A cruise vehicle speed control method based on sliding mode control

Abstract

The application discloses a cruise vehicle speed control method based on a sliding mode control, comprising the following steps: obtaining driving data, and constructing a second-order system with an actual vehicle speed x1 and an actual acceleration x2 as state variables; constructing a sliding mode controller about a vehicle speed error e1 and a sliding mode surface s based on the second-order system; calculating an expected control law based on the vehicle speed error e1 and the sliding mode surface s, loading a control law prediction model, obtaining a switching control law according to the sliding mode surface s and a derivative of the sliding mode surface, integrating the expected control law and the switching control law to obtain a total control law F t , and sending a control instruction to a commercial vehicle driver and a brake according to the total control law F t . According to the above technical scheme, the control precision of the vehicle speed control system can be improved, and the stability and control efficiency of the vehicle speed control can be ensured.

Description

Technical Field

[0001] This invention relates to the field of active safety technology for automobiles, and more specifically, to a cruise speed control method based on sliding mode control. Background Technology

[0002] As automobiles become increasingly common, improper driving behaviors such as speeding and fatigued driving seriously affect road safety. This has led to the development of active safety technologies, particularly adaptive cruise control, which ensures driving safety. Predictive adaptive cruise control, in particular, uses high-precision maps to obtain road information such as slope, curvature, and speed limits to achieve optimal control of commercial vehicle speed and gear selection, resulting in higher safety and fuel economy.

[0003] However, due to the uncertainty of parameters in the vehicle speed control system of commercial vehicles (such as vehicle mass, friction coefficient, engine efficiency, sensor monitoring error, etc.), external driving environment interference (such as other vehicles overtaking, road congestion, sudden weather changes), and controller design problems, commercial vehicles are subject to external or internal disturbances during cruising, resulting in lower control performance, weaker anti-interference ability, and worse control accuracy of the vehicle speed control system.

[0004] In the existing technology, algorithmic control is used to improve anti-interference performance. For example, the article entitled "A Ground Autonomous Vehicle Model Predictive Speed ​​Tracking Control Method" proposes a hierarchical longitudinal control method; the article entitled "Self-learning Cruise Control Based on Kernel Least Squares Strategy Iteration" proposes a combination of PI control and reinforcement learning to obtain the optimal switching strategy; and integral modules are used to handle steady-state errors. However, these control algorithms still have many shortcomings: (1) For the stability problem of commercial vehicle cruise speed control system, most traditional algorithms have high computational requirements, but the control principle of traditional algorithms is complex and difficult to debug. Therefore, in practical applications, they greatly reduce the anti-interference performance of the vehicle speed control system.

[0005] (2) In response to defect (1), sliding mode control is proposed. Its principle is simple and its anti-interference ability is strong. However, the traditional commercial vehicle cruise speed control system mostly uses first-order sliding mode control. Although it can correct the actual vehicle speed in time, it also has the problems of large actual output speed jitter and large workload of speed error debugging.

[0006] (3) Currently, the sliding mode control scheme used in commercial vehicle speed control systems is based on the vehicle speed backstepping method to design control laws in order to improve the problem of actual output vehicle speed jitter. However, when it is in a large dimension or has a long computation time (for example, it needs to process multiple inputs and state variables: vehicle speed, acceleration, road information obtained by high-precision maps and sensors, driving environment conditions, etc.), it is easy to cause differential (computational) explosion or the absence of higher-order derivatives, which leads to poor efficiency in controlling vehicle speed correction and poor sensitivity in adjusting vehicle speed error. (4) In the traditional cruise control of commercial vehicles, there is less control to suppress unknown external interference, resulting in poor control accuracy of the vehicle speed control system. Summary of the Invention

[0007] To address the above problems, this application provides a cruise speed control method based on sliding mode control, comprising the following steps:

[0008] Acquire driving data and construct a second-order system with actual vehicle speed x1 and actual acceleration x2 as state variables; the driving data includes vehicle driving parameters and desired indices, the desired indices including desired speed v. d Expected acceleration Expected variable acceleration

[0009] A sliding mode controller is constructed based on the second-order system, relating the vehicle speed error e1 to the sliding surface s.

[0010] The desired control law is calculated based on the vehicle speed error e1 and the sliding surface s.

[0011] Load the control law prediction model, based on the sliding surface s and the sliding surface derivative. Obtain the switching control law

[0012] Integrating the desired control law and switching control laws Obtain the overall control law The overall control law described by the integral Find the driving force F t According to the driving force F t It sends control commands to the commercial vehicle's drive and brake systems.

[0013] Acquiring driving data includes: obtaining vehicle driving parameters from sensors and predicting desired indicators;

[0014] Among them, the vehicle driving parameters obtained from the sensors include the current vehicle speed v and acceleration. Vehicle load Vehicle net weight (m), slope information (α), atmospheric pressure (P), temperature (T), engine torque (T) tq Engine speed n;

[0015] The predicted expected index refers to the expected index calculated based on the vehicle driving parameters and the vehicle dynamics model.

[0016] Furthermore: The second-order system is represented as:

[0017] and:

[0018]

[0019]

[0020] Where M is the total mass of the vehicle, Where m is the vehicle's additional load capacity, and F is the vehicle's unloaded mass. f For rolling resistance, F i For slope resistance, F w For air resistance, F c For unknown external interference forces on the vehicle, F t Where f is the driving force, f is the rolling resistance coefficient, and C is the rolling resistance coefficient. D Let A be the air resistance coefficient, ρ be the frontal area, ρ be the air density, and Q be the interference term.

[0021] Furthermore, a sliding mode controller is defined, whose sliding variables are the vehicle speed tracking error e1 and the system state of the second-order system, corresponding to the sliding surface s, which is used to realize the state variables vehicle speed v and acceleration. The control law that reaches s=0 in a finite amount of time.

[0022] The execution process of constructing a sliding mode controller based on the second-order system regarding the vehicle speed error e1 and the sliding surface s includes:

[0023] The speed error is defined as: e1 = x1 - x 1d ,

[0024] Where x1 is the actual vehicle speed, x 1d Let x be the desired speed. 1d =v d ;

[0025] The sliding mode function is defined based on the actual vehicle speed and the speed error. The sliding mode function is expressed as follows:

[0026] and

[0027] Based on the definition of the sliding mode function, the Lyapunov function V1 is expressed as:

[0028] The derivative of the Lyapunov function V1 is expressed as:

[0029]

[0030] Furthermore, the desired control law is calculated based on the vehicle speed error e1 and the sliding surface s under the control of the sliding mode controller. include:

[0031] Construct a first-order low-pass filter with desired acceleration and desired variable acceleration as state variables. The first-order low-pass filter is expressed as: Where α1 is the input quantity of the first-order inertial element and T2 is the random coefficient;

[0032] By using a first-order low-pass filter and the vehicle speed error e1 and sliding surface s, the expression for the derivative of the Lyapunov function V1 is updated as follows:

[0033]

[0034] Where k1 is the random coefficient, For expected control laws;

[0035] The desired control law can be obtained from the above formula. The calculation method is expressed as:

[0036]

[0037] Before loading the control law prediction model, the control law prediction model is constructed, including:

[0038] Define a BP fuzzy neural network;

[0039] Train a backpropagation (BP) fuzzy neural network to obtain optimal model parameters, including the optimal center value c. ij Width σ ij and connection weight w ij ;

[0040] Based on the optimal model parameters, construct a control law prediction model.

[0041] The control law prediction model is a two-input single-output system, with input parameter x. i This includes the sliding surface s and the derivative of the sliding surface. The output parameter is the switching control law.

[0042] Obtaining driving force F t Refers to: the switching control law and expected control law Accumulate the sums and apply the overall control law to the accumulated values. Integrate the forces; the result is the driving force F.t .

[0043] Furthermore, based on the optimal model parameters of the control law prediction model, the disturbance force can be predicted. include:

[0044] Set up an algorithm for predicting the estimated value of unknown external interference:

[0045]

[0046] Where X and ω are the input parameters of the prediction algorithm. The input parameters obtained for the j-th time corresponding to parameter X include the sliding surface s and the derivative of the sliding surface. ω T ω represents the optimal connection weights for the control law prediction model.

[0047] Furthermore, in a second-order system, the parameters such as the unknown external disturbance force experienced by the vehicle can be predicted by analyzing the disturbance force. Obtain.

[0048] According to the present invention, the vehicle speed control system can make the vehicle speed tracking control simpler and easier to debug, and can effectively correct the tracking error between the desired vehicle speed and the actual vehicle speed, while ensuring the anti-interference capability of the vehicle speed control; at the same time, it incorporates the influence of unknown external interference on the actual control output speed during the operation of commercial vehicles, improves the control accuracy of the vehicle speed control system, ensures the stability and efficiency of vehicle speed control, and reduces the dependence on MCU computing power and computing costs. Attached Figure Description

[0049] Figure 1 This is a basic step diagram of the cruise speed control method provided in the embodiment of the present invention;

[0050] Figure 2 This is a detailed step diagram of the cruise speed control method provided in the embodiment of the present invention;

[0051] Figure 3 This is a flowchart of the training process for the control law prediction model provided in an embodiment of the present invention. Detailed Implementation

[0052] The proposed solution of this invention designs a second-order system with respect to the state variables actual vehicle speed x1 and actual acceleration x2, and designs a second-order sliding model controller based on this system, incorporating the desired acceleration. and expected variable acceleration A first-order low-pass filter with parameters is used to calculate the vehicle speed control law that conforms to the current state of predictive cruise. On the other hand, based on the second-order slip model controller, a control law prediction model is constructed to predict the switching control law that can counteract the effects of unknown disturbances during adaptive cruise. The two control laws are integrated, and the target driving force is obtained through an integrator. The corresponding control commands are then applied to the brake and drive modules to adjust and correct the actual vehicle speed.

[0053] The specific implementation of the present invention will now be described in detail with reference to the accompanying drawings.

[0054] The cruise speed control method based on sliding mode control provided by this invention, such as... Figure 1 As shown, it includes the following steps:

[0055] Step S100: Acquire driving data and construct a second-order system with actual vehicle speed x1 and actual acceleration x2 as state variables; wherein, the driving data includes vehicle driving parameters and desired indicators, and the desired indicators include desired speed v. d Expected acceleration Expected variable acceleration

[0056] This step is detailed as follows: Figure 2 As shown, it includes: step S201 acquiring vehicle driving parameters from sensors, and step S202 predicting expected indicators;

[0057] In step S201, the sensor acquires vehicle driving parameters and internal operating parameters. The specific vehicle driving parameters include the current vehicle speed v and acceleration. Vehicle load Vehicle net weight (m), slope information (α), atmospheric pressure (P), temperature (T), internal operating parameters including engine torque (T) tq Engine speed n;

[0058] Based on this, the air density ρ can be calculated from the atmospheric pressure P and temperature T measured by the sensor, and the engine torque T measured by the sensor can also be calculated. tq Engine speed n and default transmission system power loss P T The mechanical efficiency η of the transmission system can be calculated.

[0059] Step S202 achieves the predicted expected index, specifically referring to: calculating the expected index based on vehicle driving parameters and the vehicle dynamics model. The expected index specifically includes: the expected speed v. d Expected acceleration Expected variable acceleration

[0060] The vehicle dynamics model used in this step is represented as follows:

[0061]

[0062] Among them, the rolling resistance coefficient f and the air resistance coefficient C D Air density ρ, transmission system mechanical efficiency η, frontal area A, gearbox ratio i g The main reducer gear ratio i0 and wheel radius r are vehicle information and default coefficients; current vehicle speed v and vehicle load are also parameters. Vehicle net weight m, slope information α, and engine torque T tq The parameter information is obtained in step S201.

[0063] Step S203: Construct a second-order system based on the actual vehicle speed and actual acceleration;

[0064] The construction process includes:

[0065] 1) Establish the vehicle dynamics model:

[0066] Due to external interference, the vehicle dynamics model involved in this step includes unknown external disturbance forces. Will Figure 2 The parameters obtained in steps S201 and S202 are used as input parameters, including: vehicle load. Vehicle net weight m, slope information α, rolling resistance coefficient f, air resistance coefficient C D Frontal area A, air density ρ, current vehicle speed v and acceleration Combined with unknown external interference forces The construction of a vehicle dynamics model can be represented as follows:

[0067]

[0068] Where: M is the total mass of the vehicle, Where m is the vehicle's additional load capacity, and F is the vehicle's unloaded mass. f For rolling resistance, F i For slope resistance, F w For air resistance, F t Where f is the driving force, f is the rolling resistance coefficient, and C is the rolling resistance coefficient. D Where A is the air resistance coefficient, ρ is the frontal area, and F is the air density. c The unknown external interference force experienced by the vehicle.

[0069] Among them, the unknown external disturbance force F experienced by the vehicle c It can be obtained through prediction models based on the current driving status.

[0070] 2) Construct a second-order system, namely: using the current vehicle speed v and acceleration... Given the system state, the state-space equation with the derivative of the driving force as input is expressed as:

[0071] and:

[0072]

[0073]

[0074] Where M is the total mass of the vehicle, Where m is the vehicle's additional load capacity, and F is the vehicle's unloaded mass. f For rolling resistance, F i For slope resistance, F w For air resistance, F c For unknown external interference forces on the vehicle, F t Where f is the driving force, f is the rolling resistance coefficient, and C is the rolling resistance coefficient. D Let A be the air resistance coefficient, ρ be the frontal area, ρ be the air density, and Q be the interference term.

[0075] Traditional commercial vehicle cruise speed control systems typically employ first-order sliding mode control. While this can promptly correct the actual vehicle speed to reduce the error between the actual and desired speeds, it suffers from significant fluctuations in the actual output speed and a large workload for adjusting speed errors. This step constructs a second-order system for the sliding mode controller in step S110, effectively addressing this problem.

[0076] Step S110: Construct a sliding mode controller based on the second-order system regarding the vehicle speed error e1 and the sliding surface s;

[0077] The construction process includes:

[0078] 1) Define the vehicle speed error as: e1 = x1 - x 1d ,

[0079] Where x1 is the actual vehicle speed, x 1d The desired vehicle speed, i.e., x 1d =v d ;

[0080] The derivative of the vehicle speed error is: Right now: For desired acceleration

[0081] 2) Define the sliding mode function as:

[0082] Taking its derivative, we get:

[0083] at this time, For desired variable acceleration

[0084] 3) Define the Lyapunov function V1 as:

[0085] Taking the derivative with respect to V1, we construct a sliding mode controller, the expression of which is:

[0086] Where s is the sliding surface. This is the derivative of the sliding surface.

[0087] In this step, the Lyapunov function is designed and differentiated to ensure that the tracking speed error and its derivative converge to 0.

[0088] In summary, the sliding mode controller constructed in this step uses the vehicle speed tracking error e1 and the system state of the second-order system as sliding variables, with the sliding surface s, used to realize the state variables vehicle speed v and acceleration. The control law that reaches s=0 in a finite amount of time.

[0089] Step S121: Calculate the desired control law based on the vehicle speed error e1 and the sliding surface s.

[0090] First, such as Figure 2 As shown in step S221, a novel sliding mode control is constructed based on Dynamic Surface Control (DSC), as detailed below:

[0091] To avoid the differential "explosion problem," we first assume a first-order inertial element that satisfies the following conditions:

[0092] Where α1 is the input quantity of the first-order inertial element and T2 is the random coefficient;

[0093] Based on this, the errors e2 and γ2 can be defined as follows:

[0094]

[0095] Where e2 is the actual acceleration value x2 and the desired acceleration value. The error, γ2 is The error between the input quantity α1 and the first-order inertial element;

[0096] The actual acceleration value x2 is obtained as: x2=e2+γ2+α1;

[0097] Substituting the actual acceleration value x2 into the derivative of the Lyapunov function: The generated expression is:

[0098]

[0099] Based on this, a virtual control law α1 is designed.

[0100] The derivative expression of the Lyapunov function is updated as follows:

[0101]

[0102] The expected control law is designed as follows:

[0103]

[0104] Where k1 and k2 are random coefficients, For expected control laws;

[0105] Then: Expected control law The calculation method is expressed as follows:

[0106]

[0107] This invention also provides a verification control law. Reasonable methods:

[0108] Expectation control law Substituting the calculation method into the novel sliding mode controller, we can obtain...

[0109]

[0110] It is evident that the total Lyapunov V2 can reach a certain level under a certain amount of e1γ2. The optimal control law is effective because it ensures that the system enters the sliding mode within a finite time.

[0111] In addition, the sliding mode controller constructed in this invention can also be used to calculate control law compensation under the influence of unknown external disturbances, i.e., to switch control laws.

[0112] Step S122: Load the control law prediction model, based on the sliding surface s and the sliding surface derivative. Obtain the switching control law

[0113] In step S121, the desired control law under ideal conditions, i.e., without considering unknown disturbances, is obtained. However, during actual vehicle operation, there are still unknown disturbances causing actual impact values. In this invention, a control law compensation value caused by unknown disturbances is defined, i.e., the switching control law. This value is predicted using a BP fuzzy neural network, mimicking the error backpropagation method of a BP network.

[0114] First, construct the control law prediction model, including the following steps:

[0115] Define a BP fuzzy neural network, with the following specific structure.

[0116] First layer:

[0117]

[0118] Second layer:

[0119] Calculate the membership degree of each input quantity to the fuzzy set of each linguistic variable value, and select a Gaussian function as the membership function:

[0120]

[0121]

[0122] Third layer:

[0123] Each node in the fuzzy inference layer represents a fuzzy rule, and the output value of each node represents the excitation strength of each fuzzy rule, expressed as follows:

[0124]

[0125] In the formula, i1∈{1,2,...,m1},i2∈{1,2,...,m2},...,i n ∈{1,2,...,m n},

[0126]

[0127] Fourth layer:

[0128] This layer is a normalization layer, and the output uses the Madmdani fuzzy rule. The expression for this layer is:

[0129]

[0130] Fifth floor:

[0131] The specific formula for the system output of the fuzzy neural network defuzzification is as follows:

[0132]

[0133]

[0134] In the formula, j = 1, 2, ..., m; i = 1, 2, ..., r.

[0135] Next, the training samples for the BP fuzzy neural network are determined. In this invention, the training samples are defined as the sliding surface s and the derivative of the sliding surface.

[0136] The control law prediction model is set as a two-input single-output system, with input parameter x. i Including the sliding surface s and the derivative of the sliding surface make x represents i The j-th language variable, i.e., the input parameter acquired through the sensor for the j-th time, where i = 1, 2; j = 1, 2, ..., m i TS blurry output y i To switch control laws The specific implementation process is as follows:

[0137] First, set the estimated value of the unknown external interference term as follows:

[0138]

[0139] Where X and ω are the input parameters of the prediction algorithm. The input parameters obtained for the j-th time corresponding to parameter X include the sliding surface s and the derivative of the sliding surface. ω T ω represents the optimal connection weights for the control law prediction model;

[0140] Fuzzification is achieved using product reasoning, and the reasoning result is:

[0141]

[0142] Using a single-valued fuzzy converter to calculate y i That is, the function value of the x-coordinate corresponding to the maximum value of the membership function:

[0143] The control law prediction model for j = 1, 2, ..., m performs a union operation on all fuzzy rules, and then uses the average defuzzifier control law prediction model to obtain the output of the fuzzy system:

[0144]

[0145] Secondly, during the training process, according to the control law y i The optimal center value c is obtained through error training. ij Width σ ij and connection weight w ij This leads to the modification of the next output control law, adjusting the control strength of the control law to suppress unknown disturbances.

[0146] like Figure 3 As shown, the training process of the BP fuzzy neural network is as follows:

[0147] Set the error cost function as follows:

[0148]

[0149] In the formula, y id For the desired output, y i This is the actual output.

[0150] Give the error backpropagation algorithm

[0151]

[0152] Then obtain

[0153]

[0154] Further calculation yields

[0155]

[0156]

[0157]

[0158] Therefore, the first gradient can be obtained as

[0159]

[0160]

[0161] Finally, the parameter adjustment learning algorithm is given as follows:

[0162]

[0163] In the formula, k is the number of iterations, λ is the width, Let λ be the learning efficiency of the center value and ρ be the connection weight, and λ be the learning efficiency of the connection weight. ρ>0. By training the network, each step is taken along the negative direction of the gradient, and this process is repeated until the network output error is minimized.

[0164] Through the above training process, the optimal model parameters are obtained, including the optimal center value c. ij Width σ ij and connection weight w ij ;

[0165] After training, the control law prediction model is constructed based on the optimal model parameters.

[0166] In this step, based on the optimal parameters obtained from training the control law prediction model and combined with the definition of the disturbance force, the generated disturbance force F is predicted. c The results are then applied to step S203 to implement the second-order system.

[0167] In this invention, such as Figure 2 As shown in step S233, load the control law prediction model, and input the sliding surface s and the sliding surface derivative in the sliding mode controller as in step S210. Switch control law based on prediction results Switching control laws Ultimately, by participating in the controller's operation, external disturbances during driving can be further considered when controlling the cruise speed, thereby improving the control accuracy of the vehicle speed control system.

[0168] Step S130: As Figure 2 As shown in step S233, the integrated desired control law is... and switching control laws Integrate again to obtain the driving force F. t According to the driving force F t Send control commands to the commercial vehicle's drive and brake systems;

[0169] Among them, the driving force F is obtained t Refers to: the switching control law and expected control law The summation is performed, and the overall control law is applied to the accumulated value. The integral is performed, and the result of the integral is the driving force F. t .

[0170] Compared with traditional cruise speed control methods, this invention proposes a second-order sliding mode speed controller, which simplifies and facilitates speed tracking control in the vehicle speed control system. It also effectively corrects the tracking error between the desired and actual vehicle speeds, keeping the tracking error within the sliding surface and ensuring the anti-interference capability of the speed control. Dynamic surface control (DSC) is used to improve the sliding mode speed controller, reducing computational load and increasing the speed error correction rate. Considering the impact of unknown external disturbances on the actual control output speed during commercial vehicle operation, a BP fuzzy neural network is used to optimize the sliding mode speed controller. Adaptive compensation for the control effects caused by unknown external disturbances is proposed, ensuring the tracking error does not deviate from the sliding surface, thus improving the control accuracy of the vehicle speed control system and guaranteeing its stability.

[0171] The above-disclosed embodiments are merely a few specific examples of the present invention. However, the present invention is not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. A cruise speed control method based on sliding mode control, characterized in that, Includes the following steps: Acquire driving data and construct a system based on actual vehicle speed. With actual acceleration This is a second-order system with state variables; wherein the driving data includes vehicle driving parameters and desired indicators, the desired indicators including desired speed. Expected acceleration Expected variable acceleration The acquisition of driving data includes: obtaining vehicle driving parameters from sensors and predicting desired indicators; the vehicle driving parameters include the current vehicle speed. With acceleration Vehicle load capacity Vehicle net weight Slope information Atmospheric pressure ,temperature Engine torque and engine speed The predicted expected index refers to the expected index calculated based on the vehicle driving parameters and the vehicle dynamics model. The second-order system is represented as: ,and: , , , , The vehicle dynamics model is represented as follows: , in, For actual vehicle speed, For actual acceleration, For rolling resistance coefficient, For vehicle unloaded weight, For interference items, For driving force, For air drag coefficient, For air density, For windward area, For slope information, For the total mass of the vehicle, For the vehicle's additional load capacity, For rolling resistance, For slope resistance, For air resistance, For unknown external interference forces on the vehicle, For the mechanical efficiency of the transmission system, For the gear ratio of the transmission, Main reducer transmission ratio, For wheel radius, For current vehicle speed, This refers to engine torque; Based on the second-order system, construct information about vehicle speed error. With sliding surface A sliding mode controller; the sliding mode variable of the sliding mode controller is defined as the vehicle speed error. The system state of the second-order system is related to the sliding surface s, which is used to realize the state variable vehicle speed. and acceleration The control law that reaches s=0 in a finite time; Based on vehicle speed error and sliding surface Calculate the expected control law ; Load the control law prediction model, based on the sliding surface and sliding surface derivative Obtain the switching control law ; Integrating the desired control law and switching control laws Obtain the overall control law The overall control law described by the integral Find the driving force According to the driving force Send control commands to the commercial vehicle's drive and brake systems; Among them, the vehicle speed error is constructed based on the second-order system. With sliding surface The sliding mode controller includes: Define the vehicle speed error as: , in, This is the actual vehicle speed. For the desired vehicle speed, and ; The sliding mode function is defined based on the actual vehicle speed and the speed error, and is expressed as follows: ,and ; Lyapunov function is defined based on the sliding mode function. , is represented as: ; For the Lyapunov function The derivative is expressed as: ; Based on vehicle speed error and sliding surface Calculate the expected control law ,include: Construct a first-order low-pass filter with desired acceleration and desired variable acceleration as state variables, expressed as: ,in, For the input quantity of the first-order inertial element, These are random coefficients; The first-order low-pass filter and the vehicle speed error are used together. and sliding surface Update the Lyapunov function The expression for differentiation is: ； in, For random coefficients, For expected control laws; Obtain the desired control law The calculation method is expressed as follows: 。 2. The cruise speed control method according to claim 1, characterized in that, Before loading the control law prediction model, the control law prediction model is constructed, including: Define a BP fuzzy neural network; Train the BP fuzzy neural network to obtain optimal model parameters, which include: optimal center value. ,width and connection weights ; Based on the optimal model parameters, a control law prediction model is constructed.

3. The cruise speed control method according to claim 2, characterized in that, The control law prediction model is a two-input single-output system, with input parameters... Including sliding surfaces With the derivative of the sliding surface The output parameter is the switching control law. .

4. The cruise speed control method according to claim 3, characterized in that, Based on the optimal model parameters, the predicted interference force is achieved. ,include: Set up an algorithm for predicting the estimated value of unknown external interference: ； in, Input parameters to the prediction algorithm, The input parameters obtained for the j-th time corresponding to parameter X include the sliding surface. With the derivative of the sliding surface , for The corresponding optimal connection weights for the control law prediction model.

5. The cruise speed control method according to claim 4, characterized in that, The unknown external disturbance force experienced by the vehicle is predicted by the disturbance force. Obtain.

6. The cruise speed control method according to claim 1, characterized in that, Obtain the driving force Refers to: the switching control law and expected control law Accumulate the sums and apply them to the overall control law. The integral is performed, and the result of the integral is the driving force. .

Images