Vehicle speed control method of hybrid electric vehicle

Abstract

The invention relates to a vehicle speed control method of a hybrid electric vehicle. The method comprises the following steps: establishing a vehicle dynamics model; designing an improved disturbance observer according to the vehicle dynamics model, and obtaining a disturbance estimation value; a speed control law is designed, the speed control law comprises feedforward control and feedback control, and control input of the improved disturbance observer is determined according to the disturbance estimation value and the speed control law; the motor torque is designed based on the control input of the improved disturbance observer, and the motor torque and the engine torque are dynamically adjusted based on a PID controller of a neural network, so that the vehicle speed dynamically approaches the target vehicle speed. The torque of the vehicle is dynamically adjusted through PID control based on the neural network, and the vehicle speed can dynamically approach the target vehicle speed.

Description

Technical Field

[0001] The present application relates to the technical field of electric vehicle speed control, and in particular to a method for controlling the speed of a hybrid electric vehicle. Background Art

[0002] Adaptive cruise control (ACC) was proposed as the core technology of autonomous driving. Although related research has a long history, longitudinal vehicle speed control is an important research content in the automotive industry because it is closely related to the challenges faced by fully automated vehicle control systems, such as safety, fuel economy, traffic congestion, and ride comfort. At the same time, automobile manufacturers have added various additional features to vehicle design, such as highway driving assistance systems and emergency braking systems. In order to meet these market demands, it is crucial to design and develop a longitudinal vehicle control algorithm that is easy to implement, efficient to develop, and meets the control performance requirements for mass production.

[0003] Generally, the longitudinal vehicle speed controller consists of two layers: an upper controller and a lower controller. The upper controller determines the speed profile that can be obtained from the speed set by the driver or from the speed planning algorithm that considers fuel economy, traffic congestion and safety. The role of the lower controller is to track the speed profile required by the upper controller while adjusting the torque demand of the propulsion system.

[0004] For the lower-level controller. To meet the performance specifications and develop an effective control system, it is necessary to obtain zero steady-state error and well suppress external disturbances. At the same time, the impact of model uncertainty should be minimized to provide the required acceleration / deceleration behavior to ensure occupant comfort. To meet these goals, many methods for longitudinal vehicle control have been proposed. In the automotive industry, the classic method is a PID controller because of its simple structure, robustness and stability. After that, many studies have been conducted on PID-based learning algorithms, sliding mode control, and artificial intelligence methods. However, most of these methods need to be verified in vehicle experiments under different driving conditions. In addition, the algorithms need to be implemented on real-time controllers.

[0005] In the related art, in order to improve the performance of the lower-level controller, a gain scheduling scheme is adopted, but this requires linearization of the nonlinear system to obtain the gain of each equilibrium point. Research based on optimal control theory, such as linear quadratic control and model predictive control, is used to design longitudinal vehicle control. However, these methods require an accurate vehicle model. The inaccuracy of the model and the complexity and variability of the vehicle power system may lead to certain uncertainties in the established model, thereby affecting the performance and control effect of the disturbance observer. In addition, real-time performance limitations are one of the key indicators of the control system, and the disturbance observer and its related controllers may require faster computing speeds and responsiveness to meet the requirements in actual driving conditions. Although the disturbance observer can compensate for model uncertainty and external disturbances to a certain extent, its performance may be affected by changes in the external environment, resulting in challenges to the stability of the control system.

[0006] Therefore, it is necessary to provide a new technical solution to improve one or more problems existing in the above solutions.

[0007] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present application, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0008] The purpose of the present application is to provide a vehicle speed control method for a hybrid electric vehicle, thereby overcoming one or more problems caused by limitations and defects of related technologies at least to a certain extent.

[0009] According to a vehicle speed control method of a hybrid electric vehicle provided in an embodiment of the present application, the method includes:

[0010] Build vehicle dynamics models;

[0011] Designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value;

[0012] Designing a speed control law, wherein the speed control law includes feedforward control and feedback control, and determining a control input of an improved disturbance observer according to the disturbance estimation value and the speed control law;

[0013] The motor torque is designed based on the control input of the improved disturbance observer, and the motor torque and the engine torque are dynamically adjusted by a PID controller based on a neural network so that the vehicle speed dynamically approaches the target vehicle speed.

[0014] In the embodiment of the present application, the expression of the vehicle dynamics model is:

[0015]

[0016] The expression of the transfer function is:

[0017]

[0018] By rewriting Formula 1 through the transfer function, the expression of the vehicle dynamics model is obtained as follows:

[0019]

[0020] Where v(s) represents the Laplace transform of vehicle speed, i.e., the complex frequency representation of the change of vehicle speed over time, m represents the vehicle weight, s represents the complex frequency variable, and w r represents the rolling resistance coefficient, f represents the tire load, ρ represents the air density, C d represents the air resistance coefficient, A represents the frontal area of ​​the vehicle, v represents the vehicle speed, g represents the acceleration due to gravity, and θ represents the slope angle.

[0021] In an embodiment of the present application, the step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value includes:

[0022] The expression of the improved disturbance observer is:

[0023] y=G nominal (s)u(s)+y d (4);

[0024] In the formula, G nominal represents the nominal model, u(s) represents the control input of the improved disturbance observer, and y d represents the disturbance response;

[0025] Rewrite formula 4 as follows:

[0026]

[0027] In the formula, G(s) represents the actual control system, C(s) represents the adjustment function, d(s) represents the input disturbance, and n(s) represents the measurement noise;

[0028] Under low frequency conditions, assuming Q(s)≈1,n(s)≈0, Formula 5 can be rewritten as:

[0029]

[0030] Where r(s) represents the Laplace transform of the input disturbance.

[0031] In an embodiment of the present application, the step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value further includes:

[0032] Design a low-pass filter, the expression of which is as follows:

[0033]

[0034] Where τ represents the time constant.

[0035] In an embodiment of the present application, the step of designing a speed control law, wherein the speed control law includes feedforward control and feedback control, and determining a control input of an improved disturbance observer according to the disturbance estimation value and the speed control law, includes:

[0036] According to the improved disturbance observer, a vehicle speed control system is obtained, and the expression of the vehicle speed control system is:

[0037] v=G nominal (s)u(s)(9);

[0038] In the formula, v represents the vehicle speed;

[0039] Through Formula 2 and Formula 9, the specific expression of the control input u(s) is obtained as follows:

[0040]

[0041] In the formula, u r represents the concrete expression of u(s), Indicates the acceleration of the vehicle;

[0042] Through the vehicle speed and target speed, the tracking error of the vehicle speed is obtained as:

[0043] e=vv re (11);

[0044]

[0045] Where, e represents the tracking error of vehicle speed, v re Indicates the target vehicle speed. represents the first-order derivative of the tracking error, that is, the error change rate, λ>0;

[0046] Furthermore, the tracking error of acceleration is:

[0047] a x =a x_r +λ(v re -v)(13);

[0048] In the formula, a x represents the tracking error of acceleration, a x_ represents the target acceleration;

[0049] According to Formula 10, Formula 11, Formula 12 and Formula 13, the specific expression of the control input is further obtained as follows:

[0050] u r =ma x_r +mλ(v re -v) (14);

[0051] In the formula, ma x_r represents the feedforward control term, mλ(v re -v) indicates feedback control item;

[0052] The control input of the improved disturbance observer is:

[0053]

[0054] In an embodiment of the present application, the step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on a neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed includes:

[0055] The expression of the designed motor torque is:

[0056]

[0057] Where, T e_err represents the engine torque error, T e_err =T eb -T e , r gearb represents the gear ratio of the powertrain, T eb Indicates the engine braking torque, T e The engine outputs effective torque, referred to as engine torque, and R represents the rotation radius of the tire.

[0058] In an embodiment of the present application, the step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on a neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed, further includes:

[0059] Design a neural network. The speed control output expression of the neural network is as follows:

[0060] y m =w1h1+w2h2+...+w m h m (17);

[0061] Where w is the weight vector, w=[w1,w2,...,w m ]T , w m represents the mth weight vector, radial basis vector H = [h1,h2,...,h m ] T ,h m Represents the output value of the mth neuron;

[0062] The expression of the performance index function of the neural network is as follows:

[0063]

[0064] In the formula, y o (k) represents the value of the output of the neural network at the kth moment, y m (k) represents the output value of the mth layer;

[0065] The sensitivity of the object's output to changes in the control input is expressed by the Jacobian matrix as follows:

[0066]

[0067] c ij represents the weight connecting neurons i and j. Δu(k) represents the change of the control variable at time k, b j represents the bias term of the jth neuron.

[0068] In an embodiment of the present application, the step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on a neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed, further includes:

[0069] The motor torque and the engine torque satisfy the total driving torque, and the expression of the total driving torque is as follows:

[0070] T total =T m +T e (20);

[0071] Where, T total Indicates the total driving torque.

[0072] In an embodiment of the present application, the step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on a neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed, further includes:

[0073] The expression for designing PID controller is as follows:

[0074]

[0075] In the formula, e(t) represents the deviation between the actual output and the expected output, K p is the proportional control coefficient, K i Indicates the integral control coefficient, K d represents the differential control coefficient, u(t) represents the control output of the PID controller, and y o (t) represents the reference torque of the motor, and y(t) represents the current output torque of the motor.

[0076] In an embodiment of the present application, the step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on a neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed, further includes:

[0077] The PID controller is optimized by a neural network to obtain a PID controller based on a neural network. The expression of the PID controller based on a neural network is as follows:

[0078]

[0079] In the formula, Δu(t) represents the change of the control variable at time t, e error (k) represents the value of the error at time k, r(k) represents the value of the reference input at time k, and y o (k) represents the value of the output of the neural network at time k;

[0080] The input variables of the neural network based PID controller are:

[0081]

[0082] In the formula, a(1) represents the adjustment coefficient of the neural network in the input layer, a(2) represents the adjustment coefficient of the neural network in the hidden layer, a(3) represents the adjustment coefficient of the neural network in the output layer, and e error (k-1) represents the value of the error at time k-1, e error (k-2) represents the value of the error at time k-2;

[0083] The control law is:

[0084]

[0085] In the formula, u(k) represents the control input at the current time k, u(k-1) represents the control input at the previous time k-1, K represents the set of control coefficient gains, and X represents the state variable of the control system;

[0086] The expression of the performance index evaluation function of the engine system is as follows:

[0087]

[0088] According to the evaluation function, the neural network is used to optimize the PID controller;

[0089] Using the control output of the neural network-based PID controller as the control input of the engine system to achieve control of the engine torque;

[0090] Based on the control of the motor torque and the generator torque, the vehicle speed is controlled.

[0091] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0092] In one embodiment of the present application, the above method is used to obtain disturbance estimation values ​​through an improved disturbance observer, and robust vehicle speed control is achieved, which can effectively estimate and compensate for system uncertainty and internal and external disturbances, so as to better adapt to the uncertainty and changes in various environments and working conditions, and facilitate the subsequent control of vehicle speed; at the same time, the speed control law is designed through feedforward control and feedback control, so that the target vehicle speed can be tracked more accurately, and real-time rapid response can be made when the vehicle speed changes, thereby improving the dynamic performance and driving comfort of the vehicle. Moreover, the torque of the vehicle is dynamically adjusted through PID control based on a neural network, so that the vehicle speed can be dynamically close to the target vehicle speed.

[0093] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0095] Figure 1 A flowchart schematically illustrates a method for controlling the vehicle speed of a hybrid electric vehicle in an exemplary embodiment of the present application;

[0096] Figure 2 A block diagram schematically shows a vehicle speed control system of a hybrid electric vehicle in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0097] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0098] In addition, the accompanying drawings are only schematic illustrations of the present application and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0099] In this exemplary embodiment, a vehicle speed control method for a hybrid electric vehicle is first provided. Figure 1 As shown in , the method may include: steps S101 to S103.

[0100] Wherein, step S101: establishing a vehicle dynamics model.

[0101] Step S102: designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value.

[0102] Step S103: designing a speed control law, which includes feedforward control and feedback control, and determining a control input of an improved disturbance observer according to the disturbance estimation value and the speed control law.

[0103] Step S103: Designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on the neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed.

[0104] In one embodiment of the present application, the above method is used to obtain disturbance estimation values ​​through an improved disturbance observer, and robust vehicle speed control is achieved, which can effectively estimate and compensate for system uncertainty and internal and external disturbances, so as to better adapt to the uncertainty and changes in various environments and working conditions, and facilitate the subsequent control of vehicle speed; at the same time, the speed control law is designed through feedforward control and feedback control, so that the target vehicle speed can be tracked more accurately, and real-time rapid response can be made when the vehicle speed changes, thereby improving the dynamic performance and driving comfort of the vehicle. Moreover, the torque of the vehicle is dynamically adjusted through PID control based on a neural network, so that the vehicle speed can be dynamically close to the target vehicle speed.

[0105] Next, we will refer to Figure 1 to Figure 2 Each step of the above method in this example implementation is described in more detail.

[0106] In step S101, before establishing the vehicle dynamics model, it is necessary to obtain the relevant parameters of the vehicle. The relevant parameters such as the vehicle mass, wheelbase, engine output, etc. in the specification table provided by the manufacturer are different from the data values ​​obtained from the field test. In this application, the theoretical values ​​provided by the manufacturer are used as the parameter values ​​designed in this application.

[0107] The theoretical value includes the vehicle weight m, m=2300 kg.

[0108] In addition to the vehicle weight parameters, there are also wheel rotation speed, vehicle speed and engine torque. The present application measures the wheel rotation speed by installing a vehicle speed sensor near the wheel hub of the vehicle; the vehicle speed sensor is installed between the vehicle's transmission system to measure the speed of the entire vehicle (and the vehicle speed); the engine torque sensor is installed in the engine control unit to measure the engine torque.

[0109] Based on the above reference, the force analysis of the vehicle is carried out, ignoring the dynamic effects of the powertrain and tire slip, and the expression of the vehicle dynamics model is:

[0110]

[0111] In the formula, F drive is the total tire force, F aero Indicates air resistance, F roll represents rolling resistance and G represents the gravity of the vehicle.

[0112] In one embodiment, the vehicle dynamics model is expressed as:

[0113]

[0114] Ignoring the dynamic models of the engine and motor, Equation 1 can be rewritten using the transfer function. The expression of the transfer function is:

[0115]

[0116] By rewriting Formula 1 through the transfer function, the expression of the vehicle dynamics model is obtained as follows:

[0117]

[0118] Where v(s) represents the Laplace transform of vehicle speed, i.e., the complex frequency representation of the change of vehicle speed over time, m represents the vehicle weight, s represents the complex frequency variable, and w rrepresents the rolling resistance coefficient, f represents the tire load, ρ represents the air density, C d represents the air resistance coefficient, A represents the frontal area of ​​the vehicle, v represents the vehicle speed, g represents the acceleration due to gravity, and θ represents the slope angle.

[0119] It is understandable that the vehicle load cannot be accurately modeled due to the uncertainty of the vehicle dynamics model and unmeasurable information. In this application, the changes in the vehicle load and vehicle weight are combined into a disturbance part. That is, the disturbance observer is used as feedback to measure the uncertainty unknown disturbance term in the vehicle system, and the disturbance information is compensated to the control system.

[0120] Secondly, the main body of the present application is a hybrid system of an electric motor and an internal combustion engine (i.e., a vehicle dynamics model), both of which provide power to the transmission system at the same time. The total tire force of the hybrid transmission system can be equivalent to the sum of the engine torque and the motor torque.

[0121]

[0122] Where R is the rolling radius of the tire, r gear represents the gear ratio of the powertrain, T e represents the engine torque, T m Indicates the motor torque.

[0123] According to the design of the vehicle dynamics model of the electric vehicle and the acquisition of vehicle parameters, the obtained parameter information is transmitted to the vehicle speed controller, the disturbance observer and the torque controller (ie, the torque) respectively.

[0124] Furthermore, after the vehicle dynamics model is established, a vehicle speed controller is designed, and the step of designing the vehicle speed controller includes step S102 and step S103.

[0125] In step S102, the electric vehicle will be subject to different degrees of internal and external disturbances due to the external environment and driving conditions during driving. The longitudinal vehicle speed control (i.e., vehicle speed control) needs to stipulate that the vehicle maintains good tracking performance under different driving conditions. In order to ensure the road speed limit specification, a disturbance observer control system is added to effectively reduce external disturbances and the uncertainty of the vehicle dynamics model caused by the change of the vehicle's own weight.

[0126] The step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value includes steps S1021 and S1022.

[0127] Step S1021: The expression of the improved disturbance observer is:

[0128] y=G nominal (s)u(s)+yd (4);

[0129] In the formula, G nominal represents the nominal model, u(s) represents the control input of the improved disturbance observer, and y d represents the disturbance response;

[0130] Rewrite formula 4 as follows:

[0131]

[0132] In the formula, G(s) represents the actual control system, C(s) represents the adjustment function, d(s) represents the input disturbance, and n(s) represents the measurement noise;

[0133] Under low frequency conditions, assuming Q(s)≈1,n(s)≈0, Formula 5 can be rewritten as:

[0134]

[0135] Where r(s) represents the Laplace transform of the input disturbance.

[0136] It can be understood that the actual closed-loop system of the disturbance observer (DOB) constitutes a closed-loop system on the nominal model and can be combined with any external loop controller, while the design of the filter can achieve the stability of the disturbance observer loop system, compensate for the uncertainty factors in the system, and improve the control performance of the system.

[0137] Step S1022: Design a low-pass filter. The expression of the low-pass filter is as follows:

[0138]

[0139] Where τ represents the time constant.

[0140] It can be understood that the low-pass filter is a third-order low-pass filter, and its model is as shown in Formula 7. From the disturbance estimate of DOB, the expression of the disturbance estimate of DOB is:

[0141]

[0142] In the formula, when the nominal model is close to the actual control system G(s), that is, Formula 27 is further expressed as:

[0143]

[0144] Where d is the actual disturbance.

[0145] If the time constant of the designed low-pass filter is small enough, that is, Q(s)≈1, the disturbance estimate is close to the actual disturbance. Furthermore, the expression of the nonlinear function is set as follows:

[0146]

[0147] In the formula, is the acceleration obtained from the acceleration sensor, and U is the coefficient matrix.

[0148] Step S1022: Input the disturbance response into the low-pass filter to obtain the disturbance estimation value of the disturbance observer. The expression of the disturbance estimation value is as follows:

[0149]

[0150] Where μ(s) represents the nonlinear function and d represents the actual disturbance.

[0151] In step S103, the vehicle speed control law is a set of rules for controlling the vehicle speed, so that the electric vehicle reaches the desired vehicle speed under different driving conditions. Therefore, the vehicle speed control law is set to coordinate the vehicle speed control. Further, the speed control law is designed, and the speed control law includes feedforward control and feedback control, and the step of determining the control input of the improved disturbance observer according to the disturbance estimation value and the speed control law includes step S1031.

[0152] Step S1031: According to the improved disturbance observer, the vehicle speed control system is obtained. The expression of the vehicle speed control system is:

[0153] v=G nominal (s)u(s)(9);

[0154] In the formula, v represents the vehicle speed;

[0155] Through Formula 2 and Formula 9, the specific expression of the control input u(s) is obtained as follows:

[0156]

[0157] In the formula, u r represents the concrete expression of u(s), Indicates the acceleration of the vehicle;

[0158] Through the vehicle speed and target speed, the tracking error of the vehicle speed is obtained as:

[0159] e=vv re (11);

[0160]

[0161] Where, e represents the tracking error of vehicle speed, vre Indicates the target vehicle speed. represents the first-order derivative of the tracking error, that is, the error change rate, λ>0;

[0162] Furthermore, the tracking error of acceleration is:

[0163] a c =a x_r +λ(v re -v)(13);

[0164] In the formula, a x represents the tracking error of acceleration, a x_ represents the target acceleration;

[0165] According to Formula 10, Formula 11, Formula 12 and Formula 13, the specific expression of the control input is further obtained as follows:

[0166] u r =ma x_r +mλ(v re -v) (14);

[0167] In the formula, ma x_r represents the feedforward control term, mλ(v re -v) indicates feedback control item;

[0168] The control input of the improved disturbance observer is:

[0169]

[0170] It can be understood that the feedforward control term is set, which is partly determined by the traction force ma x_r As a speed control for the vehicle, the reverse control item uses the difference between the target speed and the vehicle speed as reverse control to make feedback adjustments to the actual control.

[0171] The speed controller design is to achieve accurate tracking and dynamic adjustment of vehicle speed. By adding an improved disturbance observer, the disturbance is estimated and compensated, the robustness of the control system is improved, and the calculated control input u (as the speed control signal) is applied to the vehicle dynamics model to adjust the vehicle speed so that the vehicle speed is close to its target speed v. re Ensure that the actual vehicle speed dynamically tracks the target speed, while compensating for disturbances caused by the external environment and modeling errors. In this part, the vehicle speed control signal u is obtained, providing basic information for the subsequent torque coordination control of the power system.

[0172] It should be noted that the present application utilizes an improved disturbance observer to achieve robust vehicle speed control. The disturbance observer can effectively estimate and compensate for system uncertainties and internal and external disturbances in the vehicle dynamics system, thereby improving the robustness and tracking performance of the control system composed of disturbance compensation and feedback control.

[0173] In step S104, the motor torque is designed based on the control input of the improved disturbance observer, and the motor torque and the engine torque are dynamically adjusted based on the PID controller of the neural network so that the vehicle speed dynamically approaches the target vehicle speed, including steps S1041 to S1045.

[0174] Step S1041: The designed expression of the motor torque is:

[0175]

[0176] Where, T e_err represents the engine torque error, T e_err =T eb -T e , r gearb represents the gear ratio of the powertrain, T eb Indicates the engine braking torque, T e It represents the effective torque output by the engine, that is, the engine torque, and R represents the rotation radius of the tire.

[0177] It can be understood that the control input u of the improved disturbance observer obtained in step S1021 comes from the motor and the engine, and the control input is used as a control signal to convert the driving force F drive (i.e., total tire force), in the nominal model, the engine torque T e , motor torque T m and the gear ratio r of the drivetrain gear Take the optimal value.

[0178] In actual systems, engine braking fuel consumption has a significant difference in engine efficiency. The engine efficiency affects the efficiency of the hybrid system, causing the motor torque to change. The converted driving force F drive The torque of the motor and the engine of the hybrid system are further distributed. Considering the dynamic change factors of the engine response and the control input, the expression of the motor torque can be the above formula 16.

[0179] Step S1042:

[0180] Design a neural network. The speed control output expression of the neural network is as follows:

[0181] y m=w1h1+w2h2+...+w m h m (17);

[0182] Where w is the weight vector, w=[w1,w2,...,w m ] T , w m represents the mth weight vector, radial basis vector H = [h1,h2,...,h m ] T ,h m Represents the output value of the mth neuron;

[0183] The expression of the performance index function of the neural network is as follows:

[0184]

[0185] In the formula, y o (k) represents the value of the output of the neural network at the kth moment, y m (k) represents the output value of the mth layer;

[0186] The sensitivity of the object's output to changes in the control input is expressed by the Jacobian matrix as follows:

[0187]

[0188] c ij represents the weight connecting neurons i and j. Δu(k) represents the change in the control variable at time k, b j It is understood that the neural network is a radial basis function (RBF) neural network, which is a single hidden layer feed-forward neural network, and uses a radial basis function as the activation function of the hidden layer neurons.

[0189] Step S1043: The motor torque and the engine torque satisfy the total driving torque, and the expression of the total driving torque is as follows:

[0190] T total =T m +T e (20);

[0191] Where, T total Indicates the total driving torque.

[0192] It is understandable that in a hybrid system, the engine torque T e The response speed is usually slower than the motor torque T mTherefore, a PID controller is needed to dynamically adjust the torque so that the combination of the engine torque and the motor torque can accurately satisfy the above formula 20.

[0193] Step S1044: The expression for designing the PID controller is as follows:

[0194]

[0195] In the formula, e(t) represents the deviation between the actual output and the expected output, K p is the proportional control coefficient, K i Indicates the integral control coefficient, K d represents the differential control coefficient, u(t) represents the control output of the PID controller, and y o (t) represents the reference torque of the motor, and y(t) represents the current output torque of the motor.

[0196] It is understandable that for hybrid electric vehicles, the PID controller controls the torque adjustment based on the output characteristics of the motor torque. Since factors such as load, operating conditions and disturbances may cause changes in torque demand and power system characteristics during vehicle operation, the parameters of the traditional fixed PID controller may be difficult to provide optimal performance in all cases. Therefore, this application uses a PID controller based on an RBF neural network to dynamically adjust the torque to adjust the problem that the engine torque responds slower than the motor torque. The expression of the PID controller is shown in the above formula 21.

[0197] Step S1045: Optimize the PID controller through the neural network to obtain a PID controller based on the neural network. The expression of the PID controller based on the neural network is as follows:

[0198]

[0199] In the formula, Δu(t) represents the change of the control variable at time t, e error (k) represents the value of the error at time k, r(k) represents the value of the reference input at time k, and y o (k) represents the value of the output of the neural network at time k;

[0200] The input variables of the neural network based PID controller are:

[0201]

[0202] In the formula, a(1) represents the adjustment coefficient of the neural network in the input layer, a(2) represents the adjustment coefficient of the neural network in the hidden layer, a(3) represents the adjustment coefficient of the neural network in the output layer, and e error (k-1) represents the value of the error at time k-1, e error(k-2) represents the value of the error at time k-2;

[0203] The control law is:

[0204]

[0205] In the formula, u(k) represents the control input at the current time k, u(k-1) represents the control input at the previous time k-1, K represents the set of control coefficient gains, and X represents the state variable of the control system;

[0206] The expression of the performance index evaluation function of the engine system is as follows:

[0207]

[0208] According to the evaluation function, the neural network is used to optimize the PID controller;

[0209] The control output of the neural network-based PID controller is used as the control input of the engine system to achieve control of the engine torque;

[0210] Vehicle speed is controlled based on the control of motor torque and generator torque.

[0211] It can be understood that the present application introduces RBF neural network to optimize PID controller, PID controller output u(k) and actual output error y o (k), together they constitute the input of the neural network. After the neural network is trained, the Jacobian matrix shown in the above formula 19 is output, which is used for parameter tuning of the PID controller. In order to avoid the cumulative error generated by the neural network, the PID controller adopts an incremental PID controller, that is, the above formula 22.

[0212] A neural network is added to the traditional PID controller to optimize the parameters of the PID controller and realize an incremental PID controller, so that it can gradually achieve zero steady-state error, meet the fast tracking response of the engine torque, and improve the driving comfort of hybrid vehicles.

[0213] Furthermore, the performance index is obtained according to the evaluation function, and the weight of the neural network is adjusted according to the gradient descent method to realize the dynamic adjustment of the parameters of the PID controller, so that the setting index finally meets the required requirements.

[0214] Dynamic torque adjustment is achieved by balancing the motor torque and the engine torque. In a hybrid system, it can be concluded from Formula 4 that the total driving force of the powertrain is provided by the motor torque and the engine torque. Combining the vehicle dynamics model of Formula 1 and Formula 26, after torque adjustment, the control law is dynamically optimized based on the improved disturbance observer and the neural network PID controller to achieve closed-loop control of the vehicle speed. At the same time, according to Formula 11 and Formula 12, through real-time monitoring of the vehicle speed deviation, the control system adjusts the torque output so that the vehicle speed dynamically approaches the target vehicle speed v re .

[0215] The present application dynamically adjusts vehicle speed and driving torque requirements, determines engine torque and motor torque, and controls the hybrid power system.

[0216] The performance of the vehicle speed control stability and the robustness of the control system to internal and external disturbances are evaluated on actual roads and roads with different slopes to assess the performance of the control system.

[0217] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary technical means in the art that are not disclosed in the present application.

Claims

1. A method for controlling the speed of a hybrid electric vehicle, characterized in that: include: Build vehicle dynamics models; Designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value; Designing a speed control law, wherein the speed control law includes feedforward control and feedback control, and determining a control input of an improved disturbance observer according to the disturbance estimation value and the speed control law; The motor torque is designed based on the control input of the improved disturbance observer, and the motor torque and the engine torque are dynamically adjusted by a PID controller based on a neural network so that the vehicle speed dynamically approaches the target vehicle speed.

2. The vehicle speed control method of a hybrid electric vehicle according to claim 1, characterized in that: The vehicle dynamics model is expressed as: The expression of the transfer function is: By rewriting Formula 1 through the transfer function, the expression of the vehicle dynamics model is obtained as follows: Where v(s) is the Laplace transform of vehicle speed, i.e., the complex frequency representation of the time-varying vehicle speed, m is the vehicle weight, s is the complex frequency variable, Wr is the rolling resistance coefficient, f is the tire load, ρ is the air density, and C is d represents the air resistance coefficient, A represents the frontal area of ​​the vehicle, v represents the vehicle speed, g represents the acceleration due to gravity, and θ represents the slope angle.

3. The vehicle speed control method of a hybrid electric vehicle according to claim 2, characterized in that: The step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value comprises: The expression of the improved disturbance observer is: y=G nominal (s)u(s)+y d (4); In the formula, G nominal represents the nominal model, u(s) represents the control input of the improved disturbance observer, and y d represents the disturbance response; Rewrite formula 4 as follows: In the formula, G(s) represents the actual control system, C(s) represents the adjustment function, d(s) represents the input disturbance, and n(s) represents the measurement noise; Under low frequency conditions, assuming Q(s)≈1, n(s)≈0, Formula 5 can be rewritten as: Where r(s) represents the Laplace transform of the input disturbance.

4. The vehicle speed control method of a hybrid electric vehicle according to claim 3, characterized in that: The step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value further includes: Design a low-pass filter, the expression of which is as follows: Where τ represents the time constant.

5. The vehicle speed control method of a hybrid electric vehicle according to claim 4, characterized in that: The step of designing an improved disturbance observer according to the vehicle dynamics model and obtaining a disturbance estimation value further includes: The disturbance response is input into a low-pass filter to obtain the disturbance estimate of the disturbance observer. The expression of the disturbance estimate is as follows: Where μ(s) represents the nonlinear function and d represents the actual disturbance.

6. The vehicle speed control method of a hybrid electric vehicle according to claim 5, characterized in that: The step of designing a speed control law, wherein the speed control law includes feedforward control and feedback control, and determining a control input of an improved disturbance observer according to the disturbance estimation value and the speed control law, comprises: According to the improved disturbance observer, a vehicle speed control system is obtained, and the expression of the vehicle speed control system is: v=G no min al (s)u(s) (9); In the formula, v represents the vehicle speed; Through Formula 2 and Formula 9, the specific expression of the control input u(s) is obtained as follows: In the formula, u r represents the concrete expression of u(s), Indicates the acceleration of the vehicle; Through the vehicle speed and target speed, the tracking error of the vehicle speed is obtained as: e=vv re (11); Where, e represents the tracking error of vehicle speed, v re Indicates the target vehicle speed. Represents the first-order derivative of the tracking error, that is, the error change rate, λ>0; Furthermore, the tracking error of acceleration is: and x =a x_r +λ(in re -v) (13); In the formula, a x represents the tracking error of acceleration, a x_r represents the target acceleration; According to Formula 10, Formula 11, Formula 12 and Formula 13, the specific expression of the control input is further obtained as follows: at r =ma x_r +mλ(v re -v) (14); In the formula, ma x_r represents the feedforward control term, mλ(v re -v) indicates feedback control item; The control input of the improved disturbance observer is:

7. The vehicle speed control method of a hybrid electric vehicle according to claim 6, characterized in that: The step of designing the motor torque based on the control input of the improved disturbance observer, and dynamically adjusting the motor torque and the engine torque based on the neural network PID controller so that the vehicle speed dynamically approaches the target vehicle speed includes: The expression of the designed motor torque is: Where, T e_err represents the engine torque error, T e_err =T eb -T e , r gearb represents the gear ratio of the powertrain, T eb Indicates the engine braking torque, T e It represents the effective torque output by the engine, referred to as engine torque, and R represents the rotation radius of the tire.

8. The vehicle speed control method of a hybrid electric vehicle according to claim 7, characterized in that: The step of designing the motor torque based on the control input of the improved disturbance observer and dynamically adjusting the motor torque and the engine torque based on the PID controller of the neural network so that the vehicle speed dynamically approaches the target vehicle speed also includes: Design a neural network. The speed control output expression of the neural network is as follows: y m =w1h1+w2h2+...+w m h m (17); Where w is the weight vector, w = [w1, w2, ..., w m ] T , w m represents the mth weight vector, radial basis vector H = [h1, h2, ..., h m ] T ,h m Represents the output value of the mth neuron; The expression of the performance index function of the neural network is as follows: In the formula, y o (k) represents the value of the output of the neural network at the kth moment, y m (k) represents the output value of the mth layer; The sensitivity of the object's output to changes in the control input is expressed by the Jacobian matrix as follows: c ij represents the weight connecting neurons i and j, Δu(k) represents the change of the control variable at time k, and b j represents the bias term of the jth neuron.

9. The vehicle speed control method of a hybrid electric vehicle according to claim 8, characterized in that: The step of designing the motor torque based on the control input of the improved disturbance observer and dynamically adjusting the motor torque and the engine torque based on the PID controller of the neural network so that the vehicle speed dynamically approaches the target vehicle speed also includes: The motor torque and the engine torque satisfy the total driving torque, and the expression of the total driving torque is as follows: T total =T m +T e (20); Where, T total Indicates the total driving torque.

10. The vehicle speed control method of a hybrid electric vehicle according to claim 9, characterized in that: The step of designing the motor torque based on the control input of the improved disturbance observer and dynamically adjusting the motor torque and the engine torque based on the PID controller of the neural network so that the vehicle speed dynamically approaches the target vehicle speed also includes: The expression for designing PID controller is as follows: In the formula, e(t) represents the deviation between the actual output and the expected output, K p is the proportional control coefficient, K i Indicates the integral control coefficient, K d represents the differential control coefficient, u(t) represents the control output of the PID controller, and y o (t) represents the reference torque of the motor, and y(t) represents the current output torque of the motor.

11. The vehicle speed control method of a hybrid electric vehicle according to claim 10, characterized in that: The step of designing the motor torque based on the control input of the improved disturbance observer and dynamically adjusting the motor torque and the engine torque based on the PID controller of the neural network so that the vehicle speed dynamically approaches the target vehicle speed also includes: The PID controller is optimized by a neural network to obtain a PID controller based on a neural network. The expression of the PID controller based on a neural network is as follows: In the formula, Δu(t) represents the change of the control variable at time t, e error (k) represents the value of the error at time k, r(k) represents the value of the reference input at time k, and y o (k) represents the value of the output of the neural network at time k; The input variables of the neural network based PID controller are: In the formula, a(1) represents the adjustment coefficient of the neural network in the input layer, a(2) represents the adjustment coefficient of the neural network in the hidden layer, a(3) represents the adjustment coefficient of the neural network in the output layer, and e error (k-1) represents the value of the error at time k-1, e error (k-2) represents the value of the error at time k-2; The control law is: In the formula, u(k) represents the control input at the current time k, u(k-1) represents the control input at the previous time k-1, K represents the set of control coefficient gains, and X represents the state variable of the control system; The expression of the performance index evaluation function of the engine system is as follows: According to the evaluation function, the neural network is used to optimize the PID controller; Using the control output of the neural network-based PID controller as the control input of the engine system to achieve control of the engine torque; Based on the control of the motor torque and the generator torque, the vehicle speed is controlled.

Images