state quantity is calculated according to the characteristics of the curve, and the stability analysis is performed by means of the root locus. The equilibrium must satisfy the steady-state equation, which is composed of the derivative terms of the dynamic equation being equal to zero.

[0024]

[0025]

[0026] wherein R road is the road curvature. The time-varying model prediction algorithm is used to track the expected steady-state drift equilibrium state, thereby realizing auxiliary drift control. According to the road curvature and the real-time vehicle speed, the vehicle equilibrium state quantity [v xeq v yeq γ eq ] and the equilibrium control quantity [δ feq F xmeq F xreq ] are solved, and the subscript eq represents the stable value of each variable equilibrium. A linearized differential dynamic model is used as the prediction model. At the steady-state drift equilibrium state, the state transition matrix A df_c and the control matrix B df_c are obtained by means of the Jacobian matrix according to the vehicle system dynamics formula.

[0027]

[0028] wherein x(t) = [v x -v xeq v y -v yeq γ-γ eqeΔψ] T u(t)=[δ f -δ feq F xm -F xmeq F xr -F xreq ] T ,

[0029]

[0030]

[0031] The dynamics equation is linearized, and the model is discretized and iterated, and a series of control increments in the control time domain can be obtained by using a nonlinear optimal solver to roll the optimization cost function. After obtaining the control increment sequence, the actual control input of the system can be calculated by the first element of the control sequence.

[0032] Preferably, by real-time previewing the road information, when the current road appears a curve, the road curvature is calculated, and the steady-state drift state quantity and control quantity are calculated in combination with the three-degree-of-freedom three-axle vehicle model and the current vehicle speed. According to the road characteristics, the mode switching of auxiliary drift control and heading stability control is performed to realize the auxiliary drift of large mass center side slip angle in advance of entering the curve and the heading stability in advance of exiting the curve.

[0033] Preferably, when the curve is approaching the end, the auxiliary drift control is stopped, and the heading stability control integrated with active front steering / additional yaw moment is switched. The vehicle mass center side slip angle is constrained and controlled in the stable range, while the path tracking accuracy is ensured. Based on the equivalent stiffness, a two-degree-of-freedom vehicle dynamics model is established,

[0034]

[0035]

[0036] Rewrite the state equation,

[0037]

[0038] y st (t)=x st (t)

[0039] wherein M Fx is an additional yaw moment depending on the wheel longitudinal force; is the equivalent axle tire stiffness of the front, middle and rear axles (k i =k il +k ir , i = f, m, r), the superscript indicates real-time updating, x st =[v yγ XY] T u st =[δ f M Fx ] T A st_c With B st_c Let h be the state transition matrix. st_c This is an additional matrix.

[0040] The model is discretized to derive the system iterative equations, define vehicle dynamics constraints and cost functions, and solve for the required front wheel steering angle and additional yaw moment. Using rule-based average and dynamic allocation strategies, encompassing actuator and road surface constraints, the optimization objective is to minimize the weighted sum of squares of the load rates of the four tires to improve vehicle stability. The extreme values ​​of the objective function are then solved using the elimination substitution method to obtain the driving torque of each drive wheel.

[0041] Preferably, the calculated control commands are transmitted to the drive-by-wire actuator via the CAN bus for vehicle attitude control.

[0042] This invention also proposes a two-stage drift safety assistance system for commercial vehicles to cope with extreme operating conditions. It includes an intelligent transportation system (V2X), an environmental perception system, a vehicle state planning unit, a vehicle state observation unit, a vehicle parameter estimation unit, an industrial control computer (IPC) central computing unit, and a vehicle drive-by-wire and steering unit.

[0043] The intelligent transportation system (V2X) mainly includes a roadside information acquisition and transmission module (RSU) and a vehicle-side information receiving module (OBU). The RSU is responsible for collecting road curvature and pre-aiming point location information of the covered road sections, obtaining road surface adhesion based on previous driving experience, and transmitting the dynamic road information to the central computing unit of the vehicle-side industrial control computer.

[0044] The environmental perception system includes lidar, millimeter-wave radar, visual cameras, and P2 integrated navigation, which realizes the perception and precise positioning of the vehicle's surrounding environment based on multi-sensor fusion technology.

[0045] The vehicle state planning unit calculates the desired vehicle driving state quantity based on the road information provided by the intelligent transportation system, and corrects the desired state quantity based on the vehicle's lateral and heading errors.

[0046] The vehicle status observation unit combines the vehicle status information collected by the vehicle-side ESP with the expected status quantity to calculate the error, and then transmits it to the industrial control computer central computing unit (IPC).

[0047] The vehicle parameter estimation unit estimates important parameters such as tire lateral stiffness and road surface adhesion based on vehicle status information and transmits them to the industrial control computer central computing unit (IPC).

[0048] The industrial personal computer central computing unit (IPC) combines the road information and the vehicle information, uses the S1-S5 control scheme to perform control amount calculation, and transmits the calculated control instruction to the drive-by-wire execution mechanism through the CAN bus to control the vehicle posture.

[0049] The beneficial effects of the present application are:

[0050] (1) The method realizes the fusion of curve drifting and straight path tracking, improves the deficiency of the existing method that can only perform stable circular condition drifting control, and meets the control requirements of unmanned commercial vehicle complex condition drifting driving.

[0051] (2) The method can realize the dynamic drifting control of three-axle commercial vehicles for complex objects, and enriches the application objects of drifting technology.

[0052] (3) The method provides a controllable scheme for safe driving outside the vehicle dynamics stability boundary, which is of great significance to improve the safe driving performance of three-axle commercial vehicles in extreme conditions, and provides a theoretical basis for the birth of a vehicle safety auxiliary drifting system for coping with extreme conditions. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 A two-stage auxiliary drifting path tracking control scheme architecture diagram;

[0054] Figure 2 A three-axle distributed drive commercial vehicle auxiliary drifting system configuration;

[0055] Figure 3 A two-stage auxiliary drifting TruckSim simulation scene;

[0056] Figure 4 A three-axle commercial vehicle auxiliary drifting path tracking bird's eye view;

[0057] Figure 5 A three-axle commercial vehicle auxiliary drifting process vehicle state phase diagram. DETAILED DESCRIPTION

[0058] The present application provides a two-stage drifting control method and a safety auxiliary system for coping with extreme conditions of a commercial vehicle. In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described below in combination with the accompanying drawings and embodiments. However, the protection scope of the present application is not limited thereto. The present application is based on Figure 1 The two-stage auxiliary drifting path tracking control scheme architecture diagram shown in the figure is implemented, wherein the auxiliary drifting system configuration of the distributed drive three-axle commercial vehicle is shown in Figure 2 .

[0059] Step 1: seeFigure 4 The drift state vehicle path tracking bird's-eye view of the large centroid side slip angle is shown, a, b, c represent the shortest distance from the front axle, the middle axle, and the rear axle to the centroid G. G, respectively. i (i = f, m, r) represent the wheel base of the front axle, the middle axle, and the rear axle, respectively. x y , β, Ψ, r represent the longitudinal velocity, the lateral velocity, the vehicle body centroid side slip angle, the vehicle speed heading angle, and the vehicle yaw rate, respectively. xij yij , α ij represent the lateral force, the longitudinal force, and the tire side slip angle of each wheel, respectively, and the subscript ij (i = f, m, r; j = l, r) represents the wheel serial number, where i = f, m, r represent the front, the middle, and the rear, respectively, and j = l, r represent the left and the right, respectively, and δ fj (j = l, r) represent the front left and the front right wheel steering angle. Further description of other state quantities of the model is as follows

[0060]

[0061]

[0062] F xi = F xil + F xir , F yi = F yil + F yir (i = f, m, r)

[0063] In the formula, V represents the vehicle speed, δ f represents the equivalent front wheel steering angle, F xi (i = f, m, r) are the equivalent longitudinal tire forces of the front axle, the middle axle, and the rear axle, respectively, and F yi (i = f, m, r) are the equivalent lateral tire forces of the front axle, the middle axle, and the rear axle, respectively.

[0064] The vehicle shown in the bird's-eye view includes the geodetic coordinate system XOY, the vehicle body coordinate system xoy, and the curvilinear coordinate system defined with respect to the position of the required reference path, and the vehicle centroid position can be obtained by the following formula

[0065]

[0066]

[0067] ​​where X, Y denote the longitudinal and lateral coordinates of the vehicle center of mass in the absolute coordinate system XOY. The curvilinear coordinate system describes the position of the vehicle relative to the reference path, and the lateral error e is the distance from the closest point on the path to the vehicle center of mass. s is the distance traveled along the path, and Δψ is the angle between the vehicle heading and the tangent to the path at the closest point. During the drift process, because the vehicle heading angle is no longer equal to the road heading angle, Δψ = ψ - ψ ref , resulting in Δψ ≠ 0. When the drift steady state is reached, the center of mass side slip angle and the heading error remain unchanged and are opposite in sign. The path error dynamics can be expressed as

[0068]

[0069]

[0070] The double-track model balances between model complexity and accuracy. In the drift steady state solving stage, a double-track model of a three-axle vehicle is adopted, and in order to further accurately express the vehicle state under a large center of mass side slip angle, the vehicle body posture is adjusted by taking V, β, r as the model state quantity and the front wheel steering angle and the middle and rear axle driving force as the control quantity, and the following dynamics model is established,

[0071]

[0072]

[0073]

[0074] where I z denotes the vehicle moment of inertia, and m is the total mass of the vehicle.

[0075] The control layer is convenient for driving torque optimization according to the tire vertical load, and further based on the static load of each axle, the vertical load transfer of each wheel is modeled according to the vehicle lateral and longitudinal acceleration, respectively.

[0076]

[0077]

[0078]

[0079] where F Zij (i = f, m, r, j = r, l) is the vertical load of each wheel, m s is the sprung mass of the whole vehicle, m ui (i = f, m, r) are the unsprung masses of the front, middle and rear axles, respectively, and satisfy m = m s+ m uf+ m um+ m ur h gis the height of the vehicle center of mass to the roll center, P f , P r , P m are the proportions of lateral weight transfer occurring on the front, middle, and rear axles, respectively, L is the wheelbase, a x , a y are the longitudinal and lateral accelerations, respectively, which can be calculated by

[0080]

[0081] To reduce the computational load of the control algorithm, its form is simplified and the tire linear region cornering stiffness is estimated,

[0082]

[0083]

[0084] where μ is the tire road adhesion coefficient, F zij is the tire vertical load, and B, C, D, E are Magic tire model parameters. Based on tire dynamics analysis, the tire cornering angle α ij (i = f, m, r, j = r, 1) can be expressed as

[0085]

[0086]

[0087] The single track model equivalent tire cornering angle is the average of the two side tire cornering angles

[0088]

[0089] From which the tire critical cornering angle α cr can be calculated, beyond which no more lateral force can be generated, and the linear tire stiffness is updated

[0090]

[0091]

[0092] To improve the robustness of the vehicle in extreme conditions, the tire lateral force is obtained with the linear tire stiffness as

[0093]

[0094] Step 2: With real-time preview road information, when the front road appears a curve, the road curvature is calculated, and the steady-state drift state and control are calculated by combining the three-degree-of-freedom three-axis vehicle model and the current vehicle speed. The steady-state equation is composed of the derivative term of the dynamic equation equal to zero, and the constraint is solved by the iter iterative optimization algorithm. According to the road curvature and real-time vehicle speed, the vehicle equilibrium state quantity [v xeq v yeq γ eq ] and the equilibrium control quantity [δ feq F xmeq F xreq ] are solved.

[0095]

[0096]

[0097] Therefore,

[0098]

[0099] The MPC-based controller is used to track the reference steady-state drift equilibrium state. A linearized differential dynamics model is used as the prediction model, and the path tracking error during driving is considered. The state space equation is derived as follows, where A df_c and B df_c are obtained by solving the Jacobian matrix according to the vehicle system dynamics formula at the steady-state drift equilibrium state.

[0100]

[0101] Where x=[v x -v xeq v y -v yeq γ-γ eq e Δψ] T u=[δ f -δ feq F xm -F xmeq F xr -F xreq ] T ,

[0102]

[0103] So far, the linearized dynamics equation is obtained, the model is discretized, and the system iteration equation is derived, as follows.

[0104] During the drift of the three-axis commercial vehicle, it is necessary to constrain the kinematics and dynamics of the vehicle. First, the control output F xi and δf the limit value δ fmax and the rate limit value Δδ fmax The following vehicle dynamics constraints need to be satisfied.

[0105]

[0106] Furthermore, the tire longitudinal and lateral forces should satisfy the friction ellipse constraints. The front wheels have no driving force, so only the lateral force due to steering is considered, and the middle and rear axles are saturated and thus exceed the adhesion limit when drifting.

[0107]

[0108] Since there is a coupling relationship between the control input of the control system and each state, it is necessary to add a weight coefficient to balance the optimization objective. To facilitate the realization of the drift state, expand the stability boundary, the rear wheel saturation should be encouraged to make the vehicle have the tendency of over-steering, with the tire side slip angle limit value cri An exponential cost function is set as the benchmark, which is expressed as follows:

[0109]

[0110] Therefore, the final MPC optimization objective function is:

[0111]

[0112] The controller sets the prediction step Np equal to the control step Nc, i.e. Np = Nc = 15. The first term y(k) of the cost function aims to achieve body stability drift and path tracking, the second term u(k) aims to smooth the control amount, and the last term g(k) promotes the saturation of the rear wheel, g(k) = [J(a m (k)), J(a r (k))] T The matrices Q df , R df , and W df are weight matrices. To make the weight adjustment more intuitive, the Q df , R df , and W df matrices (as shown below) are diagonal matrices, which respectively weigh the state and control amounts. Each diagonal value is the square of the inverse of the acceptable maximum deviation based on the balance value.

[0113]

[0114] By using a nonlinear optimal solver to roll the optimization cost function, a series of control increments in the control time domain can be obtained

[0115] U(k) = [u(k) u(k+1)...u(k+N p -1)] T

[0116] After obtaining the control increment sequence, the first element of the control sequence is selected to calculate the actual control input of the system.

[0117] u real (k) = u eq (k) + u(k)

[0118] Step 3: According to the road characteristics, the mode switching of auxiliary drift control and heading stability control is carried out to realize the auxiliary drift of large centroid side slip angle in advance of turning and the heading stability in advance of turning. See the flowchart shown in the table below. According to the preview curvature κ road_pre of the turning road section, the total length s c_pre of the preview turning road section, and (X o , Y o ) as the vehicle coordinate point, the advance turning coordinate point (X c_in , Y c_in ) is determined by the preview distance of the auxiliary drift controller, and the advance turning coordinate point (X c_out , Y c_out ) is calculated. Due to the existence of centroid side slip angle, the vehicle heading angle is aligned in advance to the turning direction.

[0119]

[0120]

[0121] Step 4: At the end of the turning, stop the auxiliary drift control and switch to the heading stability control integrated with active front steering / additional yaw moment. The vehicle centroid side slip angle is constrained and controlled within the stable range, while ensuring the path tracking accuracy. The tire stiffness is estimated online, and the tire side slip stiffness is solved in real time. The vehicle lateral acceleration and yaw rate acceleration are modeled as a two-degree-of-freedom equation to ensure the accuracy of the model.

[0122]

[0123]

[0124] where M Fx is the additional yaw moment dependent on the wheel longitudinal force; is the equivalent front, middle and rear axle tire stiffness (k i = k il + k ir, where the superscript denotes real-time update. Since there is a large vehicle body side slip angle and yaw rate at the beginning of the end of drift, which exceeds the vehicle stability boundary, in order to reduce the coupling between state variables, the global coordinates X and Y of the vehicle are selected as state variables for path tracking, and v y To constrain the vehicle body posture correction, the continuous-time state equation is established, and the model discretization and system iteration equation are attached below,

[0125]

[0126] y st (t)=x st (t)

[0127] where x st =[v y γ X Y] T u st =[δ f M Fx ] T .

[0128]

[0129] respectively represent the equivalent tire stiffness of the front, middle and rear axles.

[0130] The limit values and the rate of change of the control outputs δ f and M Fx need to meet the following vehicle dynamics constraints: the front wheel steering angle limit is δ fmax =(25π / 180) rad, and the additional yaw moment constraint is M Fxmax =(0.2I z ) Nm, u max =[δ fmax M Fx_max ] T

[0131] u(k)≤[u max (k)-u max (k)] T

[0132] When the drift ends, reducing the center of mass side slip angle to restore the vehicle body posture is the primary task, so the lateral velocity expectation v ydes is set to 0. To ensure stability during posture adjustment and path tracking, according to the vehicle front wheel steering angle, the steady-state yaw rate γ ss ,

[0133]

[0134] where

[0135]

[0136]

[0137]

[0138] is an intermediate variable.

[0139] yaw rate γ tire is also limited by the road adhesion condition

[0140]

[0141] where μ ij is the friction coefficient of each wheel; F zij represents the vertical load of each wheel. Thus, the required yaw rate γ des can be expressed as

[0142] γ des = min{|γ ss |,|γ tire |}sgn(δ f )

[0143] path tracking collects the coordinates X, Y in front of the road through advanced perception preview technology, and assigns them to X des and Y des , respectively.

[0144] At the same time, the output of the controller is minimized, considering the constraints of the system on the objective function, and the prediction step of this controller is set to be equal to the prediction compensation, i.e. Np=Nc=10. The MPC cost function is as follows:

[0145]

[0146] where y des = [v ydes γ des X des Y des ], des represents the expected value of the relevant state quantity. Q st , R st are weight matrices, and the relevant components are as follows

[0147]

[0148] By solving the optimization problem in the formula, an optimal control sequence, i.e. U(k), can be obtained, and the first control vector is selected as the feedback u(k), i.e.

[0149]

[0150] The optimization distribution can better adapt to the dynamic performance of vehicle load transfer and road adhesion conditions, and to a certain extent, serve the upper additional yaw moment control, realize smaller centroid side slip angle and yaw angular velocity following deviation at the end of drift, and quickly restore the body posture. The actuator constraints and road constraints are included, and the weighted square sum of four tire load rates is minimized as the optimization objective, which improves the vehicle stability, and the extreme value of the objective function is solved by the elimination substitution method.

[0151] minJ = u T Φu + (χu - u ref ) T Θ(χu - u ref )

[0152] s.t.u min ≤u≤u max

[0153] In the formula: u = (T ml ,T mr ,T rl ,T rr ) T , is the vector to be optimized, T i is the torque command of the i-th tire, i = ml, mr, rl, rr, which are the middle left, middle right, rear left and rear right respectively; u min , u max are vectors composed of the upper and lower limits of the motor torque respectively; u ref = [T Fx , M Fx ], is a virtual control reference vector; T FX is the longitudinal total driving torque;

[0154]

[0155] a xdes is the target longitudinal acceleration, and the gain K vx = 8000 Ns / m is sufficient to realize speed tracking. Φ, Θ are diagonal weight matrices; χ is the efficiency matrix. Φ, Θ, χ can be expressed as follows

[0156]

[0157] Where, f m is the middle axle track, and f r is the rear axle track.

[0158] The discretization and iterative system equations of the MPC control model of the two modes described in steps 3 and 4 are described here. The formula method is used to obtain the discrete model of dynamics,

[0159]

[0160] wherein,

[0161]

[0162] wherein T s is the sampling period, ζ=(df or st), df represents the drift state, and st represents the steady-state path tracking. If ζ=df, h df_c =O.

[0163] Based on the above discrete model, the future output of the system is predicted by iterating the state equation

[0164]

[0165] (N=N, 2,...,Np)

[0166]

[0167] Where,

[0168]

[0169]

[0170]

[0171] Subsequently, the controller recalculates a new control sequence according to the current state of the vehicle, and iterates and updates continuously to achieve the rolling optimization of the control command. Similarly, by predicting the output of the next period according to the state information, the remaining input control quantity of the system can be generated in turn.

[0172] Step 5, the calculated front wheel steering angle and rear axle drive force control command are transmitted to the drive-by-wire actuator through CAN bus for vehicle attitude control, as shown in Figure 3 The two-stage auxiliary drift TruckSim simulation scene is the intuitive effect of the control system. From Figure 5 The vehicle state phase diagram in the auxiliary drift process of the three-axle commercial vehicle can be seen that the scheme expands the steady-state control range of the vehicle, and provides a controllable scheme for the safe driving of the three-axle commercial vehicle outside the stability boundary.

[0173] The above series of detailed descriptions are only specific descriptions of the feasible embodiments of the present application, and are not intended to limit the protection scope of the present application. Any equivalent means or changes made without departing from the technology of the present application shall be included in the protection scope of the present application.

Claims

1. A three-axle commercial vehicle two-stage drift control method for coping with extreme working conditions, characterized in that, The method comprises the following steps: S1: building a double-track three-axle vehicle path tracking model containing a nonlinear tire; S2: calculating a steady-state drift state quantity according to a curve feature, performing stability analysis on a state quantity root locus, tracking an expected steady-state drift balance state based on a time-varying model prediction algorithm, and realizing auxiliary drift control; S3: acquiring pre-look road information through a vehicle-mounted visual sensor, calculating a lateral error e and a heading error ΔΨ, switching control modes according to the two errors, realizing large-moment-center side slip angle auxiliary drift early into a curve, and heading stability early out of the curve; S4: stopping auxiliary drift control when a curve is close to an end, and simultaneously switching to heading stability control integrated with active front steering / additional yaw moment, so as to rapidly restore a vehicle body posture to be smooth. The implementation of S1 comprises: modeling vertical load transfer of each wheel according to vehicle lateral and longitudinal accelerations; where m s is the sprung mass of the vehicle, m ui (i = f, m, r) are the unsprung masses of the front, middle, and rear axles, h g is the height of the vehicle's center of mass to the roll center, P f , P r , P m are the proportions of lateral weight transfer that occur on the front, middle, and rear axles, respectively, L is the wheelbase, F Zij are the vertical loads on each wheel, i = f, m, r represent the front, middle, and rear, respectively, j = 1, r represent the left and right, respectively, a, b, c are the shortest distances from the front, middle, and rear axles to the center of mass, a x , a y are the longitudinal and lateral accelerations, respectively, which can be calculated by the following equations where v x , v y denote longitudinal and lateral velocity, respectively, and γ denotes the yaw rate. In order to reduce a control algorithm calculation load, a form is simplified, and tire linear zone side slip stiffness is estimated, where μ is the road adhesion coefficient of the tire, F zij is the vertical load of the tire, B, C, D, E are Magic tire model parameters, F yij (α ij ) represents the tire lateral force, k αij represents the tire linear region cornering stiffness, based on tire dynamics analysis, the tire cornering angle α ij may be represented as wherein α fl represents the side slip angle of the front left wheel, α fr represents the side slip angle of the front right wheel, α ml,r represents the side slip angle of the middle left and right wheels, α rl,r represents the side slip angle of the rear left and right wheels; t i , i = f, m, r, respectively represent the wheel base of the front axle, the middle axle, and the rear axle; δ fl represents the equivalent left front wheel steering angle, δ fr represents the equivalent right front wheel steering angle; A single-track model equivalent tire side slip angle is an average value of two-side wheel side slip angles wherein α i , i = f, m, r, represent the single-track model equivalent tire side slip angle of the front, middle, and rear wheel pairs, respectively, and α il , α ir , i = f, m, r, represent the side slip angles of the left and right wheels of the front, middle, and rear, respectively. From this the critical cornering angle a of the tire can be calculated cr , beyond which no more lateral force can be generated, while updating the linear tire stiffness In order to improve vehicle robustness in an extreme working condition, tire lateral force is obtained by means of linear tire stiffness Where μ is the tire's road adhesion coefficient, F zij B, C, D, and E represent the vertical load on the tire, while α represents the parameters of the Magic tire model. ij The tire slip angle, α crij In the α symbol, the subscripts i = f, m, r represent the front, middle, and back respectively, and j = l, r represent the left and right respectively. crij This indicates the critical sideslip angles of the left and right wheels at the front, middle, and rear, respectively. In this context, the subscripts i = f, m, r represent the front, middle, and back positions respectively, j = l, r represent the left and right positions respectively, and the superscript max represents the maximum value. This indicates the maximum lateral force of the left and right wheels at the front, middle, and rear.

2. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, The implementation of S1 further comprises: in a drift steady-state solving stage, adopting a three-axle vehicle double-track model, taking vehicle speed V, moment-center side slip angle β, and yaw rate r as model state quantities, taking front wheel steering angle and middle-rear axle driving force as control quantities to adjust the vehicle body posture, and establishing the following dynamic model where δ f denotes the equivalent front wheel steering angle, F xi F (i = f, m, r) are the front, middle, and rear axle equivalent longitudinal tire forces, respectively, F yi F (i = f, m, r) are the front, middle, and rear axle equivalent lateral tire forces, respectively, I z denotes the vehicle moment of inertia, m is the total vehicle mass, a, b, c denote the shortest distance from the front, middle, and rear axles to the center of mass, respectively, t i denote the wheel base of the front, middle, and rear axles, respectively.

3. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, The implementation of S1 further comprises: establishing a path tracking model containing a ground coordinate system XOY, a vehicle body coordinate system xoy, and a curve coordinate system defined relative to a required reference path position, and obtaining a vehicle mass center position state equation and a path error kinematics equation where ψ denotes the vehicle speed heading angle, κ road is the road curvature, X, Y denote the longitudinal and lateral coordinates of the vehicle center of mass in the absolute coordinate system XOY, the curvilinear coordinate system describes the position of the vehicle relative to the reference path, the lateral error e is the distance from the closest point on the path to the vehicle center of mass, s is the distance traveled along the path, and Δψ is the angle between the vehicle heading and the tangent to the path at the closest point.

4. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, S2 performs stability analysis on a state quantity root locus, and a balance must satisfy a steady-state equation composed of derivative terms of the dynamic equation being equal to zero: Kappa road R is the road curvature road R is the road curvature radius.

5. The two-stage drift control method for three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, S2 tracks an expected steady-state drift balance state based on a time-varying model prediction algorithm, and realizes auxiliary drift control. According to the road curvature and real-time vehicle speed, the equilibrium state variables [v xeq v yeq γ eq ] and equilibrium control variables [δ feq F xmeq F xreq ] are solved. A linearized differential dynamics model is used as a prediction model.

6. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, The implementation of S3 comprises: through real-time pre-look road information, when a current front road appears a curve, calculating a road curvature, combining a three-degree-of-freedom three-axle vehicle model with a current vehicle speed to calculate a steady-state drift state quantity and a control quantity, switching a mode of auxiliary drift control and heading stability control according to a road feature, realizing large-moment-center side slip angle auxiliary drift early into a curve, and heading stability early out of the curve.

7. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, The implementation of S4 comprises: Taking a vehicle moment-center side slip angle as a constraint, controlling the moment-center side slip angle in a stable range, guaranteeing path tracking precision, establishing a two-degree-of-freedom vehicle dynamic model based on equivalent stiffness; Discretizing the model, obtaining system iteration equations, formulating vehicle dynamic constraints and a cost function, solving required front wheel steering angle and additional yaw moment, adopting a rule-based average distribution and dynamic distribution strategy, including actuator constraints and road surface constraints, taking a weighted square sum of four tire load rates as an optimization objective, improving vehicle stability, and solving an extreme value of the objective function by a method of elimination and substitution, to obtain driving torques of each driving wheel.

8. The two-stage drift control method for a three-axle commercial vehicle in extreme working conditions according to claim 1, characterized in that, Also including S5: through the communication bus to S2, S3, S4 resulting control instruction transmission to the wheel drive assembly and steer-by-wire mechanism for real-time execution.

9. A two-stage drift safety assist system for a commercial vehicle operating in extreme conditions, characterized in that, Including intelligent transportation system (V2X), environmental perception system, vehicle state planning unit, vehicle state observation unit, vehicle parameter estimation unit, industrial computer central computing unit (IPC), vehicle steer-by-wire drive and steering unit; The intelligent transportation system (V2X) includes a roadside road information collection and transmission module (RSU) and a vehicle end information receiving module (OBU), which is responsible for collecting the road curvature and pre-aim point position information of the covered road section, and obtaining the road adhesion condition according to the previous driving experience, and transmitting the road dynamic information to the vehicle end industrial computer central computing unit; The environmental perception system includes laser radar, millimeter wave radar, visual camera and P2 combined navigation, and realizes the vehicle surrounding environment information perception and accurate positioning based on multi-sensor fusion technology; The vehicle state planning unit calculates the expected vehicle driving state quantity according to the road information provided by the intelligent transportation system, and corrects the expected state quantity according to the vehicle driving lateral error and heading error; The vehicle state observation unit combines the vehicle state information collected by the vehicle end ESP with the expected state quantity to obtain the error, and transmits it to the industrial computer central computing unit (IPC); The vehicle parameter estimation unit estimates the important parameters of tire cornering stiffness and road adhesion condition according to the vehicle state information, and transmits it to the industrial computer central computing unit (IPC); The industrial computer central computing unit (IPC) combines the above road information and vehicle information, uses the control method of claim 1 to calculate the control quantity, and transmits the calculated control instruction to the steer-by-wire execution mechanism through the CAN bus for vehicle attitude control.

Images