Chassis multi-system cooperative game control method for path tracking and stability

Abstract

The application discloses a chassis multi-system cooperative game control method for path tracking and stability, and belongs to the technical field of vehicle dynamics control, wherein the chassis multi-system comprises a DYC system, an AFS system, an ARS system and an ASS system, and the method comprises the following steps: constructing a vehicle dynamics and path tracking coordinated control model and establishing a vehicle state prediction equation; taking the DYC system, the AFS system, the ARS system and the ASS system as participants of cooperative game, and constructing a total cost function, wherein the control weight of each participant is dynamically adjusted according to the real-time vehicle state; and adopting an ADMM algorithm to distributively solve the total cost function, so as to obtain the optimal control input of the DYC system, the AFS system, the ARS system and the ASS system. The application realizes the optimal balance between the path tracking accuracy and the driving stability while ensuring the calculation efficiency, so as to realize the high-performance control of the intelligent vehicle under the extreme working condition.

Description

Technical Field

[0001] This invention belongs to the field of vehicle dynamics control technology, specifically relating to a chassis multi-system cooperative game control method oriented towards path tracking and stability. Background Technology

[0002] With the rapid development of intelligent and connected vehicles, autonomous driving and advanced driver assistance systems (ADAS) place extremely high demands on vehicle path tracking capabilities and driving stability. Path tracking control aims to ensure the vehicle accurately follows the desired reference trajectory, while stability control aims to prevent instability under extreme conditions such as high speeds, high curvature, or low-friction surfaces. However, there is an inherent contradiction between these two: excessive pursuit of path tracking accuracy may lead to overly aggressive steering, sacrificing lateral stability and even causing dangerous situations such as skidding and fishtailing; conversely, overly conservative prioritization of stability can reduce the accuracy and timeliness of tracking, affecting ride comfort and system performance. Therefore, achieving synergistic optimization between path tracking accuracy and vehicle dynamic stability in complex and ever-changing driving environments has become a key technical challenge in the field of intelligent vehicle control.

[0003] Currently, traditional control methods such as PID control, linear quadratic regulators (LQR), and model predictive control (MPC) are widely used in this field. These methods largely rely on accurate vehicle dynamics models. However, a vehicle is a highly nonlinear, time-varying, and parameter-deterministic complex system, especially under conditions approaching physical limits, where model accuracy is difficult to guarantee, leading to degraded control performance. Furthermore, the aforementioned methods typically treat path tracking and stability as two independent or simply weighted problems, lacking an effective mechanism to dynamically and intelligently coordinate their conflict. Against this backdrop, corner module vehicle control provides entirely new actuator degrees of freedom. By independently and precisely distributing the longitudinal / lateral forces and suspension forces of the four wheels, it lays the physical foundation for achieving higher-level integrated control performance. However, current control strategies have high computational complexity, making it difficult to meet the real-time requirements of the system. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies by providing a chassis multi-system collaborative game control method oriented towards path tracking and stability. It treats direct yaw moment control, active front wheel steering, active rear wheel steering, and active suspension system as four independent participants, and achieves parallel optimization and collaborative decision-making of each subsystem through the alternating direction multiplier method. In addition, this invention combines a state weight adaptive mechanism, which can dynamically adjust the control weight according to the real-time vehicle state, achieving an optimal balance between path tracking accuracy and driving stability while ensuring computational efficiency, thereby realizing high-performance control of intelligent vehicles under extreme conditions.

[0005] This invention provides the following technical solution:

[0006] A chassis multi-system cooperative game control method for path tracking and stability, wherein the chassis multi-systems include: Direct Yaw Moment Control (DYC) system, Active Front Steering (AFS) system, Active Rear Steering (ARS) system, and Active Suspension System (ASS) system, and the control method includes:

[0007] Step S1: Combine the vehicle dynamics model and the path tracking model to construct a coordinated control model for vehicle dynamics and path tracking;

[0008] Step S2: Based on the discretized vehicle dynamics and path tracking coordinated control model, establish the vehicle state prediction equation through the model predictive control algorithm;

[0009] Step S3: Based on the vehicle state prediction equation and the defined vehicle desired state, with the goal of balancing path tracking accuracy and driving stability, the DYC system, AFS system, ARS system and ASS system are used as participants in a cooperative game to construct the total cost function. When constructing the total cost function, the control weights of each participant are dynamically adjusted according to the real-time vehicle state.

[0010] Step S4: For the total cost function, the ADMM algorithm is used for distributed solution to obtain the optimal control inputs for the DYC system, AFS system, ARS system and ASS system.

[0011] Optionally, the vehicle dynamics and path tracking coordinated control model in step S1 is:

[0012] ;

[0013] in, for The first derivative, For state variables, , The sideslip angle is the angle of the centroid. The yaw rate is angular velocity. The body roll angle, The angular velocity is the roll rate. The vehicle's lateral position. For heading angle error, superscript Indicates transpose. , , and The control inputs are for the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. , , , , To add yaw moment, and These are the front and rear wheel steering angles of the car, This is the tilting moment; ;

[0014] ;

[0015] ;

[0016] ;

[0017] ;

[0018] ;

[0019] in, and These are the front and rear axle lateral stiffness, respectively. For the overall vehicle quality, For longitudinal velocity, and These are the distances from the center of mass to the front and rear axles, respectively. For the whole vehicle to be around Moment of inertia of the shaft For vehicle sprung mass Moment of inertia of the shaft For the sprung mass of the vehicle, Let be the tilt arm, denoted as the distance from the center of mass to the tilt center. It is the acceleration due to gravity. For axial roll stiffness, For roll damping, Indicates the curvature of the reference path.

[0020] Optionally, the vehicle state prediction equation in step S2 is:

[0021] ;

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] in, To the current sampling time The vehicle state prediction sequence below, Current sampling time The state sequence, To the current sampling time Predicted vehicle in future moments state, To predict the step size, The state matrix of the prediction equation, , It is the identity matrix. For discrete time, For participants in the prediction equation The control input matrix, , To the current sampling time Participants The control input sequence vector, Current sampling time The future The control input applied at all times, For the future The sequence of external disturbances for each prediction step affects the state prediction. Cumulative effects Current sampling time Next time Constant external disturbances.

[0028] Optionally, the desired vehicle state defined in step S3 includes: desired yaw rate. Desired vehicle roll angle And the expected centroid side slip angle ;

[0029] The desired yaw rate ,in, The road surface adhesion coefficient, It is the acceleration due to gravity. For longitudinal velocity, Wheelbase , For vehicle stability factors, , and These are the distances from the center of mass to the front and rear axles, respectively. For the overall vehicle quality, and These are the front and rear axle lateral stiffness, respectively. For the steering angle of the car's front wheels, It is a symbolic function;

[0030] The desired vehicle roll angle The desired centroid sideslip angle .

[0031] Optionally, step S3 specifically includes:

[0032] Step S31: Construct the cost function for each participant, specifically expressed as:

[0033] ;

[0034] in, To the current sampling time At that time, participants The cost function, To predict the step size, This is the sequence number of the prediction step. For participants The expected state matrix, To the current sampling time Predicted future moments state, , For the desired yaw rate, To achieve the desired vehicle roll angle, The desired centroid sideslip angle is indicated by the superscript. Indicates transpose. For participants State weight coefficients, For participants At the current sampling time The future moment The applied control input, For participants The weighting coefficients of the input quantities It is an L2 norm;

[0035] Step S32: Construct the total cost function based on the dynamically adjusted control weights of each participant. for:

[0036] ;

[0037] in, , , , The control weights for the four participants—DYC system, AFS system, ARS system, and ASS system—are respectively. , , and At the current sampling time The cost functions of the four participants: DYC system, AFS system, ARS system, and ASS system.

[0038] Optionally, in step S3, the step of dynamically adjusting the control weights of each participant based on the real-time vehicle status specifically involves:

[0039] Based on the normalized centroid sideslip angle deviation, yaw rate deviation, roll angle deviation, lateral displacement deviation, and heading angle deviation, a fusion weight vector based on multi-state errors is defined.

[0040] ;

[0041] in, For the normalized centroid sideslip angle deviation, For the normalized yaw rate deviation, For normalized vehicle roll angle deviation, This is the normalized lateral positional deviation. For normalized heading angle deviation, For weight normalization factor, , To prevent positive numbers from being set with a denominator of zero, The fusion weight for the centroid sideslip angle deviation, The fusion weight for yaw rate deviation, The fusion weights for vehicle roll angle deviation, The fusion weights are for the lateral positional deviation. The fusion weight for the heading angle deviation;

[0042] Calculate the base weight of the controller for each participant;

[0043] ;

[0044] in, For participants Corresponding to the controller's base weights, , , , and Participants of , , , and The corresponding weights;

[0045] The basic weights of the controllers corresponding to each participant are adaptively adjusted based on vehicle speed;

[0046] ;

[0047] ;

[0048] in, For longitudinal velocity, This represents the maximum longitudinal velocity. , , and These are the base weights of the controllers for the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. , , and The weights of the four participants—DYC system, AFS system, ARS system, and ASS system—are adjusted for vehicle speed.

[0049] The weights of each participant after speed adjustment are normalized, and first-order inertial smoothing is added to obtain the weights of each participant at the current sampling time. Control weight ;

[0050] ;

[0051] in, The weights of each participant after speed adjustment and normalization. This is the smoothing coefficient.

[0052] Optionally, step S4 specifically involves: reconstructing the total cost function into an augmented Lagrangian function and solving it using the ADMM algorithm to obtain the optimal control inputs for the DYC system, AFS system, ARS system, and ASS system, including: the additional yaw moment of the DYC system, the front wheel steering angle of the AFS system, the rear wheel steering angle of the ARS system, and the roll moment of the ASS system.

[0053] Optionally, the following steps are also included:

[0054] Step S5: Based on the additional yaw moment obtained from the solution, distribute longitudinal force to each wheel of the vehicle, and based on the roll moment obtained from the solution, distribute active suspension force to each wheel of the vehicle.

[0055] Compared with the prior art, the beneficial effects of the present invention are:

[0056] This invention significantly reduces computational complexity and greatly improves real-time performance by decomposing a complex centralized optimization problem into four parallel solvable local subproblems: a DYC system, an AFS system, an ARS system, and an ASS system. The distributed architecture ensures independent operation of each chassis system, preventing individual failures from affecting overall control and enhancing vehicle chassis reliability. Online adaptive adjustment of weight coefficients dynamically optimizes participant control weights, achieving a dynamic balance between stability, path tracking, handling, and comfort. Each system only needs to exchange control strategy information, resulting in low communication requirements and reduced demands on the in-vehicle network. The design based on cooperative game theory achieves Pareto optimality for overall vehicle performance, effectively resolving the conflict between path tracking accuracy and driving stability. Furthermore, this invention has low hardware requirements, can be implemented in existing distributed controller architectures, and possesses good economic viability and promotional value, providing an efficient solution for integrated control of intelligent vehicle chassis. Attached Figure Description

[0057] Figure 1 This is a structural block diagram of the chassis multi-system cooperative game control method for path tracking and stability according to the present invention. Detailed Implementation

[0058] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be used to limit the scope of protection of the present invention. It should be noted that the term "comprising" and any variations thereof in the specification, claims and the above-mentioned drawings of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or devices.

[0059] This embodiment provides a chassis multi-system cooperative game control method oriented towards path tracking and stability. The chassis multi-system includes: Direct Yaw Moment Control (DYC) system, Active Front Steering (AFS) system, Active Rear Steering (ARS) system, and Active Suspension (ASS) system. The structures of the DYC system, AFS system, ARS system, and ASS system refer to the prior art.

[0060] like Figure 1 As shown, the chassis multi-system cooperative game control method for path tracking and stability proposed in this application includes the following steps:

[0061] Step S1: Combine the vehicle dynamics model and the path tracking model to construct a coordinated control model for vehicle dynamics and path tracking.

[0062] In this embodiment, step S1 specifically includes steps S11-S13.

[0063] Step S11: Considering the effects of additional yaw moment, front wheel steering angle, rear wheel steering angle, and roll moment, build a vehicle dynamics model, specifically as follows:

[0064] (1);

[0065] In the formula: For the overall vehicle weight; For vehicle sprung mass; and This is the distance from the center of mass to the front and rear axles; The lateral lever arm is represented by the distance from the center of mass to the center of tilt. For vehicle sprung mass Moment of inertia of the shaft; For the whole vehicle to be around Moment of inertia of the shaft; and These are the front and rear axle lateral stiffness, respectively. This refers to the lateral force acting on the front axle. This refers to the lateral force acting on the rear axle. For axial tilt stiffness; For roll damping; This refers to the vehicle body roll angle; This refers to the yaw rate; Longitudinal velocity; Lateral velocity; It is the centroid sideslip angle; and These are the steering angles of the front and rear wheels of the car, respectively. and The slip angle of the front and rear tires; This refers to the vehicle's lateral acceleration. Differentiate the lateral velocity; This is the yaw acceleration; The angular velocity of the center of mass deflection; This refers to the roll angular velocity; This refers to the roll acceleration; To add front wheel steering angle; To add yaw moment; This is the tilting moment.

[0066] Considering that the vehicle roll angle is generally small during vehicle movement, to simplify the calculation, let:

[0067] (2);

[0068] (3);

[0069] Therefore, equation (1) can be rewritten as:

[0070] (4);

[0071] Step S12: Use lateral tracking error and heading angle error to characterize the effectiveness of path tracking and establish a vehicle path tracking model.

[0072] (5);

[0073] In the formula, This refers to the heading angle error; This is the actual heading angle; This is the actual heading angular velocity; The desired heading angle; Indicates the curvature of the reference path. This is the first derivative of the heading angle error.

[0074] vehicle lateral position The first derivative is:

[0075] (6);

[0076] Considering the path tracking process, heading angle error Generally, the size is small, so equation (6) can be simplified to:

[0077] (7);

[0078] Step S13: Combining the vehicle dynamics model and the path tracking model, i.e., formulas (4), (6), and (7), establish a coordinated control model for vehicle dynamics and path tracking expressed in the form of state-space equations:

[0079] (8);

[0080] in, for The first derivative, For state variables, superscript Indicates transpose. , , and The control inputs are for the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. , , , ; ; ;

[0081] ; ; ; .

[0082] Step S2: Based on the discretized vehicle dynamics and path tracking coordinated control model, establish the vehicle state prediction equation through the model predictive control algorithm.

[0083] In this embodiment, step S2 specifically includes steps S21 and S22.

[0084] Step S21: Discretize the vehicle dynamics and path tracking coordinated control model.

[0085] Equation (8) is discretized using an approximate discretization method:

[0086] (9);

[0087] in, , , , , ; For discretization time, It is the identity matrix; , , and These are the control input matrices for the four participants.

[0088] Step S22: Establish vehicle state prediction equations.

[0089] Set the current sampling time The prediction step size is Its vehicle state prediction equation is:

[0090] (10);

[0091] The vehicle state prediction equation, i.e., formula (10), is rearranged as follows:

[0092] (11);

[0093] in, ; ; ; ; , To the current sampling time The vehicle state prediction sequence below, Current sampling time The state sequence, To the current sampling time Predicted vehicle in future moments state, To predict the step size, The state matrix of the prediction equation, For participants in the prediction equation The control input matrix, To the current sampling time Participants The control input sequence vector, Current sampling time The future The control input applied at all times, For the future The sequence of external disturbances for each prediction step affects the state prediction. Cumulative effects Current sampling time Next time Constant external disturbances.

[0094] Step S3: Based on the vehicle state prediction equation and the defined desired vehicle state, and with the goal of balancing path tracking accuracy and driving stability, the DYC system, AFS system, ARS system, and ASS system are treated as participants in a cooperative game to construct the total cost function. During the construction of the total cost function, the control weights of each participant are dynamically adjusted according to the real-time vehicle state.

[0095] In this embodiment, the defined desired vehicle state includes: desired yaw rate. Desired vehicle roll angle And the expected centroid side slip angle .

[0096] First, the ideal yaw rate of the vehicle is calculated based on a linear two-degree-of-freedom model. and centroid side slip angle :

[0097] (12);

[0098] (13);

[0099] in, Wheelbase , For vehicle stability factors, .

[0100] Meanwhile, considering road surface adhesion limitations, the desired yaw rate is constrained:

[0101] (14);

[0102] in, This is the road surface adhesion coefficient.

[0103] Based on formulas (12) and (14), the desired yaw rate under normal operating conditions is obtained. for:

[0104] (15);

[0105] Desired vehicle roll angle:

[0106] (16);

[0107] Of course, the defined desired vehicle state also includes the desired heading angle and desired lateral position. The desired heading angle is the tangent angle of the desired path at the current position point, which is usually calculated based on preset global path points or trajectory points from local planning. The desired heading angle and desired lateral position have already been integrated into the vehicle dynamics and path tracking coordinated control model in step S1, therefore, they will not be defined separately here.

[0108] Step S3 specifically includes steps S31 and S32.

[0109] Step S31: Construct the cost function for each participant.

[0110] (17);

[0111] (18);

[0112] (19);

[0113] (20);

[0114] That is, according to formulas (17)-(20), the cost function of each participant can be uniformly expressed as:

[0115] (twenty one);

[0116] in, To the current sampling time At that time, participants The cost function, To predict the step size, For participants The expected state matrix, , , , , To the current sampling time Predicted future moments state, This is the sequence number of the prediction step. , For the desired yaw rate, To achieve the desired vehicle roll angle, To determine the desired centroid sideslip angle, For participants State weight coefficients, , , , , For participant 1, the weighting coefficient for the centroid side slip angle. The yaw rate weighting coefficient for participant 1. The lateral positional deviation weighting coefficient for participant 2. For participant 2, the heading angle deviation weighting coefficient, The weighting coefficient for the centroid sideslip angle of the participants. The yaw rate weighting coefficient for participants is 3. For the participants' 4-sided tilt angle weighting coefficients, For participants At the current sampling time The future moment The applied control input, For participants The weighting coefficients of the input quantities It is an L2 norm.

[0117] Step S32: Construct the total cost function based on the dynamically adjusted control weights of each participant. for: (twenty two);

[0118] in, , , , The control weights for the four participants—DYC system, AFS system, ARS system, and ASS system—are respectively. , , and At the current sampling time At that time, the cost functions of the four participants—DYC system, AFS system, ARS system, and ASS system—are... , , .

[0119] In this embodiment, the control weights of each participant are dynamically adjusted based on the real-time vehicle status, specifically as follows:

[0120] First, the deviations of the center of gravity sideslip angle, yaw rate, roll angle, lateral displacement, and heading angle are normalized:

[0121] (twenty three);

[0122] In the formula, For the normalized centroid sideslip angle deviation, For the normalized yaw rate deviation, For normalized vehicle roll angle deviation, This is the normalized lateral positional deviation. Normalized heading angle deviation; This represents the maximum value of the centroid sideslip angle; This represents the maximum yaw rate. This represents the maximum roll angle of the vehicle. This represents the maximum value of the horizontal position. This represents the maximum heading angle. The desired lateral position of the vehicle;

[0123] Define a fusion weight vector based on multi-state errors:

[0124] (twenty four);

[0125] in, For the normalized centroid sideslip angle deviation, For the normalized yaw rate deviation, For normalized vehicle roll angle deviation, This is the normalized lateral positional deviation. For normalized heading angle deviation, The fusion weight for the centroid sideslip angle deviation, The fusion weight for yaw rate deviation, The fusion weights for vehicle roll angle deviation, The fusion weights are for the lateral positional deviation. The fusion weight for the heading angle deviation; This is the weight normalization factor.

[0126] (25);

[0127] in, A positive number is set to prevent the denominator from being zero.

[0128] Establish the correlation strength matrix between the controller and the state error for each participant:

[0129] (26);

[0130] In the formula, The relevant weighting coefficient for participant 1; The relevant weighting coefficients for participant 2; The relevant weighting coefficients for participant 3; The relevant weighting coefficients for participant 4.

[0131] Calculate the base weight of the controller for each participant;

[0132] (27);

[0133] in, For participants Corresponding to the controller's base weights, , , , and Participants of , , , and The corresponding weights. The ADMM constraint handling mechanism can strictly guarantee the physical constraints of the actuators corresponding to each system, improving engineering practicality.

[0134] The weights are automatically adjusted based on vehicle speed:

[0135] (28);

[0136] (29);

[0137] in, For longitudinal velocity, This represents the maximum longitudinal velocity. , , and These are the base weights of the controllers for the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. , , and The weights of the four participants—DYC system, AFS system, ARS system, and ASS system—are adjusted for vehicle speed. The weight of DYC system increases at high speeds; the weight of AFS system decreases at high speeds; the weight of ARS system decreases at high speeds; and the weight of ASS system increases at high speeds.

[0138] Normalize the weights:

[0139] (30);

[0140] Add first-order inertial smoothing to avoid abrupt weight changes:

[0141] (31);

[0142] In the formula This is a smoothing coefficient, typically ranging from 0.3 to 0.7. For participants At the current sampling time Control weights.

[0143] It is worth noting that, in order to prevent the control weights of each participant from being zero, and to prevent dynamic weight adjustment from failing during sudden disturbances, weight boundary constraints are established:

[0144] (32);

[0145] Step S4: For the total cost function, the ADMM algorithm is used for distributed solution to obtain the optimal control inputs for the DYC system, AFS system, ARS system and ASS system.

[0146] The optimal control inputs include: the additional yaw moment of the DYC system, the front wheel steering angle of the AFS system, the rear wheel steering angle of the ARS system, and the roll moment of the ASS system.

[0147] Step S4 specifically includes steps S41-S43.

[0148] Step S41: For the total cost function, reconstruct it into the following form:

[0149] (33);

[0150] in, Current sampling time Lower participants The control weight, Current sampling time Participants The cost function, For participants Control the input sequence vector, As a global consensus variable, , For participants control input sequence vector At the coordination point at the global level, , , and These are the control inputs for the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. and The minimum and maximum values ​​of the input are controlled by participant 1, respectively. and These are the minimum and maximum values ​​of the input actions controlled by participant 2. and These represent the minimum and maximum values ​​of the input controlled by participant 3. and These are the minimum and maximum values ​​of the input controlled by participant 4.

[0151] Step S42: Construct the augmented Lagrangian function based on the reconstructed problem. (34);

[0152] in, For the penalty function, Denotes the square of the L2 norm. As dual variables, , , , and These measures the degree of inconsistency between the local solutions and the global consensus variables of the four participants: the DYC system, the AFS system, the ARS system, and the ASS system. To measure participants Local solutions and global consensus variables The degree of inconsistency between them The global solution represents the optimal control input for the DYC system, AFS system, ARS system, and ASS system.

[0153] Step S43: Use the Alternating Direction Multiplier Method (ADMM) for distributed solution.

[0154] Update alternately using the following three steps:

[0155] Step a: Each participant independently solves the local optimization problem with actuator constraints:

[0156] (35);

[0157] in, The iteration period is defined as follows.

[0158] Step b: Collect local solutions from all participants and update the global consensus variables:

[0159] (36);

[0160] This problem has a closed-ended solution:

[0161] (37);

[0162] Step c: Update the dual variable to penalize consensus violations:

[0163] (38);

[0164] Based on alternating updates from steps a to c, the convergence condition of the ADMM algorithm is:

[0165] (39);

[0166] in, and These are the preset original and dual tolerances.

[0167] In some other embodiments, the chassis multi-system cooperative game control method for path tracking and stability further includes step S5: allocating longitudinal force to each wheel of the vehicle based on the solved additional yaw moment, and allocating active suspension force to each wheel of the vehicle based on the solved roll moment.

[0168] Specifically, considering the peak torque of the motor, the additional yaw moment, the total desired longitudinal force, and the constraints of the road adhesion coefficient, the optimal distribution function of the vehicle's longitudinal force is designed as follows:

[0169] (40);

[0170] In the formula, This represents the peak torque of the motor. Tire radius; The total longitudinal force required by the driver; , , and These are the longitudinal forces on the left front wheel, right front wheel, left rear wheel, and right rear wheel, respectively. The total lateral force required by the driver; , , and The lateral forces are respectively the left front wheel, right front wheel, left rear wheel, and right rear wheel; To add yaw moment; Wheelbase; This represents the road surface adhesion coefficient for a single tire.

[0171] The active suspension force distribution rules are as follows:

[0172] (41);

[0173] In the formula, The wheelbase is the distance between the wheels. For the roll moment, Left front wheel active suspension force, Right front wheel active suspension force, Left rear active suspension force, Right rear wheel active suspension force.

[0174] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Those skilled in the art will clearly understand that the technologies in the embodiments of this invention can be implemented using software and necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments of this invention.

[0175] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A chassis multi-system cooperative game control method for path tracking and stability, the chassis multi-system comprising: The direct yaw moment control (DYC) system, the active front steering (AFS) system, the active rear steering (ARS) system and the active suspension (ASS) system are characterized in that the control method comprises the following steps: Step S1: a vehicle dynamics and path tracking coordination control model is constructed in combination with a vehicle dynamics model and a path tracking model; Step S2: a vehicle state prediction equation is established through a model prediction control algorithm based on the discretized vehicle dynamics and path tracking coordination control model; Step S3: based on the vehicle state prediction equation and a defined vehicle desired state, the DYC system, the AFS system, the ARS system and the ASS system are taken as participants in a cooperative game, and a total cost function is constructed with the goal of balancing path tracking accuracy and driving stability, wherein, when the total cost function is constructed, the control weights of the participants are dynamically adjusted according to real-time vehicle states; Step S4: the ADMM algorithm is used to solve the total cost function in a distributed manner to obtain optimal control inputs of the DYC system, the AFS system, the ARS system and the ASS system.

2. The chassis multi-system cooperative game control method for path tracking and stability according to claim 1, characterized in that, The vehicle dynamics and path tracking coordination control model in step S1 is: ； wherein is the first derivative of , is the state variable, , is the side slip angle of the center of mass, is the yaw rate, is the roll angle of the vehicle body, is the roll rate, is the lateral position of the vehicle, is the heading angle error, the superscript denotes the transpose, , , and are the control inputs of the four participants, DYC system, AFS system, ARS system and ASS system, respectively, and are given by , , , , is the additional yaw moment, and are the front and rear wheel steering angles, respectively, is the roll moment; ; ； ； ； ； ； wherein, and are the front and rear axle side stiffnesses, respectively, is the total vehicle mass, is the longitudinal velocity, and are the distances from the center of mass to the front and rear axles, respectively, is the moment of inertia of the total vehicle about the roll axis, is the moment of inertia of the sprung mass of the vehicle about the roll axis, is the sprung mass of the vehicle, is the roll arm, expressed as the distance from the center of mass to the roll center, is the gravitational acceleration, is the axle roll stiffness, is the roll damping, denotes the curvature of the reference path.

3. The chassis multi-system cooperative game control method for path tracking and stability according to claim 2, characterized in that, The vehicle state prediction equation in step S2 is: ； ； ； ； ； ； wherein is a predicted sequence of vehicle states at the current sampling time , is a sequence of states at the current sampling time , is a predicted vehicle state at a future time at the current sampling time , is a prediction step size, is a state matrix of the prediction equation, , is an identity matrix, is a discrete time, is a control input matrix of the participants of the prediction equation, , is a control input sequence vector of the participants at the current sampling time , is a control input applied at a future time at the current sampling time , is an influence of a sequence of external disturbances on the state prediction for a future number of prediction steps, is an external disturbance at a future time at the current sampling time .

4. The chassis multi-system cooperative game control method for path tracking and stability according to claim 1, characterized in that, The defined vehicle desired state in step S3 comprises a desired yaw rate , a desired vehicle roll angle and a desired center of mass side slip angle ; the desired yaw rate wherein is the road adhesion coefficient, is the gravitational acceleration, is the longitudinal velocity, is the wheelbase, , is the vehicle stability factor, , and are the distances of the center of mass to the front and rear axles, respectively, is the total vehicle mass, and are the front and rear axle cornering stiffnesses, respectively, is the front wheel steering angle, is the sign function; the desired vehicle roll angle ; the desired center of mass side slip angle .

5. The chassis multi-system cooperative game control method for path tracking and stability according to claim 1, characterized in that, The step S3 specifically comprises: Step S31: a cost function of each participant is constructed, which is specifically represented as: ； in, To the current sampling time At that time, participants The cost function, To predict the step size, This is the sequence number of the prediction step. For participants The expected state matrix, To the current sampling time Predicted future moments state, , For the desired yaw rate, To achieve the desired vehicle roll angle, The desired centroid sideslip angle is indicated by the superscript. Indicates transpose. For participants State weight coefficients, For participants At the current sampling time The future moment The applied control input, For participants The weighting coefficients of the input quantities It is an L2 norm; Step S32: constructing a total cost function based on the dynamically adjusted control weights of each participant is: ； in, , , , The control weights for the four participants—DYC system, AFS system, ARS system, and ASS system—are respectively. , , and At the current sampling time The cost functions of the four participants: DYC system, AFS system, ARS system, and ASS system.

6. The chassis multi-system cooperative game control method for path tracking and stability according to claim 1, characterized in that, In step S3, the control weights of the participants are dynamically adjusted according to real-time vehicle states, which specifically comprises: Based on the normalized centroid side slip angle deviation, yaw rate deviation, roll angle deviation, lateral displacement deviation and heading angle deviation, a fusion weight vector based on multiple state errors is defined; ； wherein is a normalized side slip angle error, is a normalized yaw rate error, is a normalized vehicle roll angle error, is a normalized lateral position error, is a normalized heading angle error, is a weight normalization factor, , is a positive number set to prevent the denominator from being zero, is a fusion weight for the side slip angle error, is a fusion weight for the yaw rate error, is a fusion weight for the vehicle roll angle error, is a fusion weight for the lateral position error, is a fusion weight for the heading angle error; The base weight of the controller corresponding to each participant is calculated; ； in, For participants Corresponding to the controller's base weights, , , , and Participants of , , , and The corresponding weights; The base weight of the controller corresponding to each participant is adaptively adjusted based on the vehicle speed; ； ； wherein, is the longitudinal speed, is the longitudinal speed maximum, , , and are the base weights of the controllers corresponding to the four participants DYC system, AFS system, ARS system and ASS system respectively, , , and are the weights of the four participants DYC system, AFS system, ARS system and ASS system respectively after speed adjustment. The weight adjusted by the vehicle speed of each participant is normalized, and a first-order inertia smoothing is added to obtain the control weight of each participant at the current sampling time of the current sampling time ; ； wherein, is the weight of each participant after speed adjustment and normalization, is a smoothing coefficient.

7. The chassis multi-system cooperative game control method for path tracking and stability according to claim 1, characterized in that, Step S4 specifically comprises: the total cost function is reconstructed into an augmented Lagrangian function, and the ADMM algorithm is used to solve it to obtain optimal control inputs of the DYC system, the AFS system, the ARS system and the ASS system, including: additional yaw moment of the DYC system, front wheel steering angle of the AFS system, rear wheel steering angle of the ARS system and roll moment of the ASS system.

8. The chassis multi-system cooperative game control method for path tracking and stability according to claim 7, characterized in that, Further comprising the following steps: Step S5: based on the solved additional yaw moment, longitudinal forces are distributed to each wheel of the vehicle, and based on the solved roll moment, active suspension forces are distributed to each wheel of the vehicle.

Images