Intelligent electric heavy truck fuzzy sliding mode path tracking control method and device considering crosswind interference, storage medium and automobile

Abstract

The invention discloses an electric heavy truck path tracking control method and device based on a self-adaptive interval type-2 fuzzy second-order sliding mode under the influence of crosswind, a storage medium and an automobile, and belongs to the field of intelligent traffic. Comprising the steps that 1, disturbance factors existing in the driving process of the electric heavy truck are analyzed, and an electric heavy truck dynamic model and a path tracking model are constructed; 2, estimating a course-related unknown state and lumped disturbance by using an interval type-2 fuzzy logic system; 3, selecting a proper sliding variable, establishing a second-order sliding mode kinetic equation, and describing a dynamic relationship and disturbance suppression of a system state tracking error; and 4, designing a self-adaptive interval type-2 fuzzy sliding mode controller to realize a path tracking target. The method has the advantages that 1, the method has higher robustness on parameter change and external interference; 2, the buffeting phenomenon in the control process can be reduced, and the stability and reliability of the system are improved; and 3, the dynamic response speed and the control precision of the system are improved while the stability of the system is ensured.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and specifically to a fuzzy sliding mode-based intelligent electric heavy truck path tracking control method, equipment, storage medium and automobile under the influence of side wind. Background Art

[0002] As digitalization and intelligence deeply reshape the industry ecosystem, the logistics and transportation industry is undergoing a historic transition from a traditional model to a technology-driven one. As the core vehicle of the smart logistics system, autonomous electric heavy-duty trucks, leveraging the deep integration of intelligence and electrification, are systematically addressing the structural constraints that have constrained the industry's development. Driven by the continued expansion of global trade and the upgrading of consumer patterns, demand for freight transportation has shifted from basic scale growth to a complex demand characterized by high frequency, high precision, and high reliability. However, the traditional heavy-duty truck transportation system, which relies on manual dispatching, is increasingly exposed to the triple dilemmas of efficiency lag, cost uncontrollability, and safety vulnerabilities. Regarding efficiency, intermittent manual operation and scheduling errors lead to wasted capacity, making it difficult to meet the ever-increasing transportation demand. Regarding cost, the high proportion of driver labor costs, coupled with a shortage of drivers, creates a vicious cycle of cost control. Regarding safety, fatigue risks and human errors caused by long driving hours continue to threaten transportation safety and the sustainable development of the industry. Autonomous electric heavy-duty trucks are achieving breakthroughs through full-chain technology integration: Intelligent driving systems build closed-loop capabilities for environmental perception, decision-making, and precise control, upgrading the transportation process from "strong coupling between driver and vehicle" to "autonomous intelligent execution." Electric platforms innovate powertrains and energy management mechanisms, breaking through the efficiency and environmental bottlenecks of traditional fuel-powered vehicles. This collaborative innovation not only drives a performance revolution in transportation vehicles but also, with its 24 / 7 uninterrupted operation and standardized, end-to-end control, reshapes the responsiveness and reliability of logistics networks. This provides core support for establishing an inherently safe and cost-effective modern transportation system, marking a substantial breakthrough in the logistics industry's transition from a scale-driven to a quality-driven model.

[0003] However, the actual operation of self-driving electric heavy-duty trucks still faces many severe challenges, among which crosswind interference is an issue that cannot be ignored. In the dynamic operation of self-driving electric heavy-duty trucks, crosswind disturbances are far more than ordinary environmental disturbances. Their essence is a deep game between aerodynamic characteristics, vehicle inertial mechanics, and intelligent control systems. Compared with traditional passenger cars, electric heavy-duty trucks have an ultra-high cross-section design (the vehicle height / wheelbase ratio exceeds 2.5) and an ultra-long wheelbase topology. The lateral wind-exposed area and the offset of the aerodynamic pressure center are amplified exponentially, making the disturbance effect of lateral wind pressure on the vehicle "lever-type" amplification - even moderate crosswinds (level 6-8) can trigger triple instability modes of lateral center of mass offset, heading angle oscillation, and nonlinear attenuation of tire cornering stiffness. On highways, this aerodynamic-mechanical coupling can cause a "creeping" of the vehicle's trajectory. A deviation of 0.5 meters in just three seconds can trigger a chain reaction of lane-crossing intrusions. On winding mountain roads, the combined effects of canyon effects and transient crosswinds compress the critical safety distance between the vehicle and the guardrail to the centimeter level, creating a risk of loss of control due to the "airflow scissors gap." This reveals the system-level vulnerability of autonomous electric heavy-duty trucks in dealing with complex aerodynamic disturbances, which urgently requires innovative control technologies to address.

[0004] Traditional path tracking control methods have obvious limitations when dealing with complex situations such as crosswind interference. Some control algorithms based on fixed mathematical models can achieve good control effects under ideal conditions. However, in actual applications, they fail to fully consider the dynamic changes of uncertain factors such as crosswind, resulting in a decrease in control accuracy and an inability to ensure that the vehicle can always stably track the preset path. Other control schemes that rely on high-precision sensor feedback will have serious impacts on sensor performance in severe weather conditions such as heavy rain and dense fog, making it impossible to accurately obtain the vehicle's driving status and surrounding environment information, thereby affecting the effectiveness and reliability of the entire control system. Therefore, developing a control method that can effectively overcome crosswind interference and ensure that self-driving electric heavy-duty trucks can stably and accurately track the preset path in various complex environments has become a key issue that needs to be urgently addressed in the current field of intelligent transportation and autonomous driving, and is of great significance to promoting the intelligent development of the logistics and transportation industry. Summary of the Invention

[0005] In order to solve the path tracking problem of existing electric heavy-duty trucks in complex real-world driving scenarios, the present invention studies the path tracking control of electric heavy-duty trucks from three aspects: vehicle and path tracking modeling, estimation of lumped disturbance including sidewind disturbance using interval type-2 fuzzy logic system, and design of adaptive interval type-2 fuzzy sliding mode controller. A path tracking control method for electric heavy-duty trucks based on adaptive interval type-2 fuzzy second-order sliding mode is proposed.

[0006] A path tracking control method for an electric heavy truck based on an adaptive interval type-2 fuzzy second-order sliding mode includes the following steps:

[0007] S1: Establish a five-degree-of-freedom dynamic model of an electric heavy truck under the influence of crosswinds and construct a path tracking model including disturbances;

[0008] S2: Based on the coordinate transformation method, the path tracking model is converted into a state equation in a strict feedback form that is convenient for controller design;

[0009] S3: Design fuzzy rules, fuzzy reasoning, type reduction methods and defuzzification methods in interval type-2 fuzzy logic systems;

[0010] S4: Using interval type-2 fuzzy logic system to estimate the unknown aggregate disturbance in the system including crosswind interference;

[0011] S5: Design a second-order sliding mode controller based on adaptive interval type-2 fuzzy; define and design the adaptive law parameters in the adaptive interval type-2 fuzzy sliding mode controller;

[0012] S6: Perform inverse transformation on the adaptive interval type-2 fuzzy sliding mode controller to obtain the system control input, namely the front wheel steering angle of the electric heavy truck.

[0013] This paper introduces wheel rotational freedom into the classic three-degree-of-freedom vehicle model (i.e., considering the vehicle's longitudinal, lateral, and yaw motion) for electric heavy-duty trucks, and establishes a five-degree-of-freedom vehicle model. On this basis, crosswind interference is coupled into the vehicle state equation, while the load transfer between wheels and vehicle aerodynamics are neglected. The vehicle force and torque balance equation can be obtained according to Newton's second law:

[0014]

[0015] Where, v x ,v y and ω are the longitudinal velocity, lateral velocity, and yaw rate, respectively; m is the mass of the vehicle; δ f is the front wheel turning angle; F lf , F lr are the longitudinal forces of the front and rear wheels respectively; F sf , F sr are the lateral forces of the front and rear wheels respectively; F wind Indicates the side wind force; l f , l r are the distances from the front and rear axles to the center of mass respectively; I z is the moment of inertia of the vehicle around the z-axis; M wind is the yaw moment caused by the side wind. Based on the principles of aerodynamics, the side wind force and yaw moment acting on the vehicle are established as:

[0016]

[0017] Where ρ is the air density, C d is the lateral aerodynamic drag coefficient, A is the lateral projection area of ​​the vehicle, V ω is the wind speed, β is the wind yaw angle, ψ is the vehicle heading angle, h cg is the vehicle center of mass height.

[0018] The wheel rotation freedom is introduced into the vehicle model, and the tire dynamics model is shown in the attached figure. Figure 1 As shown, we can get

[0019]

[0020] Where, ω f ,ω r are the front and rear tire angular velocities respectively; R e is the wheel rotation radius; J f , J r are the front and rear wheel moments of inertia respectively; T f , T r are the front and rear wheel torques respectively.

[0021] Combining the above equations, we can obtain the five-degree-of-freedom vehicle dynamics model under the influence of crosswind:

[0022]

[0023] Assume that the state variables are longitudinal velocity, lateral velocity, yaw angular velocity, and front and rear wheel angular velocity, that is, x = [v x ,v y ,ω,ω f ,ω r ] T ; The control input is the front wheel angle and the vehicle torque, that is, u=[δ f ,T] T ; The control output is the longitudinal velocity, lateral velocity, and yaw angular velocity, that is, y = [v x ,v y ,ω] T The vehicle dynamics model can be written in the form of the following state equation:

[0024]

[0025] Where f is a smooth function; G is a constant matrix, G = [I 3×3 ,0 3×2 ].

[0026] Considering that electric heavy trucks are affected by disturbances in actual driving scenarios, especially those caused by crosswinds, a path tracking model including disturbances is established. The specific form is as follows:

[0027]

[0028] Where: L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path. It's L os The first derivative of θ os Indicates the heading deviation between the actual driving path of the electric heavy truck and the reference path. is θ os The first derivative of , σ represents the direction coefficient, and is defined as negative clockwise, v x represents the longitudinal speed of the electric heavy truck, L represents the wheelbase of the electric heavy truck, δ f represents the front wheel steering angle of the electric heavy truck, R represents the curvature radius of the reference path, and D(t) represents the aggregate disturbance of the external interference including the influence of the side wind;

[0029] To facilitate reference path design, the curvature radius R of the reference path is converted to curvature c0, and it is assumed that the electric heavy truck travels forward along the reference path in a clockwise direction, that is, the direction coefficient σ = -1. To facilitate the processing of the unknown system state x2, the coordinate transformation method is used to further express the system equation as a state equation in a strict feedback form. The specific form is as follows:

[0030]

[0031] Where: x1 = L os , x2=v x sinθ os , x1 and x2 represent the system status, and Represent the first-order derivatives of the system states x1 and x2, respectively, u = tanδ f represents the input of the controller, represents the unknown lumped disturbance including the crosswind effect, and D1(t) represents the disturbance acting on the controller u.

[0032] The single-point fuzzification method is selected, and two fuzzy sets are defined, corresponding to the system state variables x1 and x2 respectively. The membership functions μ1 and μ2 are designed as Gaussian functions to quantify the membership of each state variable in its respective fuzzy set. The Gaussian membership function is designed in the following form:

[0033]

[0034] Where: μ1 and μ2 represent Gaussian membership functions, and is the standard form of the Gaussian function, corresponding to the adjusted values ​​of x1 and x2, respectively. i is an index variable used to represent different fuzzy sets, with values ​​from 1 to 5, which means that there are five different Gaussian membership functions for the state variables x1 and x2, respectively. Each function is adjusted by To change its position in the state space, each function corresponds to a different fuzzy set; x1 and x2 represent the system state, x1 = L os , x2=v x sinθ os , where L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path, θ os Indicates the heading deviation between the actual driving path of the electric heavy truck and the reference path;

[0035] These membership functions will generate 5×1 vectors respectively, representing the membership of each state variable on the five fuzzy sets;

[0036] Using the product reasoning mechanism, KM reduction algorithm and centroid defuzzification method, the fuzzy basis function is obtained, which is as follows:

[0037] Ψ(x)=(ξ l (x)+ξ r (x)) / 2

[0038] Where: represents the left interval vector, represents the right interval vector, Represents the i-th element of the left interval vector, whose value is f l i With all f l i The ratio of the sum of (i from 1 to N), Represents the i-th element of the right interval vector, whose value is With all The ratio of the sum of (i from 1 to N), f l i It represents the matching degree between the input data point and the antecedent of the left interval fuzzy rule, which is calculated by the KM reduction algorithm and is finally used to generate the left interval vector ξ l The normalized weight of (x), It represents the matching degree between the input data point and the antecedent of the right interval fuzzy rule, which is also calculated by the KM reduction algorithm and is finally used to generate the right interval vector ξ r (x) is the normalized weight, N represents the dimension of the vector, and T represents the transpose operation;

[0039] The output of the interval type-2 fuzzy logic system is obtained in the following form:

[0040] y(x)=θ T ψ(x)

[0041] Where: θ T represents the adaptive parameter, and Ψ(x) represents the fuzzy basis function.

[0042] In the strict feedback form of the state equation, p(t,x) and q(t,x) satisfy the following conditions, which are in the form of:

[0043]

[0044] Where: p(t,x) and q(t,x) are auxiliary functions introduced to simplify the complex solution process. They are transformations of the original function form. However, due to various uncertainties, these two functions have certain differences from the actual system. and q (t,x) are both greater than zero, represents the upper bound function of p(t,x), q (t,x) represents the lower bound function of q(t,x);

[0045] Using the output of interval type-2 fuzzy logic system, and q (t,x) is estimated, the specific form is:

[0046]

[0047] In the formula: In the formula: and θ q Represents the adaptive weight coefficient vector to be adjusted in fuzzy logic, which is dynamically updated through online learning and is used to approximate the unknown function terms in system dynamics. and ψ q (x) represents the fuzzy basis function, which is the weighted rule activation strength generated by the input variable X through the membership function, and T represents the transposition operation. and Respectively and q (t,x) is a function dynamically approximated using a fuzzy logic system.

[0048] Design a second-order sliding mode controller u based on adaptive interval type-2 fuzzy, the specific form is:

[0049]

[0050] Where: β, ω are normal parameters of the system, sign represents the sign function, s1, s2 are sliding variables, And satisfy ω≥|ζ|, ζ is the system parameter, v represents the virtual control input used to compensate for uncertainty, and its specific form is:

[0051]

[0052] Where: and θ q represents the adaptive law parameters, and Respectively and q A function (t,x) dynamically approximated using a fuzzy logic system.

[0053] Define and design the adaptive law parameters, and the adaptive interval type-2 fuzzy sliding mode controller and Depends on and θ q , the specific form is:

[0054]

[0055] Where: and Respectively and θ q The first-order derivative of , γ1 and γ2 represent positive constants, β is the control gain, and ψ q (x) represents the fuzzy basis function.

[0056] By setting the controller u = tanδ f Perform the inverse transformation to obtain the front wheel steering angle δ of the electric heavy truck f , the specific form is:

[0057]

[0058] Where: v represents a quantity in the adaptive second-order sliding mode controller that is a combination of fuzzy logic system-related functions, specific constants, and sign functions. It is used to comprehensively adjust the controller behavior to adapt to system dynamics, handle uncertainties, and achieve the desired control performance. represents the dynamic approximation using interval type-2 fuzzy logic system q (t, x), ω is the normal parameter of the system, s1 and s2 are sliding variables; by adjusting the front wheel steering angle δ f Control, the final lateral deviation L os and heading deviation θ os will converge to zero.

[0059] Beneficial effects of the present invention:

[0060] (1) The present invention integrates the advantages of fuzzy logic and sliding mode control through an interval type-II fuzzy sliding mode controller, significantly improving the system's ability to resist uncertain interference such as dynamic changes in vehicle parameters and external wind disturbances.

[0061] (2) A sliding surface is constructed based on the path tracking deviation signal, achieving high-precision disturbance suppression without the need for multi-sensor redundant configuration, significantly reducing hardware cost and system complexity.

[0062] (3) Through the dynamic response-accuracy collaborative optimization design under stability constraints, the inherent contradiction between response speed and tracking accuracy in traditional control methods is broken through, and a leap in comprehensive performance is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the tire dynamics model of the present invention;

[0064] Figure 2 Schematic diagram of the control principle of the electric heavy truck path tracking control system of the present invention;

[0065] Figure 3 This is a schematic diagram of the path tracking of an electric heavy truck;

[0066] Figure 4 It is the block diagram of interval type-2 fuzzy logic system;

[0067] Figure 5 Schematic diagram of interval type-2 fuzzy membership function. DETAILED DESCRIPTION

[0068] The present invention will be further described below with reference to the accompanying drawings.

[0069] like Figure 2 As shown in FIG, a path tracking control method for an electric heavy truck based on an interval type-2 fuzzy second-order sliding mode is described. The specific process of implementing the method is as follows:

[0070] S1: Establish a five-degree-of-freedom dynamic model of an electric heavy truck under the influence of crosswinds and construct a path tracking model including disturbances;

[0071] S1.1. Establishing the dynamic model and nonlinear tire model of electric heavy trucks

[0072] For electric heavy-duty trucks, this paper introduces wheel rotational freedom on the basis of the classic three-degree-of-freedom vehicle model (i.e., considering the vehicle's longitudinal, lateral, and yaw motion), couples crosswind interference into the vehicle state equation, and establishes a five-degree-of-freedom vehicle model that takes into account the influence of crosswind. On this basis, ignoring the load transfer between wheels and vehicle aerodynamics, the vehicle force and torque balance equation can be obtained according to Newton's second law:

[0073]

[0074] Where, v x ,v yand ω are the longitudinal velocity, lateral velocity, and yaw rate, respectively; m is the mass of the vehicle; δ f is the front wheel turning angle; F lf , F lr are the longitudinal forces of the front and rear wheels respectively; F sf , F sr are the lateral forces of the front and rear wheels respectively; F wind Indicates the lateral wind force; l f , l r are the distances from the front and rear axles to the center of mass respectively; I z is the moment of inertia of the vehicle around the z-axis; M wind is the yaw moment caused by the side wind. Based on the principles of aerodynamics, the side wind force and yaw moment acting on the vehicle are established as:

[0075]

[0076] Where ρ is the air density, C d is the lateral aerodynamic drag coefficient, A is the lateral projection area of ​​the vehicle, V ω is the wind speed, β is the wind yaw angle, ψ is the vehicle heading angle, h cg is the vehicle center of mass height.

[0077] The wheel rotation freedom is introduced into the vehicle model, and the tire dynamics model is shown in the attached figure. Figure 1 As shown, we can get

[0078]

[0079] Where, ω f ,ω r are the front and rear tire angular velocities respectively; R e is the wheel rotation radius; J f , J r are the front and rear wheel moments of inertia respectively; T f , T r are the front and rear wheel torques respectively.

[0080] Combining the above equations, we can obtain the five-degree-of-freedom vehicle dynamics model under the influence of crosswind:

[0081]

[0082] Assume that the state variables are longitudinal velocity, lateral velocity, yaw angular velocity, and front and rear wheel angular velocity, that is, x = [v x ,v y ,ω,ω f ,ω r ] T ; The control input is the front wheel angle and the vehicle torque, that is, u=[δ f ,T] T; The control output is the longitudinal velocity, lateral velocity, and yaw angular velocity, that is, y = [v x ,v y ,ω] T The vehicle dynamics model can be written in the form of the following state equation:

[0083]

[0084] Where f is a smooth function; G is a constant matrix, G = [I 3×3 ,0 3×2 ].

[0085] S1.2. Establishing a path tracking model

[0086] like Figure 3 As shown in the figure, considering that electric heavy trucks are affected by interference in actual driving scenarios, especially the interference under the influence of crosswind, a path tracking model including disturbance is established. The specific form is as follows:

[0087]

[0088] Where: L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path. It's L os The first derivative of θ os Indicates the heading deviation between the actual driving path of the electric heavy truck and the reference path. is θ os The first derivative of , σ represents the direction coefficient, and is defined as negative clockwise, v x represents the longitudinal speed of the vehicle, L represents the wheelbase of the vehicle, δ f represents the steering angle of the front wheels of the vehicle, R represents the curvature radius of the reference path, and D(t) represents the lumped disturbance of the external interference.

[0089] S2: Based on the coordinate transformation method, the path tracking model is converted into a state equation in a strict feedback form that is convenient for controller design.

[0090] To facilitate reference path design, the curvature radius R of the reference path is converted to curvature c0, and it is assumed that the electric heavy truck travels forward along the reference path in a clockwise direction, that is, the direction coefficient σ = -1. To facilitate the processing of the unknown system state x2, the system equation is further expressed using the coordinate transformation method to obtain the state equation in strict feedback form, which is as follows:

[0091]

[0092] Where: x1 = L os , x2=v x sinθ os, x1 and x2 represent the system status, and Represent the first-order derivatives of the system states x1 and x2, respectively, u=tanδ f represents the input of the controller, represents the unknown lumped disturbance including the crosswind disturbance, and D1(t) represents the disturbance acting on the controller u.

[0093] S3: Interval type-2 fuzzy logic system block diagram Figure 4 As shown, the fuzzy rules, fuzzy reasoning, type reduction method and defuzzification method in interval type-2 fuzzy logic system are designed;

[0094] S3.1. Select the single-point fuzzification method and define two fuzzy sets, corresponding to the system state variables x1 and x2, respectively. The membership functions μ1 and μ2 are designed as Gaussian functions to quantify the membership of each state variable in its respective fuzzy set. The Gaussian membership functions are designed as follows:

[0095]

[0096] Where: μ1 and μ2 represent Gaussian membership functions, and is the standard form of the Gaussian function, corresponding to the adjusted values ​​of x1 and x2, respectively. i is an index variable used to represent different fuzzy sets, with values ​​from 1 to 5, which means that there are five different Gaussian membership functions for the state variables x1 and x2, respectively. Each function is adjusted by To change its position in the state space, each function corresponds to a different fuzzy set; x1 and x2 represent the system state, x1 = L os , x2=v x sinθ os , where L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path, θ os Represents the heading deviation between the actual driving path of the electric heavy truck and the reference path; these membership functions will generate 5×1 vectors respectively, representing the membership of each state variable on the five fuzzy sets;

[0097] S3.2. Using the product reasoning mechanism, KM reduction algorithm and centroid defuzzification method, we can get the fuzzy basis function and the interval type II fuzzy membership function as follows: Figure 5 As shown, the specific form of the fuzzy basis function is as follows:

[0098] Ψ(x)=(ξ l (x)+ξ r (x)) / 2

[0099] Where: represents the left interval vector, represents the right interval vector, Represents the i-th element of the left interval vector, whose value is f l i With all f l i The ratio of the sum of (i from 1 to N), The i-th element of the right interval vector is With all The ratio of the sum of (i from 1 to N), f l i It represents the matching degree between the input data point and the antecedent of the left interval fuzzy rule, which is calculated by the KM reduction algorithm and is finally used to generate the left interval vector ξ l The normalized weight of (x), It represents the matching degree between the input data point and the antecedent of the right interval fuzzy rule, which is also calculated by the KM reduction algorithm and is finally used to generate the right interval vector ξ r (x) is the normalized weight, N represents the dimension of the vector, and T represents the transpose operation;

[0100] The output of the interval type-2 fuzzy logic system is obtained in the following form:

[0101] y(x)=θ T Ψ(x)

[0102] Where: θ T represents the adaptive parameter, and Ψ(x) represents the fuzzy basis function.

[0103] S4: Using interval type-2 fuzzy logic system to estimate the unknown lumped disturbance of the system;

[0104] S4.1. In the strict feedback form of the state equation, p(t, x) and q(t, x) satisfy the following conditions:

[0105]

[0106] Where: p(t,x) and q(t,x) are auxiliary functions introduced to simplify the complex solution process. They are transformations of the original function form. However, due to various uncertainties, these two functions have certain differences from the actual system. and q (t,x) are both greater than zero, represents the upper bound function of p(t,x), q (t,x) represents the lower bound function of q(t,x);

[0107] S4.2. Using the output of the interval type-2 fuzzy logic system in S3.2, and q (t,x) is estimated, the specific form is:

[0108]

[0109] Where: and θ q Represents the adaptive weight coefficient vector to be adjusted in fuzzy logic, which is dynamically updated through online learning and is used to approximate the unknown function terms in system dynamics. and Ψ q (x) represents the fuzzy basis function, which is the weighted rule activation strength generated by the input variable X through the membership function, and T represents the transposition operation. and Respectively and q (t,x) is a function dynamically approximated using a fuzzy logic system.

[0110] S5: Design a second-order sliding mode controller u based on adaptive interval type-2 fuzzy. The specific form is:

[0111]

[0112] Where β and ω are positive constant parameters of the system, and they satisfy ω ≥ |ζ|. v represents the virtual control input used to compensate for uncertainty. Its specific form is:

[0113]

[0114] Where: and θ q represents the adaptive law parameters, and Respectively and q A function (t,x) dynamically approximated using a fuzzy logic system.

[0115] S7: Perform inverse transformation on the adaptive interval type-2 fuzzy sliding mode controller to obtain the system control input, i.e., the front wheel steering angle of the electric heavy truck; by setting the controller u = tanδ f Perform the inverse transformation to obtain the front wheel steering angle δ of the electric heavy truck f , the specific form is:

[0116]

[0117] Where: v represents a quantity in the adaptive second-order sliding mode controller that is a combination of fuzzy logic system-related functions, specific constants, and sign functions. It is used to comprehensively adjust the controller behavior to adapt to system dynamics, handle uncertainties, and achieve the desired control performance. represents the dynamic approximation using interval type-2 fuzzy logic system q (t, x), ω is the normal parameter of the system, s1 and s2 are sliding variables; by adjusting the front wheel steering angle δ f Control, the final lateral deviation L os and heading deviation θ os will converge to zero.

[0118] Based on the above control method, the present invention also includes the following embodiments:

[0119] A control device, such as an on-vehicle ECU, includes a memory and a processor. The memory stores program code, and the processor can execute the above-mentioned control method when reading and executing the program code in the memory.

[0120] A storage medium, such as a CD, a USB flash drive, etc., stores an executable program in the storage medium, and the above-mentioned control method can be implemented when the program is read and executed.

[0121] An intelligent electric heavy-duty truck adopts the above-mentioned control method to prevent crosswind interference.

[0122] The series of detailed descriptions listed above are only specific descriptions of feasible implementation methods of the present invention. They are not intended to limit the scope of protection of the present invention. Any equivalent methods or changes that do not deviate from the technology of the present invention should be included in the scope of protection of the present invention.

Claims

1. A fuzzy sliding mode path tracking control method for intelligent electric heavy trucks considering side wind interference, characterized in that: The steps include: S1: Establish a five-degree-of-freedom dynamic model of an electric heavy truck under the influence of crosswinds and construct a path tracking model including disturbances; S2: Based on the coordinate transformation method, the path tracking model is converted into a state equation in a strict feedback form that is convenient for controller design; S3: Design fuzzy rules, fuzzy reasoning, type reduction methods and defuzzification methods in interval type-2 fuzzy logic systems; S4: Estimation of unknown aggregate disturbance including crosswind disturbance using interval type-2 fuzzy logic system; S5: Design a second-order sliding mode controller based on adaptive interval type-2 fuzzy and define and design the adaptive law parameters; S6: Perform inverse transformation on the adaptive interval type-2 fuzzy sliding mode controller to obtain the system control input, i.e., the front wheel steering angle of the electric heavy truck.

2. The path tracking control method according to claim 1, wherein: The specific method of S1 is as follows: S1.

1. Establish a dynamic model and nonlinear tire model for an electric heavy-duty truck. Based on the classic three-degree-of-freedom vehicle model that considers the longitudinal, lateral, and yaw directions of the vehicle, the wheel rotation degree of freedom is introduced to establish a five-degree-of-freedom vehicle model. The crosswind interference is coupled into the vehicle state equation, and the load transfer between wheels and vehicle aerodynamics are ignored. According to Newton's second law, the vehicle force and torque balance equation can be obtained: Where, v x ,v x and ω are the longitudinal velocity, lateral velocity, and yaw rate, respectively; m is the mass of the vehicle; δ is the front wheel turning angle; F lf , F lr are the longitudinal forces of the front and rear wheels respectively; F sf , F sr are the lateral forces of the front and rear wheels respectively; F wind Indicates the lateral wind force; l f , l r are the distances from the front and rear axles to the center of mass respectively; I z is the moment of inertia of the vehicle around the z-axis; M wind is the yaw moment caused by the crosswind; Among them, based on the principle of aerodynamics, the lateral wind force F acting on the vehicle is established wind and yaw moment M wind for: Where ρ is the air density, C d is the lateral aerodynamic drag coefficient, A is the lateral projection area of ​​the vehicle, V ω is the wind speed, β is the wind yaw angle, ψ is the vehicle heading angle, h cg is the height of the vehicle's center of mass; Introducing wheel rotation freedom into the vehicle model, we get: Where, ω f ,ω r are the front and rear tire angular velocities respectively; R e is the wheel rotation radius; J f , J r are the front and rear wheel moments of inertia respectively; T f , T r are the front and rear wheel torques respectively; Combining the above equations, we can obtain the five-degree-of-freedom vehicle dynamics model under the influence of sidewind: Assume that the state variables are longitudinal velocity, lateral velocity, yaw angular velocity, and front and rear wheel angular velocity, that is, x = [v x ,v y ,ω,ω f ,ω r ] T ; The control input is the front wheel angle and the vehicle torque, that is, u=[δ f ,T] T ; The control output is the longitudinal velocity, lateral velocity, and yaw angular velocity, that is, y = [v x ,v y ,ω] T ; The vehicle dynamics model can be written in the form of the following state equation: Where f is a smooth function; G is a constant matrix, G = [I 3×3 ,0 3×2 ]; S1.

2. Establishing a path tracking model A path tracking model including crosswind disturbance is established, and its specific form is as follows: Where: L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path. It's L os The first derivative of θ os Indicates the heading deviation between the actual driving path of the electric heavy truck and the reference path. is θ os The first derivative of , σ represents the direction coefficient, and is defined as negative clockwise, v x represents the longitudinal speed of the vehicle, L represents the wheelbase of the vehicle, δ f represents the steering angle of the front wheels of the vehicle, R represents the curvature radius of the reference path, and D(t) represents the lumped disturbance of the external interference.

3. The path tracking control method according to claim 2, wherein: The S2 is specifically as follows: The curvature radius R of the reference path is converted to curvature c0, and it is assumed that the electric heavy truck moves forward along the reference path in a clockwise direction, that is, the direction coefficient σ = -1; in order to facilitate the processing of the unknown system state x2, the coordinate transformation method is used to express the system equation as the following strict feedback form state equation: Where: x1 and x2 represent the system state, x1=L os , x2=v x sinθ os , where L os Indicates the lateral deviation between the actual driving path of the electric heavy truck and the reference path, θ os Indicates the heading deviation between the actual driving path of the electric heavy truck and the reference path; and Represent the first-order derivatives of the system states x1 and x2, respectively, u=tanδ f represents the input of the controller, represents the unknown lumped disturbance including the crosswind disturbance, and D1(t) represents the disturbance acting on the controller.

4. The path tracking control method according to claim 3, wherein: The S3 is specifically as follows: S3.

1. Select the single-point fuzzification method and define two fuzzy sets, corresponding to the system state variables x1 and x2, respectively. The membership functions μ1 and μ2 are designed as Gaussian functions to quantify the membership of each state variable in its respective fuzzy set. The specific form of the Gaussian membership function is as follows: Where: μ1 and μ2 represent Gaussian membership functions, and is the standard form of the Gaussian function, corresponding to the adjusted values ​​of x1 and x2, respectively. i is an index variable used to represent different fuzzy sets, with values ​​from 1 to 5, which means that there are five different Gaussian membership functions for the state variables x1 and x2, respectively. Each function is adjusted by To change its position in the state space, each function corresponds to a different fuzzy set. These membership functions will generate 5×1 vectors respectively, indicating the membership of each state variable on the five fuzzy sets. S3.

2. Using the product reasoning mechanism, KM reduction algorithm, and centroid defuzzification method, we can obtain the fuzzy basis function. The specific form of the fuzzy basis function is as follows: Ψ(x)=(ξ l (x)+ξ r (x)) / 2 Where: represents the left interval vector, represents the right interval vector, Represents the i-th element of the left interval vector, whose value is f l i With all f i l The ratio of the sum of (i from 1 to N), Represents the i-th element of the right interval vector, whose value is With all The ratio of the sum of (i from 1 to N), f l i It represents the matching degree between the input data point and the antecedent of the left interval fuzzy rule, which is calculated by the KM reduction algorithm and is finally used to generate the left interval vector ξ l The normalized weight of (x), It represents the matching degree between the input data point and the antecedent of the right interval fuzzy rule, which is also calculated by the KM reduction algorithm and is finally used to generate the right interval vector ξ r (x) is the normalized weight, N represents the dimension of the vector, and T represents the transpose operation; The output of the interval type-2 fuzzy logic system is obtained in the following form: y(x)=θ T Ψ(x) Where: θ T represents the adaptive parameter, and Ψ(x) represents the fuzzy basis function.

5. The path tracking control method according to claim 4, wherein: The S4 is specifically as follows: S4.

1. In the strict feedback form of the state equation, p(t, x) and q(t, x) satisfy the following conditions: Where: p(t,x) and q(t,x) are auxiliary functions introduced to simplify the complex problem solving process. represents the upper bound function of p(t,x), q (t,x) represents the lower bound function of q(t,x), and q (t,x) are both greater than zero; S4.

2. Using the output of the interval type-2 fuzzy logic system in S3.2, and q (t,x) is estimated, the specific form is: Where: and θ q Represents the adaptive weight coefficient vector to be adjusted in fuzzy logic, which is dynamically updated through online learning and is used to approximate the unknown function terms in system dynamics. and Ψ q (x) represents the fuzzy basis function, which is the weighted rule activation strength generated by the input variable X through the membership function, and T represents the transposition operation. and Respectively and q (t,x) is a function dynamically approximated using a fuzzy logic system.

6. The path tracking control method according to claim 5, characterized in that: The second-order sliding mode controller u of S5 is specifically designed as follows: Where: β, ω are normal parameters of the system, sign represents the sign function, s1, s2 are sliding variables, And satisfy ω≥|ζ|, v represents the virtual control input used to compensate for uncertainty, and its specific form is: Where: and θ q represents the adaptive law parameters, and Respectively and q(t,x) are functions dynamically approximated using a fuzzy logic system.

7. The path tracking control method according to claim 6, wherein: The S6 is specifically as follows: By setting the controller u = tanδ f Perform the inverse transformation to obtain the front wheel steering angle δ of the electric heavy truck f , the specific form is: By adjusting the front wheel steering angle δ f Control, the final lateral deviation L os and heading deviation θ os will converge to zero.

8. A control device, characterized in that: The control device includes a memory and a processor, wherein the memory stores program code, and the processor can execute the method according to any one of claims 1 to 7 when reading and executing the program code in the memory.

9. A storage medium, characterized in that: The storage medium stores an executable program, which can implement the method of any one of claims 1 to 7 when the program is read and executed.

10. An intelligent electric heavy truck, characterized in that: The intelligent electric heavy-duty truck adopts the control method of any one of claims 1-7.

Images