A type-2 fuzzy control-based intelligent semi-trailer trajectory tracking control method, system and device

Abstract

The application discloses a kind of intelligent semi-trailer trajectory tracking control method, system and equipment based on two type fuzzy control, the dynamics model of semi-trailer is established by Lagrange method, this dynamics model accurately simulates vehicle state, and builds trajectory tracking model;Design a two type fuzzy controller, in addition to handling tractor tracking error, the uncertainty interference received by trailer tracking error is covered.This double-level uncertainty processing improves the adaptability and robustness of fuzzy control system in complex environment.Using particle swarm optimization algorithm to optimize the parameters of two type fuzzy controller, effectively improve the controller precision, significantly reduce the difficulty of controller parameter design.In the actual driving process of intelligent commercial semi-trailer, the trajectory tracking accuracy is significantly improved, and the safety of intelligent commercial semi-trailer under heavy load and high speed is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent commercial vehicle automatic driving, in particular to a trajectory tracking control method, system and device for a commercial intelligent semi-trailer. BACKGROUND

[0002] Commercial semi-trailers have become one of the important means in the field of transportation due to their low transportation cost and high transportation efficiency. However, the cost of driver personnel is increasing year by year, and fatigue driving is prone to cause safety accidents, which increases the transportation cost and reduces the transportation efficiency. With the development of high-level automatic driving technology and advanced auxiliary driving technology, new ideas have been brought to commercial semi-trailers to reduce costs and increase efficiency. In high-speed driving, trajectory tracking control of intelligent semi-trailers can be achieved under the premise of pre-set driving trajectory, which realizes the automatic driving of intelligent commercial semi-trailers.

[0003] Research has found that in the case of using only the front axle front wheel steering of the tractor, the movement direction of the trailer is determined by the front wheel steering device and the hinged device, and its movement completely depends on the traction of the tractor. Therefore, the difficulty and key point of semi-trailer trajectory tracking control is to reduce the cost while improving the trajectory tracking ability of the tractor and the trailer. In addition, in actual application, the hinge is easily disturbed by the uncertainty caused by the change of the trailer mass center shift and load shift, which makes the trajectory tracking control of the trailer inconsistent with that of the tractor. With the change of speed and load, this feature will be more pronounced.

[0004] In the design of actual control system, it is difficult to coordinate and regulate multiple control targets with a single control variable, and too many control inputs increase the difficulty of controller design and reduce the precision of controller design. On the other hand, if only a single trajectory tracking control target is considered, the trajectory tracking control precision of the semi-trailer tractor and trailer body cannot be guaranteed. SUMMARY

[0005] To solve the above problems, the present application provides a trajectory tracking control method for intelligent semi-trailers based on two-type fuzzy logic control, which effectively solves the problem that the mass center shift and load shift of the trailer cause uncertain deviation between the driving trajectories of the rear trailer body and the front trailer body when the semi-trailer drives in a large curvature trajectory, and it is difficult to consider the tracking effect of both trailer bodies in trajectory tracking. Mainly includes semi-trailer vehicle nonlinear dynamics modeling part, trajectory tracking model building, two-type fuzzy controller main body building, and two-type fuzzy controller parameter optimization.

[0006] The vehicle nonlinear dynamics modeling part is a vehicle physical model established by some simplifications and assumptions, which will be used as a carrier for the controller.

[0007] The controller main part of the type 2 fuzzy controller is built to process the trailer tracking error, and the uncertainty interference received by the trailer tracking error is covered. The double-level uncertainty processing improves the adaptability and robustness of the fuzzy control system in complex environment. The particle swarm optimization algorithm is introduced to solve the problems of numerous parameters and complex design of the type 2 fuzzy logic controller, ensure the coverage of the uncertainty interference, and improve the controller precision. The improved type 2 fuzzy logic controller can make the semitrailer well process the trajectory tracking control under large curvature or curvature mutation working conditions when facing the centroid shift and load shift interference under different vehicle speeds and different loads.

[0008] The technical scheme of the present application is:

[0009] The nonlinear dynamics modeling part of the semitrailer vehicle: the semitrailer is composed of a tractor and a trailer connected by a hinged device, wherein the tractor usually has two axles, and the trailer has three axles, which are simplified as a single axle in the modeling process. In addition, the vehicle model also makes the following assumptions: 1. ignoring the influence of aerodynamics; 2. combining the left and right wheels of the same axle into an equivalent wheel; 3. assuming that the hinge angle is small; 4. ignoring the rolling resistance of the tire; 5. ignoring the roll and pitch of the vehicle; adopting a 7-DOF (Degree of Freedom) dynamics model as the reference model in the controller, and establishing a 7-DOF vehicle model considering the lateral, longitudinal and yaw of the tractor, the lateral, longitudinal and yaw of the trailer, and the relative rotation at the hinge angle according to the characteristics of the semitrailer, while considering the influence of the change of the vehicle longitudinal speed on the vehicle.

[0010] The partial derivative of the tractor variable is obtained by Lagrange method to meet the requirement of simplified model, and the source of uncertainty at the hinge is analyzed, and the specific Lagrange equation expression is as shown in the following formula:

[0011] The Lagrange equation expression is as follows:

[0012]

[0013] In the formula, L is the Lagrange quantity, defined as L=T-V, wherein T is the kinetic energy of the system, V is the potential energy of the system, q is the generalized speed, t is the time variable, and Q is the generalized force.

[0014] The coordinates of the trailer relative to the tractor can be expressed as:

[0015]

[0016] The generalized coordinates are selected as which is a function of time, and the kinetic energy is expressed as:

[0017]

[0018] wherein: m1 is the mass of the tractor, m2 is the mass of the trailer, x1 is the longitudinal displacement of the tractor, x2 is the longitudinal displacement of the trailer, is the longitudinal velocity of the tractor, is the longitudinal velocity of the trailer, y1 is the lateral displacement of the tractor, y2 is the lateral displacement of the trailer, is the lateral velocity of the tractor, is the lateral velocity of the trailer, and represent the yaw angle of the tractor and the trailer, respectively, I1, I2 represent the moment of inertia of the tractor and the trailer, respectively, f is the distance between the mass center of the tractor and the hinge point, wherein, the potential energy V = 0, so L = T. The present application only considers the partial derivative of the Lagrange equation L with respect to the longitudinal velocity, the lateral velocity, the yaw angular velocity of the tractor and the yaw angle of the trailer, and the Lagrange term can be obtained:

[0019]

[0020] The generalized force expression of each state is as follows:

[0021] Q x1 = F xf + F xr + F xs - F yf δ + F ys γ

[0022] Q y1 = F yf + F yr + F ys + F xf δ - F xs γ

[0023]

[0024] Q x1 , Q y1 , represent the longitudinal generalized force, the lateral generalized force, the yaw direction generalized force of the tractor and the generalized force in the yaw direction of the trailer, respectively.

[0025] From the above formula, it can be deduced that:

[0026]

[0027] Define the intermediate variable matrix:

[0028]

[0029] Therefore, the state equation is represented as follows:

[0030]

[0031]

[0032] F x,i = Ka * S

[0033] F yi = Ca * a

[0034] In the formula, a1, b1 are the distances between the front axle, rear axle and the center of mass of the tractor, a2, b2 are the distances between the front axle, rear axle and the center of mass of the trailer, g is the articulation angle, d represents the front wheel steering angle of the tractor, S is the slip rate of the wheel, a represents the side slip angle of the wheel, which can be obtained by Trucksim software. Ka and Ca are the longitudinal slip stiffness and side slip stiffness of the tire, respectively, which are proportional to the load size in the linear region. x1 V represents the longitudinal speed of the tractor, V y1 represents the lateral speed of the tractor, w1 represents the yaw angular velocity of the tractor, x represents the state vector.

[0035] Therefore, the actual longitudinal force and lateral force should be:

[0036] F xi = (Ka + AKa) * S

[0037] F yi = (Ca + ACa) * a

[0038] In the formula, the subscript x of F represents the longitudinal direction, y represents the lateral direction, i = f, r, s, respectively, represent the front, middle and rear three axles of the semi-trailer.

[0039] Build a trajectory tracking model: the trajectory tracking model establishes the relationship between the vehicle coordinate system and the global coordinate system. The state quantities such as the lateral and longitudinal speeds and the yaw angular velocity of the tractor can be obtained from the above dynamics model, so that the lateral position coordinate of the tractor can be calculated. It is worth noting that although the tractor is also subject to interference at the articulation, the influence on the lateral position coordinate Y1 of the tractor is small due to the short wheelbase of the tractor, and Y1 can still be represented as:

[0040] Y1 = ∫V y1 dt

[0041] The length of the trailer is large and is not connected to the tractor by a rigid connection, and will be affected by the uncertainty factors caused by load transfer and center of mass change related to vehicle speed and load. Therefore, the trailer cannot obtain the Y1 lateral position corresponding to the X1 position of the tractor. At this time, the actual lateral position coordinate Y2 of the trailer is:

[0042] Y2 = Y1 + Ay

[0043] Ay represents the lateral position deviation of the tractor and the trailer.

[0044] The kinematic model of trajectory tracking is represented as follows:

[0045]

[0046] where: p is the road curvature, Y r is the desired longitudinal coordinate, e y1 is the lateral error of the tractor from the trajectory, is the heading angle error of the tractor from the trajectory, e y2 is the lateral error of the trailer from the trajectory.

[0047] A two-input single-output type-2 fuzzy logic controller is built: let the input variable domain x = (x1, x2) T , there are P rules (p = 1, …, P), defined as follows:

[0048]

[0049] The above and are the antecedents of the type-2 fuzzy logic controller, , the output of the type-2 fuzzy logic controller, is set to be a Gaussian membership function, including the upper membership function and the lower membership function The clear input is fuzzified by the upper and lower membership functions to enhance its fuzzy characteristics.

[0050] where the upper membership function corresponding to each input is:

[0051]

[0052] where m = 1, 2, 3, …, M; M represents the number of controller inputs.

[0053] The lower membership function is:

[0054]

[0055] where c p,m and represent the centers of the Gaussian lower and upper membership functions, respectively, σ p,m and are the standard deviations of the Gaussian lower and upper membership functions, and h is the height of the lower membership function.

[0056] The output of the type-2 fuzzy logic controller is shown in the output layer part of Figure 2 , including the fuzzy upper output and the fuzzy lower output y(x) the fuzzy factor triggers the fuzzy rule The rule corresponding to the rule consequent output After operation, the fuzzy output is obtained, wherein the rule is triggered f p , Respectively:

[0057]

[0058] ∩ represents the algebraic product.

[0059] The present application selects a T-S two-type fuzzy logic controller, and the rule consequent is designed as a constant type rule consequent, that is Fuzzy upper layer output And fuzzy lower layer output y (x) is represented as:

[0060]

[0061] The final clear output y(x) is:

[0062]

[0063] The fuzzy rule table of the two-type fuzzy controller is established:

[0064] The two-type fuzzy logic controller takes the lateral deviation and the heading angle deviation as control input variables, and the front wheel steering angle of the towing vehicle as the control output variable. The control principle is that when the lateral deviation is negative, the vehicle position is deviated to the right in the global coordinate system, and the left direction steering angle is needed, and when the heading angle error is negative or zero, a larger left direction steering angle is needed, and the control variable should be positive large (PL) at this time. When the lateral deviation is positive, the vehicle position is deviated to the left in the global coordinate system, and the right direction steering angle is needed, and when the heading angle error is positive or zero, a larger right direction steering angle is needed, and the control variable should be negative large (NL) at this time. By analogy, the fuzzy rule table can be summarized as table 1.

[0065] Table 1 fuzzy rule table

[0066]

[0067] In the table, NL represents negative large, NS represents negative small, ZE represents zero, PS represents positive small, and PL represents positive large.

[0068] The lateral error e y1 of the towing vehicle and the heading angle error are taken as the input variables of the two-type fuzzy logic controller, and the front wheel steering angle δ of the towing vehicle is solved.

[0069] Parameter optimization of the two-type fuzzy controller:

[0070] The particle swarm optimization (PSO) algorithm is inspired by the observation of the behavior of bird flocking and fish schooling. These groups exhibit a kind of cooperative behavior when searching for food or migrating, i.e., individuals cooperate with each other and move in the most favorable direction to achieve better results. The essence of this group behavior is information sharing and cooperation between individuals, thus achieving the optimization goal as a whole.

[0071] The following are the basic steps of the PSO algorithm conceived by the present application in combination with the application object of the semi-trailer tractor:

[0072] Step 1. Initialization:

[0073] Randomly generate the parameters of the bivalent fuzzy logic controller as particles in the particle swarm, and randomly generate the position and speed of each particle, initialize the individual optimal solution (initial position) and global optimal solution of each particle.

[0074] Step 2. Iteration process: for each iteration, perform the following steps:

[0075] 1. Update the particle speed and position according to the formula;

[0076] 2. Evaluate fitness: calculate the fitness value of each particle, which is determined by the objective cost function defined by the problem;

[0077] 3. Update individual and global optimal solutions: for each particle, compare the fitness values according to the current position and individual optimal solution, and update the individual optimal solution. At the same time, compare the fitness values of all particles and update the global optimal solution.

[0078] Step 3. Termination condition:

[0079] When the predetermined number of iterations is reached or the desired state is met, terminate the algorithm.

[0080] Based on the above control method, the present application also proposes an intelligent semi-trailer trajectory tracking control system based on bivalent fuzzy control, comprising:

[0081] A semi-trailer vehicle nonlinear dynamics model part, which is the semi-trailer vehicle nonlinear dynamics model established by the above method;

[0082] A trajectory tracking model part, which is the trajectory tracking model established by the above method;

[0083] A bivalent fuzzy logic controller part, which is the bivalent fuzzy logic controller established by the above method, and the parameters are optimized according to the above optimization method. Based on the optimized parameters, the optimal front wheel steering angle of the tractor is calculated to realize trajectory tracking.

[0084] The application further provides a control device for an intelligent semitrailer, comprising a processor and a memory, wherein the memory is used for storing instruction codes, and the processor is used for reading the instructions of the memory and executing the instructions, and the processor can realize the trajectory tracking control method when executing the instructions.

[0085] The application has the following beneficial effects:

[0086] 1) The application establishes a dynamic model of the semitrailer by using the Lagrange method, which accurately simulates the state of the vehicle and builds a trajectory tracking model, and analyzes the reasons why the trajectory tracking accuracy of the intelligent commercial semitrailer is affected under the influence of uncertainty.

[0087] 2) For the semitrailer with a tractor and a trailer, it is difficult to consider the tracking accuracy of both under the influence of uncertainty. A two-type fuzzy controller is designed to cover the uncertainty interference of the trailer tracking error in addition to the tracking error of the tractor. This two-level uncertainty processing improves the adaptability and robustness of the fuzzy control system in complex environments.

[0088] 3) Considering that the two-type fuzzy controller has many parameters and is complex to design, the particle swarm optimization algorithm is used to optimize the parameters, which effectively improves the controller accuracy and significantly reduces the difficulty of controller parameter design. In the actual driving process of the intelligent commercial semitrailer, the trajectory tracking accuracy is significantly improved, and the safety of the intelligent commercial semitrailer under heavy load and high speed is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0089] Figure 1 The semitrailer dynamic model schematic diagram provided for the embodiment of the application;

[0090] Figure 2 The detailed flowchart of the semitrailer trajectory tracking control method provided for the embodiment of the application.

[0091] Figure 3 The schematic diagram of the two-type fuzzy controller provided for the embodiment of the application. DETAILED DESCRIPTION

[0092] The application will be further described below with reference to the accompanying drawings.

[0093] As Figure 1 shown, the dynamic model of the semitrailer is established

[0094] Without considering the roll, pitch and vertical motion of the vehicle, the vehicle only has motion in the x-o-y plane. The vehicle is a tractor with front wheel steering, and the vehicle body coordinate system is in the vehicle left-right symmetry plane. The origin of the vehicle mass center is o, the x-axis is the vehicle longitudinal axis, the positive direction is the vehicle head direction, the z-axis positive direction is perpendicular to upward, and the y-axis points to the vehicle side direction, and the positive direction satisfies the right-hand rule.

[0095] The Lagrange equation is expressed as follows:

[0096]

[0097] wherein, L is the Lagrange quantity, defined as L=T-V, wherein T is the kinetic energy of the system, V is the potential energy of the system, q is the generalized velocity, and Q is the generalized force.

[0098] The coordinates of the trailer relative to the tractor can be expressed as:

[0099]

[0100] Select as the generalized coordinate system, the generalized coordinate is a function of time, and the kinetic energy is expressed as:

[0101]

[0102] wherein m1 is the mass of the tractor, m2 is the mass of the trailer, is the longitudinal speed of the tractor, is the longitudinal speed of the trailer, is the lateral speed of the tractor, is the lateral speed of the trailer, and γ is the hinge angle, and represent the yaw angle of the tractor and the trailer, respectively, I1 and I2 represent the moment of inertia of the tractor and the trailer, respectively, and f is the distance between the mass center of the tractor and the hinge point, wherein the potential energy V=0, so L=T. The present application only considers the partial derivative of the Lagrange equation L with respect to the longitudinal and lateral speeds of the tractor, the yaw angle of the tractor, and the yaw angle of the trailer, and the Lagrange term can be obtained as:

[0103]

[0104] The generalized force expression of each state is as follows:

[0105] Q x =F xf +F xr +F xs -F yf δ+F ys γ

[0106] Q y1 =F yf +F yr +F ys +F xf δ-F xs γ

[0107]

[0108] From the above formula, it can be deduced that

[0109]

[0110] F x,i = Ka * S

[0111] F yi = Ca * a

[0112] In the formula, a1 and b1 are the distances between the front and rear axles of the tractor and the center of mass, a2 and b2 are the distances between the front and rear axles of the trailer and the center of mass, S is the slip rate of the wheel, a is the side slip angle of the wheel, Ka and Ca are the longitudinal slip stiffness and side slip stiffness of the tire, respectively, which are directly proportional to the load size in the linear region.

[0113] Therefore, the actual longitudinal force and lateral force should be:

[0114] F xi = (Ka + AKa) * S

[0115] F yi = (Ca + ACa) * a

[0116] The trajectory tracking model establishes the relationship between the vehicle coordinate system and the global coordinate system. The state quantities such as the lateral and longitudinal speeds and the yaw angular velocity of the tractor can be obtained from the above dynamics model, so that the lateral coordinate position of the tractor can be calculated. It is worth noting that although the tractor is also subject to interference at the hinge, the influence on the lateral position coordinate of the tractor is smaller due to the shorter wheelbase of the tractor, and the lateral position coordinate of the tractor can still be represented as:

[0117] Y1 = ∫V y1 dt

[0118] At this time, the actual lateral position of the trailer is:

[0119] Y2 = Y1 + Ay

[0120] The kinematic model of trajectory tracking is represented as follows:

[0121]

[0122] In the formula: p is the road curvature, Y r is the desired longitudinal coordinate, is the tractor yaw angle, e y1 is the lateral error of the tractor and the trajectory, is the tractor yaw angle deviation, e y2 is the lateral error of the trailer and the trajectory.

[0123] As shown in Figure 2 , 3 , Figure 2is the overall flowchart of the implementation of the present application, Figure 3 is a schematic diagram of a two-type fuzzy controller, the present application takes the deviation between the planned path trajectory and the current vehicle position as the input of the controller, and optimizes the parameters of the controller through a particle swarm algorithm. Finally, the optimized two-type fuzzy controller is used for trajectory tracking control.

[0124] The specific process is as follows:

[0125] The upper membership degree of the two-type fuzzy controller is as follows:

[0126]

[0127] The lower membership degree is as follows:

[0128]

[0129] wherein c p,m and represent the center of the Gaussian type lower and upper membership functions, σ p,m and are the standard deviations of the Gaussian type lower and upper membership functions, and h is the height of the lower membership function.

[0130]

[0131] ∩ represents the algebraic product.

[0132] The fuzzy upper output and the fuzzy lower output y (x) are as follows:

[0133]

[0134] The final defuzzified output y(x) is as follows:

[0135]

[0136] Design of pso optimization algorithm

[0137] Randomly generate the parameters of the two-type fuzzy controller as particles in the particle swarm, and randomly set the initial position and velocity for each particle. In addition, initialize the individual optimal solution (initial position) and the group optimal solution of each particle.

[0138] In each iteration, the following operations are performed:

[0139] Update the velocity and position of the particle;

[0140] Evaluate fitness: Calculate the fitness value of each particle, which is equal to the value of the defined objective cost function;

[0141] After the controller is built, the particle swarm algorithm is used to optimize the parameters of the controller to achieve the trajectory tracking accuracy of the semitrailer under the influence of uncertain factors, including the following:

[0142] Establish the cost function

[0143] The cost function of the optimization algorithm is:

[0144] J=k1J1 2 +k2J2 2 +k3J3 2

[0145] k1, k2 and k3 represent the weighting coefficients.

[0146] The above formula considers the lateral displacement deviation and the heading angle deviation of the entire vehicle body at the same time, wherein:

[0147] J1=e y =n1e y1 +n2e y2

[0148] J1 is the lateral displacement deviation e y1 of the tractor and the lateral displacement deviation e y2 of the trailer, which are added in proportion to the error sum as the lateral displacement deviation of the vehicle body, and n1 and n2 represent the weighting coefficients of the two deviations e y1 and e y2 .

[0149]

[0150] represents the yaw angle of the tractor.

[0151]

[0152] In the formula, it is set that in order to expect to obtain 85% uncertainty coverage. In the formula, x q represents the lateral displacement error of the tractor as the first lateral deviation input e y1 of the controller, x g represents the trailer as the first lateral deviation e y2 of the controller, is the fuzzy output, and y represent the upper and lower outputs of the fuzzy layer, respectively.

[0153] Updating individual and global optimal solution: for each particle, the fitness value of the current position is compared with its individual optimal solution to update the individual optimal solution. At the same time, the fitness values of all particles are compared to update the global optimal solution.

[0154] When the set number of iterations is reached or the specific stop condition is met, the algorithm is terminated.

[0155] Finally, the optimized parameters will be obtained through the particle swarm optimization algorithm c p,m and The optimized parameters are used to update the membership functions of the fuzzy controller, the membership functions of the fuzzy controller are updated as follows: σ p,m , and h are given to the bivalent fuzzy controller, and the bivalent fuzzy controller outputs the optimal front wheel steering angle of the tractor based on the optimized parameters to realize trajectory tracking.

[0156] The above detailed description is only a specific description of the feasible embodiments of the present application, and is not used 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 two-dimensional fuzzy control-based intelligent semi-trailer trajectory tracking control method, characterized in that, The application relates to a method for controlling a tractor-trailer vehicle to track a given trajectory, comprising the following steps: S1, establishing a semi-trailer vehicle nonlinear dynamics model; S2, establishing a trajectory tracking model; The trajectory tracking model in the step S2 is to establish the relationship between a vehicle coordinate system and a global coordinate system, the lateral speed, the longitudinal speed and the yaw angular speed of the tractor are obtained from the vehicle dynamics model, so that the lateral position coordinate of the tractor can be calculated: Y1 = ∫V y1 dt; The lateral position coordinate of the trailer is Y2=Y1+△y Therefore, the trajectory tracking model is expressed as follows: S3, establishing a two-type fuzzy logic controller; where: p is the road curvature, Y r is the desired longitudinal coordinate, e y1 is the tractor lateral error from the trajectory, is the tractor heading angle deviation, e y2 is the trailer lateral error from the trajectory; The upper-layer membership degree corresponding to each input is: The bivalent fuzzy logic controller in S3 is a T-S bivalent fuzzy logic controller with two inputs and one output, taking the lateral deviation e y and the heading angle deviation The front wheel steering angle δ of the towing vehicle is the control input variable, and the control output variable; the input variable domain x = (x1, x2) T There are P rules, defined as follows, where p = 1, …, P: The above And As the antecedent of the bimodal fuzzy logic controller, the application is provided with a Gaussian membership function, including an upper membership function And a lower membership function The clear input is fuzzified by the upper and lower membership functions to enhance its fuzzy characteristics; The lower-layer membership degree is: The symbol represents an algebraic product; where c p,m and represent the center of the Gaussian type lower and upper membership functions, σ p,m and are the standard deviation of the Gaussian type lower and upper membership functions, h is the height of the lower membership function; Output term of the bimodal fuzzy logic controller, comprising a fuzzy upper output and a fuzzy lower output y(x), the fuzzified factors triggering fuzzy rules the rule corresponding to the rule consequent output the fuzzy output obtained after operation, wherein the fuzzy rules are respectively: The final clear output y(x) is: The rule consequent is designed as a constant type rule consequent, i.e. Fuzzy upper layer output and fuzzy lower layer output is represented as: S4, parameter optimization of the two-type fuzzy logic controller, and outputting the optimal front wheel steering angle of the tractor based on the optimized parameters to realize trajectory tracking; The parameter optimization of the two-type fuzzy logic controller in the step S4 is realized by using a particle swarm algorithm, and comprises the following steps: Step 1. Initialization: Randomly generate the two-type fuzzy logic controller parameters as particles in the particle swarm, and randomly generate the position and speed of each particle, initialize the individual optimal solution and the global optimal solution of each particle, wherein the individual optimal solution is the initial position; Step 2. Iteration process: for each iteration, the following steps are executed: Step 2.1, update the particle speed and position according to the formula; Step 2.2, evaluate the fitness: calculate the fitness value of each particle, which is equal to the value of the defined target cost function; Step 2.3, update the individual and global optimal solutions: for each particle, compare the fitness values according to the current position and the individual optimal solution, and update the individual optimal solution; meanwhile, compare the fitness values of all particles, and update the global optimal solution; Step 3. Iteration termination: when a predetermined iteration number is reached or a stop condition is met, the iteration is terminated. The semi-trailer vehicle nonlinear dynamics model in the step S1 adopts a seven-degree-of-freedom model, and the model is realized by taking the partial derivative of the Lagrange method to the tractor variables.

2. The intelligent semi-trailer trajectory tracking control method based on two-mode fuzzy control according to claim 1, characterized in that, The specific modeling method of the nonlinear dynamics model is as follows:

3. The intelligent semi-trailer trajectory tracking control method based on two-mode fuzzy control according to claim 1 or 2, characterized in that, The established Lagrange equation expression is shown in the following formula: In the formula, L is the Lagrange quantity, defined as L=T-V, wherein T is the kinetic energy of the system, V is the potential energy of the system, q is the generalized speed, t is the time variable, and Q is the generalized force; The coordinates of the trailer relative to the tractor are shown as follows: The partial derivative of the Lagrange equation L to the longitudinal speed, the lateral speed, the yaw angle of the tractor and the yaw angle of the trailer can obtain the Lagrange term: Selecting As a generalized coordinate system, the generalized coordinates are functions of time, then the kinetic energy is expressed as: where m1is the mass of the tractor, m2is the mass of the trailer, is the longitudinal speed of the tractor, is the longitudinal speed of the trailer, is the lateral speed of the tractor, is the lateral speed of the trailer, and represent the yaw angle of the tractor and trailer, respectively, I1, I2are the moments of inertia of the tractor and trailer, respectively, and f is the distance of the center of mass of the tractor from the articulation point, where the potential energy V = 0, so L = T. The generalized force expression of each state is as follows: Define the intermediate matrix variable as follows: Q x = F xf + F xr + F xs - F yf δ + F ys γ Q y1 = F yf + F yr + F ys + F xf δ-F xs γ The longitudinal force and the lateral force are: F x,i = Ka * S F yi = Ca*α where a1, b1 are the distances between the front and rear axles and the center of mass of the tractor, a2, b2 are the distances between the front and rear axles and the center of mass of the trailer, γ is the articulation angle, S is the slip of the wheels, a is the side slip angle of the wheels, Ka and Ca are the longitudinal and lateral stiffness of the tires, respectively, which are proportional to the load in the linear region; V x1 denotes the longitudinal velocity of the tractor, V y1 denotes the lateral velocity of the tractor, ω1 denotes the yaw angular velocity of the tractor, x denotes the state vector; The subscript x of F represents the longitudinal direction, y represents the lateral direction, i=f, r, s respectively represent the front, middle and rear three axles of the semi-trailer. F xi = (Ka+△Ka)*S F yi = (Ca+△Ca)*α The rule table establishment method of the two-type fuzzy logic controller is as follows:

4. The intelligent semi-trailer trajectory tracking control method based on two-mode fuzzy control according to claim 1, characterized in that, ​ The control principle of the type-2 fuzzy logic controller is as follows: when the lateral deviation is negative, the vehicle position is right in the global coordinate system, and a left turning angle is needed; when the heading angle error is negative or zero, a larger left turning angle is needed, and the control quantity should be a positive large PL; when the lateral deviation is positive, the vehicle position is left in the global coordinate system, and a right turning angle is needed; when the heading angle error is positive or zero, a larger right turning angle is needed, and the control quantity should be a negative large NL; and the like, to establish the following fuzzy rules: When for NL, e y The values of the front wheel turning angle are NL, NL, NS, NS, and NS, respectively. When for NS, e y The value of the front wheel turning angle is NL, NL, NL, NL, and ZE, respectively, when NL, NS, ZE, PS, and PL, respectively. When for ZE, e y The values of the front wheel turning angle are NS, NS, ZE, PS, and PL, respectively. When for PS, e y The values of the front wheel turning angle are NS, ZE, PS, PS, and PL, respectively, when NL, NS, ZE, PS, and PL, respectively. When for PL, e y The values of the front wheel turning angle are PS, PS, PL, PL, and PL, respectively, when NL, NS, ZE, PS, and PL, respectively.

5. The intelligent semi-trailer trajectory tracking control method based on two-mode fuzzy control according to claim 1, characterized in that, The cost function in step 2.2 is as follows: J = k1J1 2 + k2J2 2 + k3J3 2 The lateral displacement deviation and the heading angle deviation of the entire vehicle body are considered simultaneously through the above formula, wherein: J1 = e y = n1e y1 + n2e y2 J1 is the sum of the error of the proportional addition of the lateral displacement deviation of the tractor and the lateral displacement deviation of the trailer, as the lateral displacement deviation of the vehicle body; to describe the yaw angle of a tractor; Wherein:

6. A control system for implementing the intelligent semi-trailer trajectory tracking control method based on two-mode fuzzy control according to claim 1, characterized in that, including: a semi-trailer vehicle nonlinear dynamics model part, which is the model established in S1 of claim 1; a trajectory tracking model part, which is the model established in S2 of claim 1; a type-2 fuzzy logic controller part, which is the type-2 fuzzy logic controller established in S3 of claim 1, and the parameter optimization is realized according to the method in S4 of claim 1, and the optimized parameters are used to calculate the optimal front wheel turning angle of the tractor to realize trajectory tracking.

7. A control device for an intelligent semitrailer, characterized by including a processor and a memory, wherein the memory is used to store instructions, the processor is used to read and execute the instructions of the memory, and the processor can realize the trajectory tracking control method of claim 1 when executing the instructions.

Images