Vehicle active braking control system and method based on extended state observation

Abstract

The invention discloses a vehicle active braking control method based on extended state observation, and the method comprises the steps: estimating unknown parameters of a vehicle through constructing an extended state observer, obtaining the output torque of the vehicle through an accessibility control strategy, obtaining the actual vehicle speed of the vehicle, and determining a vehicle speed estimation error. Calculating a discrete estimation performance index through a discrete estimation performance index function of the extended state observer, iteratively optimizing the extended state observer through a particle swarm optimization algorithm, and iteratively obtaining an optimal solution of a parameter vector of the extended state observer so as to minimize the discrete estimation performance index of the extended state observer. And the unknown parameters of the vehicle are optimally estimated to obtain the optimal output torque of the vehicle, so that the active braking control of the vehicle is realized, and the stable operation of the vehicle under the optimal output torque is ensured. The dynamic behavior can be monitored and adjusted in real time in the running process of the vehicle, and therefore the braking performance and safety of the vehicle are improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle active braking control, and in particular to a vehicle active braking control system and method based on extended state observation. Background Art

[0002] With the rapid development of high-speed railways, vehicles have put forward higher requirements for transportation efficiency and passenger comfort. However, during high-speed operation, the braking safety and braking distance of vehicles have become technical problems that need to be solved urgently. Traditional braking systems usually rely on fixed control parameters, but under the conditions of vehicle load changes, complex operating environment and uncertain track conditions, this fixed parameter control method will lead to unstable braking performance and reduced response speed.

[0003] In recent years, researchers have begun to explore control technologies based on dynamic parameter adjustment, aiming to improve the adaptability and real-time performance of braking systems. Among them, the extended state observer (ESO), as an effective nonlinear system state estimation tool, is widely used in the estimation of unknown disturbances and system parameters. However, the existing extended state observers often face the challenges of difficult parameter adjustment and insufficient estimation accuracy when dealing with the complex nonlinear characteristics and dynamic changes of vehicles. At the same time, as a key factor affecting the braking performance of the vehicle, accurate control of the slip rate can effectively prevent the wheel axle from locking, thereby improving the braking effect. At present, the slip rate control method based on accessibility control has gradually attracted attention, but in practical applications, how to adjust the control strategy in real time to cope with the changing operating conditions is still a problem that needs to be solved urgently.

[0004] Therefore, it is of great significance to develop an advanced technology that can estimate unknown parameters in real time and dynamically adjust the slip ratio control, which not only helps to improve the braking safety of the vehicle and shorten the braking distance, but also optimizes the overall operating performance of the vehicle. Summary of the invention

[0005] The purpose of the present invention is to provide a control strategy that extends the state observer and accessibility, performs real-time estimation of unknown parameters of the vehicle and dynamic slip rate control, and can monitor and adjust the dynamic behavior in real time during vehicle operation, thereby improving the braking performance and safety of the vehicle.

[0006] In order to solve the above technical problems, the technical solution of the present invention is:

[0007] In a first aspect, the present invention provides a vehicle active braking control method based on extended state observation, comprising the following steps:

[0008] Step 1: Construct an extended state observer based on the vehicle dynamics model;

[0009] Step 2: Estimating unknown parameters of the vehicle by using an extended state observer, wherein the unknown parameters include vehicle speed and slip ratio;

[0010] Step 3: Obtain the vehicle output torque through the accessibility control strategy, and obtain the actual vehicle speed under the vehicle output torque;

[0011] Step 4: Based on the estimated vehicle speed obtained in step 2 and the actual vehicle speed obtained in step 3, determine the vehicle speed estimation error, and calculate the discrete estimation performance index by extending the discrete estimation performance index function of the state observer;

[0012] Step 5: Based on the discrete estimation performance index, the extended state observer is optimized by a particle swarm optimization algorithm to calculate a new extended state observer parameter vector;

[0013] Step 6: based on the new extended state observer parameter vector obtained in step 5, the extended state observer in step 1 is updated, and steps 2 to 5 are repeatedly performed using the updated extended state observer until the optimal solution of the extended state observer parameter vector is obtained by iterative optimization to minimize the discrete estimation performance index;

[0014] Step 7: Based on the optimal solution of the extended state observer parameter vector, execute steps 2 to 3 to optimally estimate the unknown parameters of the vehicle and obtain the optimal output torque of the vehicle to realize active braking control of the vehicle and ensure that the vehicle runs stably under the optimal output torque.

[0015] Preferably, in step 1, the vehicle dynamics model includes a vehicle longitudinal dynamics model, a wheel angular velocity dynamics model and a magic tire friction dynamics model.

[0016] Preferably, in step 1, constructing an extended state observer comprises the following steps:

[0017] Step 1.1: Build a vehicle system model;

[0018] Let x 1 =v t , x 2 =w,u=T b , the vehicle system model is expressed as follows:

[0019]

[0020]

[0021] Among them, x 1 =v t represents the longitudinal speed of the vehicle, x 2 =w represents the wheel angular velocity, is the vehicle longitudinal acceleration, Represents the wheel angular acceleration, u=T b represents the vehicle braking input, ρ represents the wind resistance coefficient, m represents one-fourth of the total mass of the vehicle, g represents the acceleration due to gravity, μ represents the adhesion coefficient between the wheel and the contact surface, r represents the wheel radius, and j represents the moment of inertia of the wheel;

[0022] Step 1.2: Introduce uncertain terms and model errors into the vehicle system model, obtain the state equation of the extended model, and calculate the output of the extended model; the uncertain terms are defined as The model error is defined as Δf 1 and Δf 2 ;

[0023] The state equation of the extended model is expressed as follows:

[0024]

[0025]

[0026]

[0027] Then, the output y of the extended model is:

[0028] y=C[x 1 ,x 2 ,x 3 ] T

[0029] Among them, x 1 =v t represents the longitudinal speed of the vehicle, x 2 =w represents the wheel angular velocity, is the vehicle longitudinal acceleration, is the wheel angular acceleration, x 3 =F(λ)=μ represents the adhesion coefficient, represents the rate of change of adhesion coefficient with slip rate, ρ represents wind resistance coefficient, g represents gravitational acceleration, Δf 1 represents the longitudinal speed error of the vehicle system model caused by the uncertainty term, Δf 2 represents the wheel angular velocity error of the vehicle system model caused by the uncertainty term, m represents one quarter of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, C represents the output matrix and C = [1,0,0], [x 1 ,x 2 ,x 3 ] T represents the vehicle system state vector;

[0030] Step 1.3: Introduce the vehicle system state vector in step 1.2 into the extended model to construct an extended state model;

[0031] Let X = [x 1 ,x 2 ,x 3 ] T Represents the vehicle system state vector, then the extended state model is expressed as follows:

[0032]

[0033] in, represents the derivative of the vehicle system state vector X, A represents the state matrix, B represents the input matrix, G represents the extended disturbance term, u=T b represents the vehicle braking input, y represents the output of the extended model, C represents the output matrix and C=[1,0,0];

[0034] The state matrix A, input matrix B and extended disturbance term G are expressed as follows:

[0035]

[0036] Where: ρ represents the wind resistance coefficient, g represents the acceleration of gravity, m represents one-fourth of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, Δf 1 represents the longitudinal speed error of the vehicle system model caused by the uncertainty term, Δf 2 represents the wheel angular velocity error of the vehicle system model caused by the uncertainty term, and f(t) represents the rate of change of the adhesion coefficient with the slip rate;

[0037] Step 1.4: Introduce the vehicle system extended state estimation vector into the extended state model to construct an extended state observer;

[0038] The estimate of the vehicle system state vector X is defined as the vehicle system extended state estimate vector The formula is as follows:

[0039]

[0040] The extended state observer is expressed as follows:

[0041]

[0042] in, represents the estimated longitudinal speed of the vehicle, represents the estimated wheel angular velocity, represents the estimated adhesion coefficient, represents the vehicle system extended state estimation vector The derivative of , A represents the state matrix, B represents the input matrix, u = T b Represents vehicle braking input, L=[l 1 ,l 2 ,l 3 ] T represents the extended state observer parameter vector, i.e., the extended state observer gain matrix, l 1 ,l 2 ,l 3 Respectively represent the first to third elements of L, y represents the output of the extended model, C represents the output matrix and C = [1,0,0].

[0043] Preferably, in step 3, obtaining the vehicle output torque through a control strategy of accessibility comprises the following steps:

[0044] Step 3.1: Based on the expected slip ratio and the actual vehicle speed, obtain the expected wheel angular velocity;

[0045] The desired wheel angular velocity is calculated as follows:

[0046]

[0047] in, represents the expected wheel angular velocity in the kth cycle, λ * represents the expected slip rate, r represents the wheel radius, represents the actual longitudinal speed of the vehicle at the final moment of the kth cycle;

[0048] Step 3.2: In a given vehicle braking cycle, the actual wheel angular velocity is adjusted to the desired wheel angular velocity in step 3.1, and the desired wheel angular acceleration is calculated;

[0049] The expected wheel angular acceleration is calculated as follows:

[0050]

[0051] in, represents the expected wheel angular acceleration of the kth cycle, represents the expected wheel angular velocity in the kth cycle, represents the actual wheel angular velocity of the kth cycle, and Δt represents the rate of change of the wheel angular velocity during each vehicle braking cycle;

[0052] Step 3.3: In the kth cycle, the vehicle output torque is calculated by the reachability controller based on the desired wheel angular acceleration obtained in step 3.2;

[0053] The vehicle output torque is calculated as follows:

[0054]

[0055] Among them, T k represents the vehicle output torque of the kth cycle, j represents the wheel moment of inertia, T f represents the friction torque, represents the expected wheel angular acceleration at the kth cycle.

[0056] Preferably, the step 3 further includes initializing the reachability controller, and calculating the output torque of the vehicle at the initial moment based on the ideal wheel angular velocity at the initial moment, and the calculation formula is as follows:

[0057]

[0058] Among them, T 0 represents the vehicle output torque at the initial moment, j represents the wheel moment of inertia, represents the ideal wheel angular acceleration at the initial moment.

[0059] Preferably, in step S4, the discrete estimation performance indicator function of the extended state observer is expressed as follows:

[0060]

[0061] Among them, J(L) represents the discrete estimation performance index of the extended state observer, n represents the number of discrete estimation performance index sampling, and N represents the total number of discrete estimation performance index sampling. represents the estimation error of the vehicle longitudinal speed, and L represents the extended state observer parameter vector.

[0062] Preferably, in step S5, the particle swarm optimization algorithm comprises the following steps:

[0063] Step 5.1: Randomly initialize the position and velocity of each particle in the solution space of the extended state observer parameter vector;

[0064] Step 5.2: Define the individual optimal fitness of each particle and the global optimal fitness of the group by extending the discrete estimation performance index of the state observer, and initialize the local optimal position pbest of each particle and the global optimal position gbest of the group;

[0065] Step 5.3: Use the particle position update equation and velocity update equation to iteratively update the position and velocity of each particle, and recalculate the fitness of each particle after the update;

[0066] Step 5.4: Determine whether the fitness of each particle is less than its individual optimal fitness one by one;

[0067] If it is less than, update the individual optimal fitness and local optimal position pbest of the particle;

[0068] If not less than, do not update;

[0069] Step 5.5: Determine whether the fitness of each particle is less than the global optimal fitness of the group one by one;

[0070] If it is less than, update the global optimal fitness and global optimal position gbest of the group;

[0071] If not less than, do not update;

[0072] Step 5.6: Loop through steps 5.3 to 5.5 until the maximum number of iterations is reached and then execute step 5.7;

[0073] Step 5.7: Output the global optimal position gbest of the group, that is, the updated extended state observer parameter vector.

[0074] Preferably, in step S5.3, the particle position update equation is as follows:

[0075] x i (n+1)=x i (n)+v i (n+1)

[0076] Among them, x i (n+1) represents the position of particle i at the n+1th iteration, x i (n) represents the position of particle i at the nth iteration, v i (n+1) represents the velocity of particle i at the n+1th iteration;

[0077] The particle velocity update equation is as follows:

[0078] v i (n+1)=a·v i (n)+c 1 ·r 1 ·(pbest i -x i (n))+c 2 ·r 2 ·(gbest-x i (n))

[0079] Among them, v i (n+1) represents the velocity of particle i at the n+1th iteration, v i (n) represents the velocity of particle i at the nth iteration, a is the inertia weight, c 1 represents the weight of the particle moving to the local optimal position, c2 represents the weight of the particle moving to the global optimal position, r 1 Represents a random array of [0,1] under the influence of the local optimal position of the particle in the current iteration, r 2 pbest represents a random array of [0,1] under the influence of the global optimal position of the particle in the current iteration. i represents the local optimal position of particle i, gbest represents the global optimal position, and x i (n) represents the position of particle i at the nth iteration.

[0080] In a second aspect, the present invention provides a vehicle active braking control system based on extended state observation, including a vehicle system module, a particle swarm optimization module, an extended state observer module and a reachability controller module;

[0081] Vehicle system module, used to simulate the vehicle's motion performance;

[0082] a reachability controller module to calculate the torque required to brake the vehicle;

[0083] An extended state observer module is used to estimate unknown parameters during vehicle braking;

[0084] The particle swarm optimization module is used to iteratively optimize the extended state observer and determine the optimal solution of the extended state observer parameter vector.

[0085] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) by updating the extended state observer, it is used to estimate the unknown disturbances and system parameters of the vehicle under complex operating conditions in real time, which can effectively cope with vehicle load changes and environmental uncertainties, and significantly improve the accuracy and reliability of parameter estimation; (2) combined with the adaptive ability of the particle swarm optimization algorithm, the extended state observer can be adjusted online, and the extended state observer parameters can be automatically adjusted according to the real-time operating state of the vehicle, thereby ensuring that the best estimated performance indicators are continuously provided under different operating conditions; (3) based on the estimated slip rate, the control strategy of accessibility is adopted to dynamically adjust the braking strategy, and the dynamic slip rate optimization can stabilize the slip rate near the expected value, prevent the wheel axle from locking, further optimize the braking distance, and improve the braking efficiency and safety. The present invention can respond quickly during the actual operation of the vehicle, effectively reduce the braking delay, and improve the overall efficiency and reliability of the braking system. The present invention significantly improves the braking performance of the vehicle under various loads and operating conditions, enhances the safety and stability of the system, and provides more reliable technical support for the safe operation of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0087] Figure 1 It is a flow chart of the vehicle active braking control method based on extended state observation of the present invention;

[0088] Figure 2 is a curve diagram showing the relationship between the vehicle slip rate and the road adhesion coefficient under different road conditions of the present invention;

[0089] Figure 3 It is a schematic diagram of the optimization algorithm flow of the extended state observer of the present invention;

[0090] Figure 4 A schematic diagram of the change of the adhesion coefficient during the vehicle braking process of the experimental example of the present invention;

[0091] Figure 5 A schematic diagram of vehicle deceleration during vehicle braking in an experimental example of the present invention;

[0092] Figure 6 It is a schematic diagram of the change of the slip ratio during the vehicle braking process of the experimental example of the present invention;

[0093] Figure 7 is the first element l of the extended state observer parameter vector L during the vehicle braking process of the experimental example of the present invention. 1 Schematic diagram of the changes;

[0094] Figure 8 is the second element l of the extended state observer parameter vector L during the vehicle braking process of the experimental example of the present invention 2 Schematic diagram of the changes;

[0095] Fig. 9 is the third element l of the extended state observer parameter vector L during the vehicle braking process of the experimental example of the present invention. 3 Schematic diagram of the changes. DETAILED DESCRIPTION

[0096] The following is a further detailed description of the present invention in conjunction with the accompanying drawings. The terminal technical solution of the embodiment of the present invention is clearly and completely described. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0097] like Figure 1 As shown, the present invention provides a vehicle active braking control method based on extended state observation, comprising the following steps:

[0098] Step 1: Based on the vehicle dynamics model, construct an extended state observer (ESO).

[0099] In this embodiment, the vehicle dynamics model includes a vehicle longitudinal dynamics model, a wheel angular velocity dynamics model and a magic tire friction dynamics model.

[0100] (1) Vehicle longitudinal dynamics model

[0101] A simple and effective quarter vehicle model is used to construct the vehicle longitudinal dynamics model. In the longitudinal dynamics analysis, the longitudinal motion of the vehicle is mainly affected by the friction between the vehicle and the ground and the air resistance.

[0102] The vehicle longitudinal dynamic model is expressed as follows:

[0103]

[0104] F R =mρgv t

[0105] in, represents the vehicle longitudinal acceleration, v t represents the longitudinal speed of the vehicle, m represents one quarter of the total mass of the vehicle, ρ represents the wind resistance coefficient, g represents the acceleration of gravity, and F f represents the friction force, F R Represents the total wind resistance acting on the vehicle.

[0106] (2) Wheel angular velocity dynamics model

[0107] The wheel angular velocity is mainly affected by the friction torque and the vehicle output torque. During braking, the vehicle stops the wheels by applying output torque, while the friction torque is caused by the friction force generated by the relative sliding between the wheels and the ground, which acts as a braking torque.

[0108] The wheel angular velocity dynamics model is expressed as follows:

[0109]

[0110] T f =F f r

[0111] F f =mgμ

[0112] Where j represents the wheel moment of inertia, is the wheel angular acceleration, T f represents friction torque, T represents vehicle output torque, F f represents the friction force, r represents the wheel radius, m represents one quarter of the total mass of the vehicle, g represents the acceleration of gravity, and μ represents the adhesion coefficient between the wheel and the contact surface.

[0113] (3) Magic Tire Friction Dynamics Model

[0114] The adhesion coefficient μ depends on the contact characteristics between the wheel and the contact surface. The braking performance of the vehicle is significantly affected by the friction characteristics between the wheel and the contact surface and external environmental factors. In this embodiment, the magic tire friction dynamics model is used to describe the friction characteristics between the wheel and the contact surface, which can effectively capture the transient behavior of the vehicle's adhesion and adapt to different contact conditions between the wheel and the contact surface. The relationship curve between the vehicle adhesion coefficient and the slip rate is shown in Figure 2 shown.

[0115] The friction dynamics model of the magic tire is expressed as follows:

[0116] μ=F(λ)=Dsin{Carctan[Bλ-E(Bλ-arctanBλ)]}

[0117] Where μ represents the adhesion coefficient between the wheel and the contact surface, F(λ) represents the function between slip rate and friction force, B represents the stiffness factor, C represents the shape factor, D represents the peak factor, and E represents the curvature factor. λ is the vehicle slip rate.

[0118] The vehicle slip ratio λ is the ratio of the absolute linear velocity of the wheel to the velocity of the vehicle, and is calculated as follows:

[0119]

[0120] Among them, v t represents the longitudinal speed of the vehicle, w represents the angular velocity of the wheel, and r represents the radius of the wheel.

[0121] Using the magic tire friction dynamics model, the relationship curve between the wheel slip rate λ and the road adhesion coefficient μ under different road conditions can be plotted by numerical methods, such as Figure 2 As shown in the figure. During the vehicle braking process, the deceleration effect of the vehicle depends on the active braking force, which is closely related to the adhesion coefficient. As the wheel slip rate gradually approaches 1 (i.e. the wheel is locked), the adhesion coefficient gradually decreases, which will lead to partial or complete loss of active braking force and may cause permanent damage to the wheel and track.

[0122] The key to vehicle active braking control is to adjust the wheel slip ratio to the optimal value to minimize the braking distance and avoid wheel wear. Although the relationship between the adhesion coefficient and the slip ratio is nonlinear under different road conditions, generally speaking, the adhesion coefficient reaches its peak value when the slip ratio is about 0.15. Therefore, in this embodiment, the expected slip ratio λ is preset. * The value is 0.15 to achieve the best vehicle braking effect.

[0123] The construction of the extended state observer comprises the following steps:

[0124] Step 1.1: Combine the above vehicle longitudinal dynamics model, wheel angular velocity dynamics model and magic tire friction dynamics model to construct a vehicle system model.

[0125] The wheel angular velocity w, the vehicle longitudinal velocity v t and adhesion coefficient μ are defined as the state variables of the vehicle system, let x 1 =v t , x 2 =w,u=T b , the vehicle system model is expressed as follows:

[0126]

[0127]

[0128] Among them, x 1 =v t represents the longitudinal speed of the vehicle, x 2 =w represents the wheel angular velocity, is the vehicle longitudinal acceleration, Represents the wheel angular acceleration, u=T b represents the vehicle braking input, ρ represents the wind resistance coefficient, g represents the acceleration of gravity, μ represents the adhesion coefficient between the wheel and the contact surface, r represents the wheel radius, and j represents the wheel moment of inertia.

[0129] Step 1.2: Introduce uncertain terms and model errors into the vehicle system model, obtain the state equation of the extended model, and calculate the output of the extended model; the uncertain terms are defined as The model error is defined as Δf 1 and Δf 2 .

[0130] Since the vehicle system may be affected by aging, wear and other effects in actual operation, the adhesion coefficient will undergo uncertain changes, which is introduced as an uncertainty term and defined as At the same time, in order to consider the impact of uncertainties on the vehicle system, the model error caused by the change of uncertainties is also introduced, and the model error is defined as Δf 1 and Δf 2 .

[0131] The state equation of the extended model is expressed as follows:

[0132]

[0133]

[0134] Then, the output y of the extended model is:

[0135] y=C[x 1 ,x 2 ,x 3 ] T

[0136] Among them, x 1 =v t represents the longitudinal speed of the vehicle, x 2 =w represents the wheel angular velocity, is the vehicle longitudinal acceleration, is the wheel angular acceleration, x 3 =F(λ)=μ represents the adhesion coefficient, represents the rate of change of adhesion coefficient with slip rate, ρ represents wind resistance coefficient, g represents gravitational acceleration, Δf 1 represents the longitudinal speed error of the vehicle system model caused by the uncertainty term, Δf 2 represents the wheel angular velocity error of the vehicle system model caused by the uncertainty term, m represents one quarter of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, C is the output matrix and C = [1,0,0] means that only the longitudinal speed of the vehicle is measurable, while the wheel angular velocity and adhesion coefficient cannot be directly measured, [x 1 ,x 2 ,x 3 ] T Represents the vehicle system state vector.

[0137] Step 1.3: Introduce the vehicle system state vector output from step 1.2 into the model to construct an extended state model.

[0138] Let X = [x 1 ,x 2 ,x 3 ] T Represents the vehicle system state vector, then the extended state model is expressed as follows:

[0139]

[0140] in, represents the derivative of the vehicle system state vector X, A represents the state matrix, B represents the input matrix, G represents the extended disturbance term, u=T b represents the vehicle braking input, y represents the output of the extended model, C represents the output matrix and C=[1,0,0].

[0141] The state matrix A, input matrix B and extended disturbance term G are expressed as follows:

[0142]

[0143] Where: ρ represents the wind resistance coefficient, g represents the acceleration of gravity, m represents one-fourth of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, Δf 1 represents the longitudinal speed error of the vehicle system model caused by the uncertainty term, Δf 2 represents the wheel angular velocity error of the vehicle system model caused by the uncertainty term, and f(t) represents the rate of change of the adhesion coefficient with the slip rate.

[0144] Step 1.4: Introduce the vehicle system extended state estimation vector into the extended state model to construct an extended state observer;

[0145] The estimate of the vehicle system state vector X is defined as the vehicle system extended state estimate vector The formula is as follows:

[0146]

[0147] The extended state observer is expressed as follows:

[0148]

[0149] in, represents the estimated longitudinal speed of the vehicle, represents the estimated wheel angular velocity, represents the estimated adhesion coefficient, represents the vehicle system extended state estimation vector The derivative of , A represents the state matrix, B represents the input matrix, u = T b Represents vehicle braking input, L=[l 1 ,l 2 ,l 3 ] T represents the extended state observer parameter vector, i.e., the extended state observer gain matrix, l 1 ,l 2 ,l 3Represent the first to third elements of L respectively, y represents the output of the extended model, C represents the output matrix and C=[1,0,0].

[0150] Step 2: Estimate the unknown parameters of the vehicle through the extended state observer, and the unknown parameters include vehicle speed and slip ratio.

[0151] Step 3: Obtain the vehicle output torque through the accessibility control strategy, obtain the actual vehicle speed under the vehicle output torque, sample and record.

[0152] In this embodiment, the vehicle output torque is obtained by using a control strategy of reachability, including the following steps:

[0153] Step 3.1: Based on the expected slip ratio and actual vehicle speed, obtain the expected wheel angular velocity.

[0154] During braking, the friction force varies with the vehicle slip rate. The wheel can be regarded as a first-order time-varying system. The control variable is the vehicle output torque, and the controlled variable is the wheel rotation. The control process of vehicle slip rate optimization is discretized, and the expected wheel angular velocity can be calculated based on the expected slip rate and the actual longitudinal speed of the vehicle.

[0155] The desired wheel angular velocity is calculated as follows:

[0156]

[0157] in, represents the expected wheel angular velocity in the kth cycle, λ * represents the expected slip rate, r represents the wheel radius, Indicates the final actual longitudinal speed of the vehicle in the kth cycle.

[0158] Step 3.2: In a given vehicle braking cycle, the actual wheel angular velocity is adjusted to the desired wheel angular velocity in step 3.1, and the desired wheel angular acceleration is calculated;

[0159] The expected wheel angular acceleration is calculated as follows:

[0160]

[0161] in, represents the expected wheel angular acceleration of the kth cycle, represents the expected wheel angular velocity in the kth cycle, represents the actual wheel angular velocity of the kth cycle, and Δt represents the rate of change of the wheel angular velocity during each vehicle braking cycle;

[0162] Step 3.3: In the kth cycle, the vehicle output torque is calculated by the reachability controller based on the desired wheel angular acceleration obtained in step 3.2;

[0163] The vehicle output torque is calculated as follows:

[0164]

[0165] Among them, T k represents the vehicle output torque T of the kth cycle, j represents the wheel moment of inertia, T f represents the friction torque, represents the expected wheel angular acceleration at the kth cycle.

[0166] In this embodiment, step 3 also includes initializing the reachability controller, and calculating the output torque of the vehicle at the initial moment based on the ideal wheel angular velocity at the initial moment, and the calculation formula is as follows:

[0167]

[0168] Among them, T 0 represents the vehicle output torque T at the initial moment, j represents the wheel moment of inertia, represents the ideal wheel angular acceleration at the initial moment.

[0169] Through the above-mentioned accessibility control strategy, the vehicle system can adjust and control the vehicle output torque in real time according to the actual braking needs of the vehicle, and optimize the vehicle slip rate and wheel angular velocity during braking in real time.

[0170] Step 4: Based on the estimated vehicle speed obtained in step 2 and the actual vehicle speed obtained in step 3, determine the vehicle speed estimation error, and calculate the discrete estimation performance index by extending the discrete estimation performance index function of the state observer.

[0171] In this embodiment, the absolute value integral of the vehicle longitudinal speed estimation error is selected as the estimation performance indicator, focusing on the tracking performance of the extended state observer, then:

[0172] The discrete estimation performance index function of the extended state observer is expressed as follows:

[0173]

[0174] Among them, J(L) represents the discrete estimation performance index of the extended state observer, n represents the number of discrete estimation performance index sampling, and N represents the total number of discrete estimation performance index sampling. expresses the estimation error of the longitudinal speed of the vehicle, L represents the parameter vector of the extended state observer, represents the nth discrete estimation performance index sampling x of the extended state observer parameter vector L 1 The error in the estimated value of .

[0175] The discrete estimation performance indicator can be regarded as a mapping between the extended state observer parameter vector and the estimation performance indicator.

[0176] Step 5: Based on the discrete estimation performance index, the extended state observer is optimized by a particle swarm optimization (PSO) algorithm to calculate a new extended state observer parameter vector.

[0177] Particle swarm optimization (PSO) algorithm is a global optimization method that simulates swarm intelligence, inspired by the foraging behavior of bird flocks. PSO was proposed by James Kennedy and Russell Eberhart in 1995 and was originally used to solve global optimization problems. PSO can efficiently search in complex, multi-peak, and nonlinear problem spaces. Its basic principle is to simulate the foraging process of bird flocks. Each "bird" (i.e., particle) represents a solution, and each particle flies in the solution space and gradually approaches the optimal solution. The movement of the particle is determined by its historical optimal position (individual experience) and the historical optimal position of all particles (group experience).

[0178] The core idea of ​​the PSO algorithm is to find the global optimal solution by iteratively updating the position and velocity of the particles. Each particle has two main attributes in the solution space: position (indicating the current solution) and velocity (controlling the movement of the particle in the solution space).

[0179] like Figure 3 As shown, the particle swarm optimization algorithm comprises the following steps:

[0180] Step 5.1: Randomly initialize the position and velocity of each particle in the solution space of the extended state observer parameter vector L.

[0181] Step 5.2: Define the individual optimal fitness of each particle and the global optimal fitness of the group by extending the discrete estimation performance index of the state observer, and initialize the local optimal position pbest of each particle and the global optimal position gbest of the group.

[0182] Step 5.3: Use the particle position update equation and velocity update equation to iteratively update the position and velocity of each particle, and calculate the fitness of each particle after the update. The calculation formula is as follows:

[0183]

[0184] Among them, pnew irepresents the fitness of particle i after update, L new represents the updated extended state observer parameter vector, n represents the nth discrete estimation performance index sampling, N represents the total number of discrete estimation performance index sampling, Expresses the estimated error of the vehicle's longitudinal speed.

[0185] The particle adjusts its position by updating its velocity. The velocity update equation of the particle is as follows:

[0186] v i (n+1)=a·v i (n)+c 1 ·r 1 ·(pbest i -x i (n))+c 2 ·r 2 ·(gbest-x i (n))

[0187] Among them, v i (n+1) represents the velocity of particle i at the n+1th iteration, v i (n) represents the velocity of particle i at the nth iteration, a is the inertia weight, c 1 represents the weight of the particle moving to the local optimal position, c 2 represents the weight of the particle moving to the global optimal position, r 1 Represents a random array of [0,1] under the influence of the local optimal position of the particle in the current iteration, r 2 pbest represents a random array of [0,1] under the influence of the global optimal position of the particle in the current iteration. i represents the local optimal position of particle i, gbest represents the global optimal position, and x i (n) represents the position of particle i at the nth iteration.

[0188] The particle position update equation is as follows:

[0189] x i (n+1)=x i (n)+v i (n+1)

[0190] Among them, x i (n+1) represents the position of particle i at the n+1th iteration, x i (n) represents the position of particle i at the nth iteration, v i (n+1) represents the velocity of particle i at the n+1th iteration;

[0191] The inertia weight a controls the flying range of the particle in the solution space. A larger inertia weight helps the particle to explore the solution space extensively, while a smaller inertia weight helps the particle to converge more quickly. 1 and c 2 It is used to balance the particle's dependence on the local optimal position and the global optimal position, and is usually set to a constant value (for example: 2) to ensure an appropriate balance between exploration and exploitation.

[0192] Step 5.4: Determine whether the fitness of each particle is less than its individual optimal fitness one by one;

[0193] If it is less than, update the individual optimal fitness and local optimal position pbest of the particle;

[0194] If not less than, do not update;

[0195] Step 5.5: Determine whether the fitness of each particle is less than the global optimal fitness of the group one by one;

[0196] If it is less than, update the global optimal fitness and global optimal position gbest of the group;

[0197] If not less than, do not update;

[0198] Step 5.6: Loop through steps 5.3 to 5.5 until the maximum number of iterations is reached and then execute step 5.7;

[0199] Step 5.7: Output the global optimal position gbest of the group, that is, the updated extended state observer parameter vector.

[0200] Step 6: Update the extended state observer in step 1 based on the new extended state observer parameter vector obtained in step 5, and repeat steps 2 to 5 using the updated extended state observer until the optimal solution of the extended state observer parameter vector is obtained by iterative optimization to minimize the discrete estimation performance index.

[0201] Step 7: Based on the optimal solution of the extended state observer parameter vector, execute steps 2 to 3 to optimally estimate the unknown parameters of the vehicle and obtain the optimal output torque of the vehicle to realize active braking control of the vehicle and ensure that the vehicle runs stably under the optimal output torque.

[0202] The present invention also provides a vehicle active braking control system based on extended state observation, comprising a vehicle system module, a particle swarm optimization (PSO) module, an extended state observer module and a reachability controller module;

[0203] Vehicle system module, used to simulate the vehicle's motion performance;

[0204] a reachability controller module to calculate the torque required to brake the vehicle;

[0205] An extended state observer module is used to estimate unknown parameters during vehicle braking;

[0206] The particle swarm optimization (PSO) module is used to iteratively optimize the extended state observer and determine the optimal solution of the extended state observer parameter vector so that the extended state observer has better estimation performance.

[0207] The effectiveness of the embodiments of the present invention is verified by the following experimental examples.

[0208] 1. Initialize the extended state observer

[0209] Firstly, according to the set initial parameters of the extended state observer, the unknown parameters of the vehicle are estimated, and the estimated longitudinal speed of the vehicle is recorded; then, the vehicle output torque is calculated through the accessibility control strategy, the vehicle is operated under the output torque, the actual longitudinal speed of the vehicle under the vehicle output torque is sampled and recorded, the estimated longitudinal speed of the vehicle is compared with the actual longitudinal speed of the vehicle, the longitudinal speed estimation error of the vehicle is calculated, and according to the longitudinal estimation error of the vehicle, the discrete estimation performance index J(L) of the current extended state observer is calculated through the discrete estimation performance index function of the extended state observer, and the discrete estimation performance index J(L) is sampled and recorded.

[0210] 2. Extended state observer iterative optimization:

[0211] Based on the discrete estimation performance index J(L) obtained by sampling, the extended state observer is iteratively optimized. The particle swarm optimization (PSO) algorithm is used to calculate the new extended state observer gain matrix L, which is used to update the extended state observer. The updated extended state observer is used to estimate the unknown parameters of the vehicle again, and the vehicle is controlled based on the control strategy of accessibility. The vehicle continues to drive and samples in real time, and the discrete estimation performance index of the extended state observer is calculated at this time. The adhesion coefficient of the vehicle when driving is as follows Figure 4 As shown in the figure, the longitudinal speed of the vehicle is as follows: Figure 5 As shown in Figure 2, the slip rate of the vehicle when driving is as follows: Figure 6 shown.

[0212] 3. Data comparison and decision-making:

[0213] The particle fitness of the current iteration calculated based on the discrete estimation performance index of the extended state observer is compared with the particle fitness of the previous iteration. According to the comparison results, the local optimal position pbest and the global optimal position gbest of each particle are updated, and the extended state observer parameter vector L required for the next iteration is calculated.

[0214] 4. Iteration loop:

[0215] Repeat the steps of extended state observer iterative optimization, data comparison and decision-making, and dynamic slip rate optimization until the optimal extended state observer gain matrix L is found. The first element of the extended state observer gain matrix L is updated as follows: Figure 7 As shown, the second element is updated as Figure 8 As shown, the third element is updated as shown in 9.

[0216] 5. Output the optimal extended state observer gain matrix L:

[0217] The optimal extended state observer gain matrix L is used to update the extended state observer, the unknown parameters of the vehicle are optimally estimated, and the optimal output torque of the vehicle is obtained by using the reachability controller to control the vehicle and ensure its stable operation.

[0218] The parts not involved in the present invention are the same as the prior art or are implemented by using the prior art.

[0219] The above contents are further detailed descriptions of the present invention in combination with specific implementation methods, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field of the present invention, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the scope of protection of the present invention.

Claims

1. A vehicle active braking control method based on extended state observation, characterized in that: The steps include: Step 1: Construct an extended state observer based on the vehicle dynamics model; Step 2: Estimating unknown parameters of the vehicle by using an extended state observer, wherein the unknown parameters include vehicle speed and slip ratio; Step 3: Obtain the vehicle output torque through the accessibility control strategy, and obtain the actual vehicle speed under the vehicle output torque; Step 4: Based on the estimated vehicle speed obtained in step 2 and the actual vehicle speed obtained in step 3, determine the vehicle speed estimation error, and calculate the discrete estimation performance index by extending the discrete estimation performance index function of the state observer; Step 5: Based on the discrete estimation performance index, the extended state observer is optimized by a particle swarm optimization algorithm to calculate a new extended state observer parameter vector; Step 6: based on the new extended state observer parameter vector obtained in step 5, the extended state observer in step 1 is updated, and steps 2 to 5 are repeatedly performed using the updated extended state observer until the optimal solution of the extended state observer parameter vector is obtained by iterative optimization to minimize the discrete estimation performance index; Step 7: Based on the optimal solution of the extended state observer parameter vector, execute steps 2 to 3 to optimally estimate the unknown parameters of the vehicle and obtain the optimal output torque of the vehicle to realize active braking control of the vehicle and ensure that the vehicle runs stably under the optimal output torque.

2. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: In step 1, the vehicle dynamics model includes a vehicle longitudinal dynamics model, a wheel angular velocity dynamics model and a magic tire friction dynamics model.

3. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: In step 1, constructing an extended state observer includes the following steps: Step 1.1: Build a vehicle system model; Let x1 = v t , x2=w,u=T b , the vehicle system model is expressed as follows: Where x1 = v t represents the longitudinal speed of the vehicle, x2=w represents the angular velocity of the wheel, is the vehicle longitudinal acceleration, Represents the wheel angular acceleration, u=T b represents the vehicle braking input, ρ represents the wind resistance coefficient, m represents one-fourth of the total mass of the vehicle, g represents the acceleration due to gravity, μ represents the adhesion coefficient between the wheel and the contact surface, r represents the wheel radius, and j represents the moment of inertia of the wheel; Step 1.2: Introduce uncertain terms and model errors into the vehicle system model, obtain the state equation of the extended model, and calculate the output of the extended model; the uncertain terms are defined as x3 = F(λ) = μ, The model errors are defined as Δf1 and Δf2; The state equation of the extended model is expressed as follows: Then, the output y of the extended model is: y=C[x1,x2,x3] T Where x1 = v t represents the longitudinal speed of the vehicle, x2=w represents the angular velocity of the wheel, is the vehicle longitudinal acceleration, represents the wheel angular acceleration, x3=F(λ)=μ represents the adhesion coefficient, represents the rate of change of adhesion coefficient with slip rate, ρ represents wind resistance coefficient, g represents gravity acceleration, Δf1 represents the longitudinal speed error of vehicle system model caused by uncertain terms, Δf2 represents the wheel angular velocity error of vehicle system model caused by uncertain terms, m represents one quarter of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, C represents the output matrix and C=[1,0,0],[x1,x2,x3] T represents the vehicle system state vector; Step 1.3: Introduce the vehicle system state vector in step 1.2 into the extended model to construct an extended state model; Let X = [x1, x2, x3] T Represents the vehicle system state vector, then the extended state model is expressed as follows: in, represents the derivative of the vehicle system state vector X, A represents the state matrix, B represents the input matrix, G represents the extended disturbance term, and u=T b represents the vehicle braking input, y represents the output of the extended model, C represents the output matrix and C=[1,0,0]; The state matrix A, input matrix B and extended disturbance term G are expressed as follows: Where: ρ represents the wind resistance coefficient, g represents the acceleration of gravity, m represents one-fourth of the total mass of the vehicle, r represents the wheel radius, j represents the wheel moment of inertia, Δf1 represents the longitudinal speed error of the vehicle system model caused by the uncertain term, Δf2 represents the wheel angular velocity error of the vehicle system model caused by the uncertain term, and f(t) represents the rate of change of the adhesion coefficient with the slip rate; Step 1.4: Introduce the vehicle system extended state estimation vector into the extended state model to construct an extended state observer; The estimate of the vehicle system state vector X is defined as the vehicle system extended state estimate vector The formula is as follows: The extended state observer is expressed as follows: in, represents the estimated longitudinal speed of the vehicle, represents the estimated wheel angular velocity, represents the estimated adhesion coefficient, represents the vehicle system extended state estimation vector The derivative of , A represents the state matrix, B represents the input matrix, u = T b Represents vehicle braking input, L = [l1, l2, l3] T represents the extended state observer parameter vector, that is, the extended state observer gain matrix, l1, l2, l3 represent the first to third elements of L respectively, y represents the output of the extended model, C represents the output matrix and C = [1, 0, 0].

4. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: In step 3, the vehicle output torque is obtained by using a control strategy of accessibility, which includes the following steps: Step 3.1: Based on the expected slip ratio and the actual vehicle speed, obtain the expected wheel angular velocity; The desired wheel angular velocity is calculated as follows: in, represents the expected wheel angular velocity in the kth cycle, λ * represents the expected slip rate, r represents the wheel radius, represents the actual longitudinal speed of the vehicle at the final moment of the kth cycle; Step 3.2: In a given vehicle braking cycle, the actual wheel angular velocity is adjusted to the desired wheel angular velocity in step 3.1, and the desired wheel angular acceleration is calculated; The expected wheel angular acceleration is calculated as follows: in, represents the expected wheel angular acceleration of the kth cycle, represents the expected wheel angular velocity in the kth cycle, represents the actual wheel angular velocity of the kth cycle, and Δt represents the rate of change of the wheel angular velocity during each vehicle braking cycle; Step 3.3: In the kth cycle, the vehicle output torque is calculated by the reachability controller based on the desired wheel angular acceleration obtained in step 3.2; The vehicle output torque is calculated as follows: Among them, T k represents the vehicle output torque of the kth cycle, j represents the wheel moment of inertia, T f represents the friction torque, represents the expected wheel angular acceleration at the kth cycle.

5. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: The step 3 also includes initializing the reachability controller, and calculating the output torque of the vehicle at the initial moment based on the ideal wheel angular velocity at the initial moment, and the calculation formula is as follows: Where T0 represents the vehicle output torque at the initial moment, j represents the wheel moment of inertia, represents the ideal wheel angular acceleration at the initial moment.

6. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: In step S4, the discrete estimation performance index function of the extended state observer is expressed as follows: Among them, J(L) represents the discrete estimation performance index of the extended state observer, n represents the number of discrete estimation performance index sampling, and N represents the total number of discrete estimation performance index sampling. represents the estimation error of the vehicle longitudinal speed, and L represents the extended state observer parameter vector.

7. The vehicle active braking control method based on extended state observation according to claim 1, characterized in that: In step S5, the particle swarm optimization algorithm comprises the following steps: Step 5.1: Randomly initialize the position and velocity of each particle in the solution space of the extended state observer parameter vector; Step 5.2: Define the individual optimal fitness of each particle and the global optimal fitness of the group by extending the discrete estimation performance index of the state observer, and initialize the local optimal position pbest of each particle and the global optimal position gbest of the group; Step 5.3: Use the particle position update equation and velocity update equation to iteratively update the position and velocity of each particle, and recalculate the fitness of each particle after the update; Step 5.4: Determine whether the fitness of each particle is less than its individual optimal fitness one by one; If it is less than, update the individual optimal fitness and local optimal position pbest of the particle; If not less than, do not update; Step 5.5: Determine whether the fitness of each particle is less than the global optimal fitness of the group one by one; If it is less than, update the global optimal fitness and global optimal position gbest of the group; If not less than, do not update; Step 5.6: Loop through steps 5.3 to 5.5 until the maximum number of iterations is reached and then execute step 5.7; Step 5.7: Output the global optimal position gbest of the group, that is, the updated extended state observer parameter vector.

8. The vehicle active braking control method based on extended state observation according to claim 7, characterized in that: In step S5.3, the particle position update equation is as follows: x i (n+1)=x i (n)+v i (n+1) Among them, x i (n+1) represents the position of particle i at the n+1th iteration, x i (n) represents the position of particle i at the nth iteration, v i (n+1) represents the velocity of particle i at the n+1th iteration; The particle velocity update equation is as follows: v i (n+1)=a·v i (n)+c1·r1·(pbest i -x i (n))+c2·r2·(gbest-x i (n)) Among them, v i (n+1) represents the velocity of particle i at the n+1th iteration, v i (n) represents the velocity of particle i at the nth iteration, a is the inertia weight, c1 represents the weight of the particle moving to the local optimal position, c2 represents the weight of the particle moving to the global optimal position, r1 represents the random array [0,1] of the particle under the influence of the local optimal position in the current iteration, r2 represents the random array [0,1] of the particle under the influence of the global optimal position in the current iteration, pbest i represents the local optimal position of particle i, gbest represents the global optimal position, and x i (n) represents the position of particle i at the nth iteration.

9. A vehicle active braking control system based on extended state observation, characterized in that: It includes vehicle system module, particle swarm optimization module, extended state observer module and reachability controller module; Vehicle system module, used to simulate the vehicle's motion performance; a reachability controller module to calculate the torque required to brake the vehicle; An extended state observer module is used to estimate unknown parameters during vehicle braking; The particle swarm optimization module is used to iteratively optimize the extended state observer and determine the optimal solution of the extended state observer parameter vector.

Images