Lateral path tracking control method for multi-axis distributed driving vehicle

Abstract

The invention discloses a lateral path tracking control method for a multi-axis distributed driving vehicle. The method comprises the following steps: constructing a lateral path tracking control model based on LPV; acquiring local feedback control gain of each subsystem based on H-infinity robust control; solving a lateral path tracking control law; calculating a double-front-axle wheel turning angle; lateral path tracking control is performed. According to the invention, a lateral path tracking error model is established based on a monorail vehicle dynamics model containing uncertain disturbance terms; the designed lateral path tracking control method based on LPV and H-infinity robust control can effectively weaken the influence of external disturbance, system modeling errors and uncertainty of parameters of a multi-axis distributed driving vehicle, and the lateral path tracking precision and robustness of the vehicle are ensured. The problem of path tracking control is solved based on H-infinity robust control, and a thought is provided for path tracking control of distributed driving vehicles under uncertainty, strong nonlinearity and various external interferences.

Description

Technical Field

[0001] The present invention belongs to the field of vehicle motion control technology, and specifically relates to a method based on linear parameter varying (LPV) and H ∞ The robust control method for lateral path tracking of multi-axis distributed drive vehicles is particularly suitable for lateral path tracking of multi-axis distributed drive vehicles under nonlinear working conditions with parameter uncertainty and external disturbances. Background Art

[0002] With the technological innovation of power drive systems, distributed drive strategies with advantages such as independent controllable torque and angle and rapid motor response have been widely used in unmanned ground vehicles, which can significantly improve their flexibility, stability and work efficiency. The maneuverability of vehicles is a key factor in ensuring efficient transportation and delivery. However, during the lateral path tracking process, multi-axis distributed vehicles are often affected by various uncertain factors such as parameter fluctuations, modeling errors and external disturbances, which can easily lead to a decrease in the vehicle's lateral path tracking accuracy. Compared with light vehicles such as two-axle passenger cars, the dynamic characteristics of multi-axis distributed drive vehicles are more complex. Under the influence of external disturbances, system modeling errors and the uncertainty of their own parameters, the path tracking control accuracy of multi-axis vehicles is low and the robustness is poor.

[0003] Traditional path-tracking control algorithms, such as proportional-integral-derivative control, linear quadratic control, and sliding mode control, struggle to adapt to nonlinear conditions at high speeds or with large curvatures. They often require a large number of accurate vehicle dynamic states and are sensitive to external disturbances such as changes in path curvature. Compared to light vehicles such as two-axle passenger cars, multi-axle distributed drive vehicles have more complex dynamic characteristics, with strong uncertainty in their characteristic parameters, slow control response, and weak anti-interference capabilities. Furthermore, external factors such as large variations in road adhesion, measurement noise, and crosswinds can interfere with vehicle control quality. The vehicle must maintain high lateral path tracking accuracy despite various disturbances. This is crucial for autonomous driving systems and advanced driver assistance systems.

[0004] Compared with other control algorithms, robust control is less affected by parameter uncertainty and external interference. The dynamic characteristics of the vehicle are taken into account in the design of the control law, and the tracking effect is better under nonlinear working conditions such as high speed, large curvature, and low road adhesion. At the same time, due to the time-varying and uncertainty of vehicle parameters, the non-time-varying output feedback H ∞ The robustness of the controller is not enough to meet the performance requirements. Considering the bounded uncertainty parameters and the changes of measurable time-varying parameters within a certain range, the LPV system is introduced, which is better than the fixed gain H ∞ The controller has better robustness.

[0005] In summary, in order to achieve accurate tracking of the reference path by a multi-axis distributed vehicle under parameter uncertainty, system modeling error and external environmental interference, it is necessary to invent a lateral path tracking control method for a multi-axis distributed drive vehicle to achieve lateral path tracking control with good real-time performance, strong anti-interference ability and high precision under all working conditions. Summary of the Invention

[0006] In order to solve the above problems in the prior art, the present invention aims to provide a method based on LPV and H ∞ A robustly controlled lateral path tracking control method for a multi-axis distributed drive vehicle enables the vehicle to accurately track the reference trajectory under parameter uncertainty, system modeling errors, and external environmental interference.

[0007] In order to achieve the above object, the technical solution of the present invention is as follows:

[0008] A multi-axis distributed drive vehicle lateral path tracking control method, which utilizes a multi-axis distributed drive vehicle lateral path tracking control system for control, comprises the following steps:

[0009] Step 1: Initialization

[0010] The initialization module is responsible for checking whether the perception module, mapping and positioning module, decision-making and planning module, and actuator signal transmission and reception are normal, loading the parameters of the single-track vehicle dynamics model containing uncertain disturbance terms, loading the parameters of the lateral path tracking model that takes into account uncertain disturbance terms, loading the parameters of the lateral path tracking controller, and receiving the position information, curvature, and heading information at the desired path matching point issued by the decision-making and planning module. The decision-making and planning module is the upper module of the lateral path tracking control system, which is used to generate the desired path information of the multi-axis distributed drive vehicle. The desired path matching point is the projection point of the current position of the multi-axis distributed drive vehicle on the desired path. The projection point is the point at which the multi-axis distributed drive vehicle is closest to the desired path.

[0011] Step 2: Build a LPV-based lateral path tracking control model

[0012] A lateral path tracking control model based on LPV is constructed to effectively describe the nonlinear dynamic characteristics of the vehicle during actual operation.

[0013] Step 3: Get H-based ∞ Local feedback control gain of each subsystem in robust control

[0014] By constructing H ∞ After solving the linear matrix inequality with performance constraints, the local feedback control gains of each subsystem in the LPV-based convex polyhedron system are obtained offline.

[0015] Step 4: Solve the lateral path tracking control law

[0016] Based on the weight coefficients of the LPV-based convex polyhedron system obtained online in step 2 and the local feedback control gains of each subsystem in the LPV-based convex polyhedron system obtained offline in step 3, the lateral path tracking control law is obtained.

[0017] Step 5: Solve the wheel angles of the dual front axles

[0018] Based on the Ackermann geometry relationship, the first-axle wheel angle obtained in step 4 is used to obtain the second-axle wheel angle, and then the final dual front axle wheel angle control law is obtained.

[0019] Step 6: Perform Lateral Path Following Control

[0020] The dual front axle wheel angles δ1 and δ2 determined in steps 4 and 5 are subjected to control constraints and filtering. These are then transmitted as control signals at a fixed frequency to the multi-axle distributed drive vehicle. The vehicle actuators use these control signals to perform lateral path tracking control of the multi-axle distributed drive vehicle. Furthermore, a determination is made as to whether the vehicle has successfully reached its destination. If so, the vehicle completes the control task; otherwise, the process proceeds to step 1.

[0021] Furthermore, the multi-axis distributed drive vehicle lateral path tracking control system, mapping and positioning module and decision-making planning module are all burned into the controller, and the controller is connected to the perception module and actuator respectively through a data bus.

[0022] Furthermore, the multi-axis distributed drive vehicle lateral path tracking control system includes an initialization module, a control module based on LPV and H ∞ The lateral path tracking control module of robust control and the angle calculation module based on Ackermann geometry. The multi-axis distributed drive vehicle lateral path tracking control system is hereinafter referred to as the lateral path tracking control system; the LPV and H ∞ The lateral path tracking control module of robust control is hereinafter referred to as the lateral path tracking control module, the LPV is the abbreviation of linear parameter time-varying; the angle calculation module based on Ackermann geometric relationship is hereinafter referred to as the angle calculation module.

[0023] The initialization module is responsible for checking the normal functioning of the perception module, mapping and localization module, decision-making and planning module, and actuator signal transmission and reception. It also loads the parameters of the single-track vehicle dynamics model containing uncertain disturbance terms, the parameters of the lateral path tracking error model, and the parameters and control constraints of the lateral path tracking control module. These modules, along with the perception module, mapping and localization module, and decision-making and planning module, are common modules outside the lateral path tracking control system and work together to ensure the normal operation of the multi-axle distributed drive vehicle.

[0024] The lateral path tracking control module is composed of a lateral path tracking control model based on LPV and a lateral path tracking control model based on H ∞ The robust control system is composed of a lateral path tracking controller, which effectively suppresses various uncertain disturbances and accurately tracks the path. These uncertain disturbances are the sum of disturbances in the multi-axis distributed drive vehicle's path tracking process. These include disturbances caused by perturbations in vehicle mass and moment of inertia, modeling errors and disturbances caused by unmodeled external dynamic characteristics, sensor signal transmission delays, actuator response lags, and even disturbances triggered by faults. This sum of disturbances is the sum of disturbances in the lateral motion channel and the yaw motion channel.

[0025] The steering angle calculation module solves the second axis steering angle through the first axis steering angle based on the Ackermann steering geometry relationship to obtain the final steering angle control law.

[0026] Furthermore, the single-track vehicle dynamics model containing the uncertain disturbance term in step 1 is established according to Newton's second law as follows:

[0027]

[0028] Where m is the vehicle mass; v x is the longitudinal velocity of the vehicle; v y is the vehicle lateral speed, is the derivative of the vehicle's lateral velocity; is the vehicle heading angle, is the first-order derivative of the vehicle heading angle, the yaw rate, is the second-order derivative of the vehicle heading angle; δ1 is the first-axis wheel angle; I z is the moment of inertia of the vehicle's center of mass around the Z axis; L i is the distance from each axis to the center of mass. For ease of expression, L3<0, L4<0; C αi is the tire cornering stiffness of each axle; α i is the tire slip angle; a s is the angle proportional coefficient; d1 and d2 are uncertainty disturbances; i = 1 to 4, corresponding to the four axes of the vehicle respectively.

[0029] Furthermore, the lateral path tracking error model considering the uncertain disturbance term in step 1 is expressed as follows:

[0030]

[0031] In the formula, x(t) is the state quantity, is the rate of change of the state quantity; u(t) is the control input δ1; y(t) is the output quantity, w(t)

[0032] is the total disturbance, w(t)=[d1 d2 d3 -k r v x +d4] T , κ r is the road curvature, d3 and d4 are the uncertainty disturbances of the lateral and directional motion channels respectively; l d is the preview distance;

[0033]

[0034]

[0035] Furthermore, the method for constructing the LPV-based lateral path tracking control model in step 2 is as follows:

[0036] Bounded uncertainty variables are selected as the scheduling variables of LPV, and the bounded uncertainty variables include vehicle mass, moment of inertia, tire cornering stiffness and longitudinal speed; it is assumed that the lower and upper bounds of vehicle mass, moment of inertia, tire cornering stiffness of the front two axles, tire cornering stiffness of the rear two axles and longitudinal speed are Among them, the lateral stiffness of the front two axle tires C α1 =C α2 =C αf , rear two axle tire cornering stiffness C α3 =C α4 =C αr , represents the upper bound of the corresponding scheduling variable, * Indicates the lower bound of the corresponding scheduling variable.

[0037] The lateral path tracking control model is described as the following LPV-based convex polyhedron system, which consists of 32 subsystems:

[0038]

[0039] The matrix set describing the LPV-based convex polytope system is defined as Ξ = {A, B1, B2}, and satisfies the following formula:

[0040]

[0041] Among them, j ={A j ,B 1j ,B 2j} represents the vertices of the convex polyhedron, α j is the weight coefficient of the convex polyhedron system based on LPV solved in real time, which is related to the changes of each scheduling variable. j≥0, j=1,…,32, corresponding to the 32 subsystems of LPV, and the vertices and weight coefficients of the convex polyhedron satisfy the following formula:

[0042]

[0043] Furthermore, the step 3 is to obtain the ∞ The method of local feedback control gain of each subsystem of robust control is as follows:

[0044] For each subsystem in the LPV-based convex polyhedron system, if the following linear matrix inequality conditions are satisfied, then the closed-loop system composed of each subsystem and the vehicle has the following H ∞ performance:

[0045]

[0046] Where, scalar γ>0, P and Q are both positive definite matrices, P inv represents the inverse matrix of P, * represents the symmetric part of the matrix.

[0047] By constructing H ∞ After solving the linear matrix inequality under performance constraints, the local feedback control gains of each subsystem in the LPV-based convex polyhedron system are obtained offline as follows:

[0048] K j =Q j P j

[0049] Where K j is the local feedback control gain of each subsystem.

[0050] Furthermore, the method for solving the lateral path tracking control law in step 4 is as follows:

[0051] Based on the weight coefficients of the LPV-based convex polyhedron system obtained online in step 2 and the local feedback control gains of each subsystem in the LPV-based convex polyhedron system obtained offline in step 3, the following lateral path tracking control law is obtained, that is, the first-axis wheel angle described in step 1:

[0052]

[0053] The state equation of the lateral path tracking control system is further obtained as follows:

[0054]

[0055] Where A cj =A j -B 1jK(α).

[0056] Furthermore, the dual front axle wheel angle control law in step 5 is as follows:

[0057]

[0058] The present invention has the following beneficial effects:

[0059] 1. Based on the single-track vehicle dynamics model with uncertain disturbance terms, the paper establishes a lateral path tracking error model. ∞ The robust control lateral path tracking control method can effectively reduce the impact of external disturbances, system modeling errors and the uncertainty of the vehicle's own parameters on the multi-axis distributed drive vehicle, ensuring the vehicle's lateral path tracking accuracy and robustness.

[0060] 2. The present invention is based on H ∞ Robust control solves the path tracking control problem and provides ideas for distributed drive vehicles, especially heavy-duty distributed drive vehicles, in the face of path tracking control under uncertainty, strong nonlinearity and various external interferences. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic diagram of the lateral path tracking control method of the present invention.

[0062] Figure 2 Schematic diagram of the lateral path tracking error model of the present invention.

[0063] Figure 3 It is an implementation flow chart of the present invention.

[0064] Figure 4 This is a schematic diagram of a multi-axle distributed drive vehicle deviating from the lane center due to uncertain disturbances.

[0065] Figure 5 This is a schematic diagram of a multi-axle distributed drive vehicle colliding with a road boundary due to uncertain disturbances.

[0066] Figure 6 It is a schematic diagram of the single-track vehicle dynamics model of a multi-axle distributed drive vehicle. DETAILED DESCRIPTION

[0067] The following reference Figure 1-6 , a lateral path tracking control method for a multi-axis distributed drive vehicle of the present invention is further described. Figure 1 FIG. 1 is a schematic diagram of the lateral path tracking control method of the present invention, which mainly consists of two modules:

[0068] (1) Based on LPV and H ∞Robust control lateral path tracking control module: The single-track vehicle dynamics model sends the vehicle's real-time state information, including the real-time position information of the multi-axis distributed drive vehicle, road curvature information, heading information and speed information, to the lateral path tracking error model that considers uncertain disturbance terms. At the same time, the vehicle's upper-level decision-making and planning module also sends the position information, curvature and heading information of the expected path matching point to the lateral path tracking error model that considers uncertain disturbance terms. Due to the existence of various uncertain disturbance terms, the vehicle deviates from the lane centerline during the path tracking process, such as Figure 4 It may even hit the road edge, such as Figure 5 In order to achieve accurate tracking of the reference trajectory by the vehicle under the uncertainty of vehicle parameters, system modeling errors and external environment interference, a vehicle tracking system based on LPV and H is designed. ∞ The lateral path tracking control module of the robust control controls the wheel angle of the dual front axles. By introducing a closed-loop system with H ∞ The performance condition can be used to transform the path tracking problem into a generalized eigenvalue problem to solve the LMI constraint. The closed-loop control is achieved by solving the state feedback control gain matrix offline. The LMI is the abbreviation of linear matrix inequality. Due to the time-varying and uncertain nature of vehicle parameters, the non-time-varying output feedback H ∞ The controller's robustness is insufficient to meet performance requirements. Considering the bounded uncertainty parameters and measurable time-varying parameters within a certain range, a polytope-based LPV system is introduced. Vehicle mass, moment of inertia, front and rear tire cornering stiffnesses, and longitudinal speed are selected as the system's dispatching variables. Time-varying feedback control can be achieved by solving the corresponding weight function online. The final first-axis steering angle control law is obtained by multiplying the local feedback control gains of each subsystem in the LPV-based convex polytope system by the corresponding weight coefficients.

[0069] (2) Angle calculation module: Based on the Ackermann geometry relationship, the second-axis angle is solved using the first-axis angle obtained in the lateral path tracking control module, and then the final dual front axle angle control law is obtained.

[0070] A multi-axis distributed drive vehicle lateral path tracking control system includes an initialization module, a lateral path tracking control module and a turning angle calculation module.

[0071] like Figure 1 As shown, the steps of the present invention are as follows:

[0072] Step 1: Establish a single-track vehicle dynamics model with uncertain disturbance terms

[0073] Define O-XYZ, o-xyz, and o w -x w y w zw They are the earth coordinate system, vehicle coordinate system and tire coordinate system respectively. Figure 6 The two-degree-of-freedom single-track vehicle dynamics model shown in the figure has a steering system using a dual front axle steering mechanism. The corresponding dynamic differential equation is:

[0074]

[0075] Where, F xi and F yi are the longitudinal force and lateral force of each axle tire respectively; ΔM z is the additional yaw moment.

[0076] Assuming the tire cornering characteristics are linear, the lateral force F of each axle tire is yi for

[0077] F yi =-C αi α i

[0078] Where, tire slip angle α i Expressed as

[0079]

[0080] Where R w is the tire rolling radius; the longitudinal speed v of each tire wheel center xwi It can be approximated as the vehicle longitudinal velocity v x .

[0081] The steering center of the dual front axle steering is taken on the extension line of the center line of the third and fourth axes, that is, the longitudinal distance between the instantaneous steering center o' and the fourth axis is d c =0.5(|L4||-|L3|). According to Ackerman geometry, we get

[0082]

[0083] The relationship between the dual front axle angles is obtained as follows:

[0084] δ2=a s δ1

[0085] In the formula, the angle proportional coefficient is

[0086] Under the small angle assumption, the single-track vehicle dynamics model with uncertain disturbance terms can be obtained:

[0087]

[0088] Step 2: Design a lateral path tracking error model with uncertain disturbance terms

[0089] Define the Frenet coordinate system τ r o r n r , establish as Figure 2 The lateral path tracking error model is shown in Figure 2. To track the desired path, the lateral offset error and heading error are defined as e dL and In t i and t i+1 At the moment, there is

[0090]

[0091] Where Δt = t i+1 -t i ; is the desired heading angle, and its first-order differential form is

[0092]

[0093] Further, the differential forms of lateral error and heading error can be obtained:

[0094]

[0095] Combined with the single-track vehicle dynamics model obtained in step 1, the lateral path tracking control state space equation containing uncertain disturbance terms can be obtained:

[0096]

[0097] Step 3: Design LPV-based lateral path tracking control model

[0098] In practical application scenarios such as transshipment and delivery, the curb mass, moment of inertia, tire pressure, etc. of multi-axle distributed drive vehicles will change, and the lateral characteristics of their tires will also be nonlinear. Therefore, based on the lateral path tracking error model in step 2, a lateral path tracking control model based on LPV is constructed to effectively describe the nonlinear dynamic characteristics existing in the actual operation of the vehicle. Bounded uncertainty variables such as vehicle mass, moment of inertia, lateral stiffness of the tires of the front two axles, lateral stiffness of the tires of the rear two axles, and longitudinal speed are selected as scheduling variables of the LPV system. Assume that their lower and upper bounds are

[0099] The path tracking control model described in step 2 can be described as a convex polyhedron system based on LPV

[0100]

[0101] Define the set of matrices describing the system as Ξ={A,B1,B2}, and satisfy

[0102]

[0103] Among them, j ={A j ,B 1j ,B 2j} represents the vertices of the convex polyhedron, α j is the weight coefficient of the convex polyhedron system based on LPV solved in real time, which is related to the changes of each scheduling variable. j ≥0, j=1,…,32, corresponding to the 32 subsystems of the LPV system, and the vertices and weight coefficients of the polytope satisfy the following formula:

[0104]

[0105] and

[0106] Step 4: Design based on H ∞ Robust control of lateral path tracking controller

[0107] Design the following lateral tracking control law

[0108]

[0109] The corresponding state space equation of the lateral path tracking control module can be further obtained:

[0110]

[0111] Where A cj =A j -B 1j K(α).

[0112] Each subsystem in the LPV-based convex polyhedron system meets the following performance conditions:

[0113] 1) When w(t) = 0,

[0114]

[0115] 2) When w(t)∈L1[0,∞),

[0116]

[0117] Where, ||y(t)|| 2 =y T (t)y(t);λ max (P) represents the maximum eigenvalue of the real positive definite matrix P, P = P T >0.

[0118] In order to facilitate the ∞The lateral path tracking controller of robust control is designed and analyzed, and the following lemma is given:

[0119] Schur's complement theorem: For a given symmetric matrix Among them S 11 is an r×r dimensional non-singular matrix, then the following conditions are equivalent:

[0120] 1) S < 0;

[0121] 2)S 11 <0,

[0122] 3)S 22 <0,

[0123] Through this lemma, the nonlinear matrix inequality can be transformed into LMI solution. The closed-loop system composed of each subsystem and the vehicle has H ∞ Theorem of performance conditions and its proof:

[0124] Assume that the scalar γ>0, the closed-loop system has H ∞ The performance condition is that there exist positive definite matrices P, Q such that the following LMI holds

[0125]

[0126] Proof: Define the Lyapunov function as V = x T Px, whose derivative is

[0127]

[0128] Introducing evaluation index

[0129]

[0130] Where,

[0131] If J H <0 is established, which can ensure the system's H ∞ Performance. For the convenience of calculation, we further multiply Λ1 by Get 0

[0132]

[0133] By Lemma 1, Λ1<0 is equivalent to

[0134]

[0135] Let P inv =P -1 , And Ac =A-B1K Substituting into the above formula we can get

[0136]

[0137] For a positive scalar γ, if there exists a positive definite matrix P, Q such that Λ2<0, then Λ1<0, that is,

[0138]

[0139] The points can be obtained

[0140]

[0141] in,

[0142]

[0143] Further sorting can be obtained

[0144]

[0145] From the above derivation, we can know that Λ2<0 holds true, which means that the system's H ∞ Performance guaranteed. Proven.

[0146] This theorem can be used to obtain the local feedback control gains of each subsystem in the LPV-based convex polyhedron system offline.

[0147] K j =Q j P j

[0148] Step 5: Solve the lateral path tracking control law

[0149] Combining the weight coefficients of the LPV-based convex polyhedron system obtained online in step 3 and the local feedback control gains of each subsystem in the LPV-based convex polyhedron system obtained offline in step 4, the overall lateral path tracking control law of the lateral path tracking control module can be further obtained, that is, the first axis wheel angle described in step 1 is

[0150]

[0151] Step 6: Calculate the wheel angles of the dual front axles

[0152] Based on the Ackermann geometry, the wheel angle of the second axle can be obtained by using the wheel angle of the first axle obtained in step 5, and then the final dual front axle wheel angle control law can be obtained.

[0153]

[0154] Step 7: Perform Lateral Path Following Control

[0155] The dual front axle wheel angles δ1 and δ2 determined in steps 5 and 6 are subjected to control constraints and filtering. These are then transmitted as control signals at a fixed frequency to the multi-axle distributed drive vehicle. The vehicle actuators perform lateral path tracking control based on the angle signals. Furthermore, a determination is made as to whether the vehicle has successfully reached its destination. If so, the vehicle completes the control task; otherwise, the process proceeds to step 1.

[0156] Furthermore, the multi-axis distributed drive vehicle lateral path tracking control system, mapping and positioning module and decision-making planning module are all burned into the controller, and the controller is connected to the perception module and actuator respectively through a data bus.

[0157] The present invention is not limited to the details of the above embodiments. Any equivalent concepts or modifications within the technical scope disclosed by the present invention are included in the protection scope of the present invention.

Claims

1. A lateral path tracking control method for a multi-axis distributed drive vehicle, characterized by: The multi-axis distributed drive vehicle lateral path tracking control system is used for control, including the following steps: Step 1: Initialization The initialization module is responsible for checking whether the perception module, mapping and positioning module, decision-making and planning module, and actuator signal transmission and reception are normal, loading the parameters of the single-track vehicle dynamics model containing uncertain disturbance terms, loading the parameters of the lateral path tracking model considering uncertain disturbance terms, loading the parameters of the lateral path tracking controller, and receiving the position information, curvature and heading information at the expected path matching point issued by the decision-making and planning module; the decision-making and planning module is the upper module of the lateral path tracking control system, which is used to generate the expected path information of the multi-axis distributed drive vehicle; the expected path matching point is the projection point of the current position of the multi-axis distributed drive vehicle on the expected path, and the projection point is the point of the multi-axis distributed drive vehicle closest to the expected path; Step 2: Build a LPV-based lateral path tracking control model Construct a lateral path tracking control model based on LPV to effectively describe the nonlinear dynamic characteristics of the vehicle during actual operation; Step 3: Get H-based ∞ Local feedback control gain of each subsystem in robust control By constructing H ∞ After solving the linear matrix inequality with performance constraints, the local feedback control gains of each subsystem in the LPV-based convex polyhedron system are obtained offline; Step 4: Solve the lateral path tracking control law Based on the weight coefficients of the LPV-based convex polyhedron system obtained online in step 2 and the local feedback control gains of each subsystem in the LPV-based convex polyhedron system obtained offline in step 3, the lateral path tracking control law is obtained; Step 5: Solve the wheel angles of the dual front axles Based on the Ackermann geometry, the wheel angle of the first axle obtained in step 4 is used to obtain the wheel angle of the second axle, and then the final dual front axle wheel angle control law is obtained; Step 6: Perform Lateral Path Following Control Control constraints are applied to the dual front axle wheel angles δ1 and δ2 determined in steps 4 and 5, and filtered. These are then sent as control signals at a fixed frequency to the multi-axle distributed drive vehicle. The vehicle actuator performs lateral path tracking control on the multi-axle distributed drive vehicle based on the angle signals. Furthermore, a determination is made as to whether the vehicle has arrived at the destination correctly. If so, the vehicle completes the control task; otherwise, the process proceeds to step 1.

2. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The multi-axis distributed drive vehicle lateral path tracking control system, mapping and positioning module, and decision-making planning module are all burned into the controller, and the controller is connected to the perception module and actuator respectively through a data bus.

3. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The multi-axis distributed drive vehicle lateral path tracking control system includes an initialization module, a LPV and H ∞ Robust control lateral path tracking control module and Ackerman geometry-based angle calculation module; the multi-axis distributed drive vehicle lateral path tracking control system hereinafter referred to as the lateral path tracking control system; the LPV and H-based ∞ The lateral path tracking control module of the robust control is hereinafter referred to as the lateral path tracking control module, the LPV is the abbreviation of linear parameter time-varying; the angle calculation module based on the Ackermann geometric relationship is hereinafter referred to as the angle calculation module; The initialization module is responsible for checking whether the perception module, mapping and positioning module, decision-making and planning module, and actuator signal transmission and reception are normal, loading the parameters of the single-track vehicle dynamics model containing uncertain disturbance terms, loading the parameters of the lateral path tracking error model, and loading the parameters and control constraints of the lateral path tracking control module; the perception module, mapping and positioning module, and decision-making and planning module are common modules outside the lateral path tracking control system, and work together with the lateral path tracking control system to ensure the normal operation of the multi-axis distributed drive vehicle; The lateral path tracking control module is composed of a lateral path tracking control model based on LPV and a lateral path tracking control model based on H ∞ A robust control lateral path tracking controller is constructed to effectively suppress various uncertain disturbance terms and accurately track the path. The uncertain disturbance term is the sum of disturbances in the multi-axis distributed drive vehicle path tracking process, including disturbances caused by perturbations in vehicle mass and moment of inertia, disturbances caused by modeling errors and unmodeled external dynamic characteristics, delays in sensor signal transmission, and disturbances triggered by actuator response lags or even failures. The sum of disturbances is the sum of disturbances in the lateral motion channel and disturbances in the heading motion channel. The steering angle calculation module solves the second axis steering angle through the first axis steering angle based on the Ackermann steering geometry relationship to obtain the final steering angle control law.

4. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The single-track vehicle dynamics model with uncertain disturbance terms described in step 1 is established according to Newton's second law as follows: Where m is the vehicle mass; v x is the longitudinal velocity of the vehicle; v y is the vehicle lateral speed, is the derivative of the vehicle's lateral velocity; is the vehicle heading angle, is the first-order derivative of the vehicle heading angle, the yaw rate, is the second-order derivative of the vehicle heading angle; δ1 is the wheel angle of the first axis; I z is the moment of inertia of the vehicle's center of mass around the Z axis; L i is the distance from each axis to the center of mass. For ease of expression, L3<0, L4<0; C αi is the tire cornering stiffness of each axle; α i is the tire slip angle; a s is the angle proportional coefficient; d1 and d2 are uncertainty disturbances; i = 1 to 4, corresponding to the four axes of the vehicle respectively.

5. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The lateral path tracking error model considering the uncertain disturbance term described in step 1 is expressed as follows: In the formula, x(t) is the state quantity, is the rate of change of the state quantity; u(t) is the control input δ1; y(t) is the output quantity, w(t) is the total disturbance, w(t)=[d1d2 d3 -κ r v x +d4] T , k r is the road curvature, d3 and d4 are the uncertainty disturbances of the lateral and directional motion channels respectively; l d is the preview distance; 6. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The method for constructing the LPV-based lateral path tracking control model in step 2 is as follows: Bounded uncertainty variables are selected as the scheduling variables of LPV, and the bounded uncertainty variables include vehicle mass, moment of inertia, tire cornering stiffness and longitudinal speed; it is assumed that the lower and upper bounds of vehicle mass, moment of inertia, tire cornering stiffness of the front two axles, tire cornering stiffness of the rear two axles and longitudinal speed are Among them, the lateral stiffness of the front two axle tires C α1 =C α2 =C αf , rear two axle tire cornering stiffness C α3 =C α4 =C αr , represents the upper bound of the corresponding scheduling variable, * represents the lower bound of the corresponding scheduling variable; The lateral path tracking control model is described as the following LPV-based convex polyhedron system, which consists of 32 subsystems: The matrix set describing the LPV-based convex polytope system is defined as Ξ = {A, B1, B2}, and satisfies the following formula: Among them, j ={A j ,B 1j ,B 2j } represents the vertices of the convex polyhedron, α j is the weight coefficient of the convex polyhedron system based on LPV solved in real time, which is related to the changes of each scheduling variable. j ≥0, j=1,…,32, corresponding to the 32 subsystems of LPV, and the vertices and weight coefficients of the convex polyhedron satisfy the following formula:

7. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: Step 3 obtains the ∞ The method of local feedback control gain of each subsystem of robust control is as follows: For each subsystem in the LPV-based convex polyhedron system, if the following linear matrix inequality conditions are satisfied, then the closed-loop system composed of each subsystem and the vehicle has the following H ∞ performance: Where, scalar γ>0, P and Q are both positive definite matrices, P inv represents the inverse matrix of P, * represents the symmetric part of the matrix; By constructing H ∞ After solving the linear matrix inequality under performance constraints, the local feedback control gains of each subsystem in the LPV-based convex polyhedron system are obtained offline as follows: K j =Q j P j Where K j is the local feedback control gain of each subsystem.

8. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The method for solving the lateral path tracking control law described in step 4 is as follows: Based on the weight coefficients of the LPV-based convex polyhedron system obtained online in step 2 and the local feedback control gains of each subsystem in the LPV-based convex polyhedron system obtained offline in step 3, the following lateral path tracking control law is obtained, that is, the first-axis wheel angle described in step 1: The state equation of the lateral path tracking control system is further obtained as follows: Where, A cj = A j - B 1j K(α).

9. The lateral path tracking control method for a multi-axis distributed drive vehicle according to claim 1, characterized in that: The dual front axle wheel angle control law in step 5 is as follows:

Images