Linear active-disturbance-rejection control method of automobile steer-by-wire system and related device

Abstract

The invention discloses an automobile steer-by-wire system linear active-disturbance-rejection control method and a related device. The method comprises the following steps: establishing an automobile steer-by-wire system model; obtaining a parrot group intelligent search algorithm; the parrot group intelligent search algorithm is improved according to the vehicle driving conditions of the vehicle under different working conditions, and an improved parrot group intelligent search algorithm is obtained; and calculating the optimal fitness of the improved parrot group intelligent search algorithm, and carrying out optimal selection on parameters of the linear active disturbance rejection controller. According to the invention, the control effect of the controller can be adaptively adjusted under different working conditions, so that the automobile has a better driving effect.

Description

Technical Field

[0001] The present invention relates to the field of steer-by-wire systems for vehicles with linear active disturbance rejection control, and particularly to a method and related device for linear active disturbance rejection control of a steer-by-wire system for vehicles optimized based on an improved parrot algorithm. Background Art

[0002] With the rapid development of the automotive industry, the steer-by-wire (SbW) system, as an advanced steering technology, has gradually become a research hotspot in the fields of intelligent vehicles and autonomous driving. The steer-by-wire system replaces the traditional mechanical connection with an electronic signal, achieving the decoupling of the steering wheel and the steering actuator, and has the advantages of fast response speed, adjustable steering characteristics, and flexible spatial layout. However, the control performance of the steer-by-wire system is affected by factors such as vehicle dynamics nonlinearity, external disturbances, and system parameter uncertainties. Traditional control methods (such as PID control) are difficult to meet the control requirements of high precision and high robustness.

[0003] The linear active disturbance rejection controller (LADRC) is an advanced control algorithm based on the extended state observer (ESO), which can estimate and compensate the internal and external disturbances of the system in real time, and has strong anti-disturbance ability and adaptability. LADRC simplifies the controller design process and reduces the dependence on the system model by unifying the nonlinearity, uncertainty, and external disturbances in the system model as the "total disturbance" and using the ESO for real-time estimation and compensation. However, the performance of LADRC highly depends on the selection of controller parameters, including the ESO bandwidth, controller gain, etc. Traditional parameter adjustment methods usually rely on empirical trial and error or manual debugging, and it is difficult to achieve the optimal configuration of parameters, especially in a complex vehicle dynamics environment.

[0004] In recent years, intelligent optimization algorithms have been widely used in the optimization of controller parameters. Traditional optimization algorithms (such as genetic algorithms, particle swarm optimization) have problems such as slow convergence speed and easy to fall into local optimum when solving high-dimensional, nonlinear optimization problems, and the existing parrot algorithm also has the problem of easy to fall into local optimum, which is difficult to meet the requirements of real-time control of the steer-by-wire system. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and related device for linear active disturbance rejection control of a steer-by-wire system for vehicles to overcome the problems existing in the prior art. The present invention can effectively cope with the nonlinearity of vehicle dynamics and external disturbances, improve the robustness of the system, and can also adaptively adjust the parameters of the controller to adapt to different road conditions in the face of different road conditions.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A linear active disturbance rejection control method for an automotive steer-by-wire system, comprising the following steps:

[0008] Step 1: Establish a model of the automotive steer-by-wire system;

[0009] Step 2: Obtain the parrot flock intelligent search algorithm;

[0010] Step 3: According to the control effect of the steer-by-wire system of the vehicle under different road conditions, improve the parrot flock intelligent search algorithm obtained in Step 2 to obtain an improved parrot flock intelligent search algorithm, and calculate the fitness value of the improved parrot flock intelligent search algorithm;

[0011] Step 4: According to the improved parrot flock intelligent search algorithm in Step 3, select the optimal fitness after iterative processing, and assign the best position coordinates corresponding to the optimal fitness to the linear active disturbance rejection controller parameters, where the linear active disturbance rejection controller parameters include the observer bandwidth w0, the controller bandwidth w c and the system gain estimation value b0;

[0012] Step 5: Substitute the linear active disturbance rejection controller parameters obtained in Step 4 into the linear active disturbance rejection controller, and apply them to the steer-by-wire system model established in Step 1 to control the vehicle to complete various different working conditions.

[0013] Further, the automotive steer-by-wire system model in Step 1 is specifically as follows:

[0014]

[0015]

[0016] T m = C m i m (5)

[0017]

[0018] Among them, formulas (1) and (2) are the dynamic models of the steering wheel assembly. In the formulas, T s , J s , θ s , and T fs are respectively the torque, moment of inertia, rotation angle, angular velocity, angular acceleration and frictional torque of the steering wheel; B s is the steering column damping coefficient; T tbs , K tbs are the torque value and torsional stiffness of the torsion bar sensor; θrim is the road feel motor rotation angle; G is the reduction ratio of the reducer; Equation (3) is the dynamic model of the steering actuator motor and its reducer, where T m is the steering motor torque; J m is the moment of inertia of the steering motor shaft; B m is the damping coefficient of the steering motor shaft; K m is the torsional stiffness of the steering motor shaft; are respectively the rotation angle, angular velocity and angular acceleration of the steering motor shaft; G m is the reduction ratio of the steering motor reducer; R p is the radius of the pinion of the pitch circle of the steering gear; X r is the rack displacement; T fm is the equivalent frictional torque of the steering motor; Equations (4) and (5) are the electromotive force balance equations in the armature winding of the steering actuator motor. In the equations, u m is the armature voltage of the steering actuator motor; i m 、 are respectively the armature current and current change rate of the steering actuator motor; R m is the armature winding resistance of the steering actuator motor; L m is the armature winding inductance of the steering actuator motor; K b is the back electromotive force proportional coefficient of the steering actuator motor; C m is the electromagnetic torque coefficient of the steering actuator motor; Equation (6) is the dynamic model of the rack and pinion steering gear. In the equation, M r 、B r are the rack mass and damping coefficient; K kt is the kingpin torsional stiffness coefficient; G w1 and G w2 are the transmission ratios from the rack to the left and right front wheels; and are respectively the rack translation speed and acceleration; θ w1 and θ w2 are the left and right steering wheel rotation angles; F s is the steering gear friction; Equations (7) and (8) are the dynamic models of the steering wheels. In the equations, J w 、B w are the moment of inertia and damping coefficient of the front wheel rotating around the kingpin; T w1 and T w2 are the left and right front wheel self-aligning torques; are respectively the angular velocity and angular acceleration of the left steering wheel; are respectively the angular velocity and angular acceleration of the right steering wheel.

[0019] Furthermore, the parrot flock intelligent search algorithm in step 2 includes the following steps:

[0020] Step 2.1: Randomly initialize the parrot population;

[0021] Step 2.2: Randomly select a behavior for position update. The behaviors include four types, namely foraging behavior, staying behavior, communication behavior, and fear behavior towards strangers.

[0022] Step 2.3: Calculate the fitness values of all parrot individuals after position update and select the optimal fitness value.

[0023] Step 2.4: Determine whether the maximum number of iterations is reached. If it is reached, output the position coordinates and fitness of the optimal parrot individual; if not, return to Step 2.2.

[0024] Furthermore, in Step 2.2, the parrot estimates the position of the food by observing the position of the food or considering the position of the owner, and then flies towards its respective position. The position update of the foraging behavior is as follows:

[0025]

[0026] where \(X^{(t)}\) is the current position of the parrot, \(X^{(t + 1)}\) is the updated position, \(\overline{X}^{(t)}\) is the average position of the current population, Levy(dim) is the Levy flight used to describe the flight of the parrot, dim is the dimension of the problem space, \(X_{best}\) is the best position from initialization to the current search and the current position of the owner, t is the current iteration number, rand(0, 1) is a random number within [0, 1], and Max is the maximum number of iterations. i (t) is the current position of the parrot, X i (t + 1) is the updated position, X M (t) is the average position of the current population, Levy(dim) is the Levy flight used to describe the flight of the parrot, dim is the dimension of the problem space, X best is the best position from initialization to the current search and the current position of the owner, t is the current iteration number, rand(0, 1) is a random number within [0, 1], Max iter is the maximum number of iterations;

[0027] The staying behavior means that the parrot suddenly flies towards the owner and stays at a certain part of the body for a period of time. Its position update is as follows:

[0028] \(X^{(t + 1)} = X^{(t)} + \alpha\cdot Levy(dim) + rand(0, 1)\cdot ones(1, dim)\) i (t + 1) = X i (t) + X best ·Levy(dim) + rand(0, 1)·ones(1, dim)

[0029] where ones(1, dim) is a vector of all 1s with dimension dim, \(\alpha\cdot Levy(dim)\) represents the process of flying towards the owner, and rand(0, 1)·ones(1, dim) represents randomly staying at a certain part of the owner's body; best ·Levy(dim) represents the process of flying towards the owner, and rand(0, 1)·ones(1, dim) represents randomly staying at a certain part of the owner's body;

[0030] The communication behavior includes communication of flying towards the flock and not flying towards the flock. Its mathematical model is as follows:

[0031]

[0032] where 0.2·rand(0,1)·(1 - t / Max iter )·(X i (t)-X M (t)) represents the process of a parrot individual joining the group for communication, 0.2·rand(0,1)·exp(-t / rand(0,1)·Max iter ) represents the process of a parrot flying away immediately after communication, P is a random number within the range of [0,1], determining which of the behaviors occurs;

[0033] The position update of the fear behavior towards strangers is as follows:

[0034]

[0035] where rand(0,1)·cos(0.5πt / Max iter )·(X best -X i (t)) represents the process of reorienting and flying towards the owner, cos(rand(0,1)·π)·(t / Max iter ) 2 / Maxiter ·(X i (t)-X best ) represents the process of moving away from strangers.

[0036] Furthermore, in step 3, the parrot swarm intelligent search algorithm is improved to obtain an improved parrot swarm intelligent search algorithm, which specifically includes the following steps:

[0037] Step 3.1: Initialize the parrot population using the monogamous cohabitation strategy, and regard the three-dimensional coordinates of the parrot as the three parameters w0, w c and b0 of the linear active disturbance rejection controller;

[0038] Step 3.2: Randomly select a behavior for each parrot individual and start iterative update;

[0039] Step 3.3: Select corresponding strategies for different behaviors for optimization, which are updating the position through the stay behavior by introducing a stay factor; updating the position through the communication behavior by fusing the salp swarm algorithm; updating the position of the fear behavior towards strangers by the golden sine algorithm;

[0040] Step 3.4: After each parrot individual updates its position, calculate the fitness of the parrot population, and select the optimal parrot individual according to the size of the fitness;

[0041] Step 3.5: Judge whether the maximum number of iterations is reached. If so, output the position coordinates and fitness value of the optimal parrot individual. If not, return to step 3.2.

[0042] Furthermore, in step 3.1, the parrot population is initialized using the perfect mate dual habitat strategy, and the formula is as follows:

[0043]

[0044] a1(i,j)=lb+ub-a(i,j)

[0045] Where α(i, j) is the initial position of the population, i is the individual index, i.e., the i-th candidate solution in the population, j is the dimension index, i.e., the j-th dimension, α1(i, j) is the updated initial position of the pairs of parrots after mutual learning, lb is the lower bound, and ub is the upper bound.

[0046] Furthermore, the specific formula of the retention factor in step 3.3 is as follows:

[0047]

[0048] Where s is the retention factor; μ min and μ max are the minimum and maximum values of the stay factor respectively; the mathematical model corresponding to the updated stay behavior is:

[0049] X i (t+1)=s·X i (t)+X best ·Levy(dim)+rand(0,1)·ones(1,dim)

[0050] Where, X i (t) is the current position of the parrot, X i (t+1) is the updated position, Levy(dim) is the Levy flight, rand(0,1) is a random number in [0,1], X best is the best position from initialization to the current search;

[0051] The leader role in the salp algorithm is introduced into the communication behavior of the parrot optimization algorithm to update the parrot position. The specific formula is:

[0052]

[0053] Where rand(0,1) is a random number in the range [0,1], and k1 is the convergence factor, which is expressed as:

[0054]

[0055] The position is updated by the fear of strangers behavior using the golden sine algorithm. The specific formula is:

[0056]

[0057] X i (t + 1) = X i (t)·|sin(r1)| - r2·sin(r1)·|b1·X best (t) - b2·X i (t)|

[0058] Wherein, b1 and b2 are golden section coefficients, r1 is a random number in the range of [0, 2π], and r2 is a random number in the range of [0, π].

[0059] The linear active disturbance rejection control system of the steer-by-wire system of an automobile includes:

[0060] Model establishment module: used to establish the model of the steer-by-wire system of an automobile;

[0061] Algorithm acquisition module: used to acquire the intelligent search algorithm of the parrot flock;

[0062] Algorithm improvement module: used to improve the obtained intelligent search algorithm of the parrot flock according to the control effect of the steer-by-wire system of the automobile under different road conditions, obtain the improved intelligent search algorithm of the parrot flock, and calculate the fitness value of the improved intelligent search algorithm of the parrot flock;

[0063] Iterative processing module: used to select the optimal fitness after iterative processing according to the improved intelligent search algorithm of the parrot flock, and assign the best position coordinates corresponding to the optimal fitness to the parameters of the linear active disturbance rejection controller, and the parameters of the linear active disturbance rejection controller include the observer bandwidth w0, the controller bandwidth w c And the system gain estimation value b0;

[0064] Substitution module: used to substitute the obtained parameters of the linear active disturbance rejection controller into the linear active disturbance rejection controller, and apply them to the steer-by-wire system model established in step 1 to control the automobile to complete various different working conditions.

[0065] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the linear active disturbance rejection control method of the steer-by-wire system of the automobile are implemented.

[0066] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the linear active disturbance rejection control method of the steer-by-wire system of the automobile are implemented.

[0067] Compared with the prior art, the present invention has the following beneficial technical effects:

[0068] Under the premise of ensuring rapid optimization, the present invention adjusts the controller parameters in real time by improving the parrot algorithm (i.e., the parrot swarm intelligent search algorithm), so that the parameters of the linear active disturbance rejection controller remain dynamically optimal, solving the uncertainty problem of manually setting parameters based on experience and the problem of non-adjustability in the face of multiple working conditions. The present invention enables the automobile to have a better control effect when facing different working conditions.

[0069] The present invention uses a good-pair dual-habitat initialization strategy to expand the location of the initialized population, making the population evenly distributed while improving the population quality; designs a stay factor to improve the length of time the parrot stays, which helps the parrot algorithm to escape from the local optimal solution; combines the rapid convergence of the golden sine algorithm with the parrot's fear of strangers, balances the convergence speed and algorithm diversity, avoids premature convergence, and can also find a higher quality optimal solution; introduces the leader role of the salp into the communication behavior stage of the parrot algorithm, improving the overall performance of the parrot algorithm.

[0070] Compared with the traditional Parrot algorithm, the present invention has a faster convergence speed, smoother changes in yaw angular velocity and center of mass sideslip angle, and smaller maximum values compared with other control methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The drawings in the specification are used to provide further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0072] Figure 1 It is a schematic flow diagram of the present invention;

[0073] Figure 2 is a flow chart of the improved parrot algorithm of the present invention;

[0074] Figure 3 : These are the results of automobile simulation under double lane-changing conditions of the present invention, wherein (a) is a comparison curve of lateral displacement between a conventional steering system and a constructed steer-by-wire system under high road adhesion, (b) is a comparison curve of yaw velocity between a conventional steering system and a constructed steer-by-wire system under high road adhesion, (c) is a comparison curve of center-of-mass slip angle between a conventional steering system and a constructed steer-by-wire system under high road adhesion, (d) is a comparison curve of lateral displacement of different control methods under uphill conditions on a slippery road, (e) is a comparison curve of yaw velocity of different control methods under uphill conditions on a slippery road, and (f) is a comparison curve of center-of-mass slip angle of different control methods under uphill conditions on a slippery road;

[0075] Figure 4It is the simulation result diagram of the vehicle under the sine input condition of the steering wheel. Among them, (a) is the comparison curve of the front wheel angle under different control methods on a wet and slippery road surface, (b) is the comparison curve of the yaw rate under different control methods on a wet and slippery road surface, (c) is the comparison curve of the sideslip angle of the center of mass under different control methods on a wet and slippery road surface, (d) is the comparison curve of the front wheel angle under different control methods on an ice and snow road surface, (e) is the comparison curve of the yaw rate under different control methods on an ice and snow road surface, and (f) is the comparison curve of the sideslip angle of the center of mass under different control methods on an ice and snow road surface. Detailed implementation manner

[0076] The present invention will be further described in detail below with reference to the accompanying drawings:

[0077] See Figure 1 is the flow chart of the present invention. The present invention optimizes and improves the basic parrot algorithm, and proposes a parrot algorithm integrating good lattice points, opposition-based learning strategy, golden sine strategy and salp swarm algorithm (Parrot Swarm Intelligence Search Algorithm Integrating Good Lattice Points Opposition-Based Learning Strategy Golden Sine Strategy and Salp Swarm Algorithm, GOGS-PO), as Figure 2 , which is applied to optimize the linear active disturbance rejection controller. First, a vehicle-by-wire steering system model is established; secondly, a good-couple double-dwelling initialization strategy is designed to expand the position of the initialized population; then, a stay factor is designed to help the algorithm jump out of the local optimal solution; finally, the golden sine algorithm is used to optimize the fear-of-strangers behavior, and the communication behavior of the salp swarm algorithm is integrated to optimize the parrot algorithm, reducing the time to find the optimal solution and improving the search efficiency. It can be seen from Figure 3 and Figure 4 that the improved parrot algorithm has better control effect on the vehicle compared with the original algorithm.

[0078] The specific steps are as follows:

[0079] Step 1: Model the vehicle-by-wire steering system model as follows:

[0080]

[0081] T tbs =K tbs (θ s -θ rim / G)

[0082]

[0083]

[0084] T m = C m i m

[0085]

[0086]

[0087]

[0088] In the formula, the first formula and the second formula are the dynamic models of the steering wheel assembly. In the formula, T s is the torque of the steering wheel, J s is the moment of inertia of the steering wheel, θ s is the steering angle of the steering wheel, and T fs is the frictional torque of the steering wheel; B s is the damping coefficient of the steering column; T tbs is the torque value of the torsion bar sensor, K tbs is the torsional stiffness of the torsion bar sensor; θ rim is the road feel motor steering angle; G is the reduction ratio of the reducer. The third formula is the dynamic model of the steering actuator motor and its reducer. In the formula, T m is the torque of the steering motor; J m is the moment of inertia of the steering motor shaft; B m is the damping coefficient of the steering motor shaft; K m is the torsional stiffness of the steering motor shaft; θ m is the steering angle of the steering motor shaft; G m is the reduction ratio of the steering motor reducer; R p is the radius of the pinion of the pitch circle of the steering gear; X r is the rack displacement; T fm is the equivalent frictional torque of the steering motor; The fourth formula and the fifth formula are the electromotive force balance equations in the armature winding of the steering actuator motor. In the formula, u m is the armature voltage of the steering actuator motor; i m is the armature current of the steering actuator motor; R m is the armature winding resistance of the steering actuator motor; L m is the armature winding inductance of the steering actuator motor; K b is the back electromotive force proportionality coefficient of the steering actuator motor; C m is the electromagnetic torque coefficient of the steering actuator motor; The sixth formula is the dynamic model of the rack and pinion steering gear. In the formula, M r , B r is the mass and damping coefficient of the rack; K kt is the kingpin torsional stiffness coefficient; Gw1 and G w2 are the transmission ratios from the rack to the left and right front wheels; θ w1 and θ w2 are the steering angles of the left and right steering wheels; F s is the frictional force of the steering gear; The seventh and eighth formulas are the dynamic models of the steering wheels. In the formulas, J w , B w are the moment of inertia and damping coefficient of the front wheel rotating around the kingpin; T w1 and T w2 are the self-aligning torques of the left and right front wheels.

[0089] Step 2: The basic parrot search algorithm is divided into the following steps:

[0090] Foraging behavior: Parrots mainly estimate the approximate position of food by observing the position of the food or considering the position of the owner, and then fly towards their respective positions. The position update of the foraging behavior is as follows:

[0091]

[0092] In the formula, X i (t) is the current position of the parrot, X i (t + 1) is the updated position, X M (t) is the average position of the current population, Levy(D) is the Levy flight, which is used to describe the flight of the parrot, X best is the best position from initialization to the current search and the current position of the owner, t is the current iteration number, rand(0, 1) is a random number within [0, 1], Max iter is the maximum number of iterations.

[0093] Staying behavior: The parrot suddenly flies towards the owner and stays at a certain part of the body for a period of time. The position update of the staying behavior is as follows:

[0094] X i (t + 1) = X i (t) + X best ·Levy(dim) + rand(0, 1)·ones(1, dim)

[0095] In the formula, ones(1, dim) is a vector of all 1s with dimension dim, X best ·Levy(D) represents the process of flying towards the owner, and rand(0, 1)·ones(1, dim) represents randomly staying at a certain part of the owner's body.

[0096] Communication behavior: The communication behavior of the parrot includes communication of flying towards the flock and not flying towards the flock. The position update of the communication behavior is as follows:

[0097]

[0098] In the formula, 0.2·rand(0,1)·(1 - t / Max iter )·(X i (t) - X M (t)) represents the process of a parrot individual joining the group for communication, and 0.2·rand(0,1)·exp(-t / rand(0,1)·Max iter ) represents the process of flying away immediately after the parrot's communication. P is a random number within the range of [0,1], which determines which of these behaviors occurs.

[0099] Regarding the fear behavior towards strangers, the parrot will have a fear psychology towards strangers and jointly look for a safe place with the owner. The position update of the fear behavior towards strangers is as follows:

[0100]

[0101] In the formula, rand(0,1)·cos(0.5πt / Max iter )·(X best - X i (t)) represents the process of reorienting and flying towards the owner, and cos(rand(0,1)·π)·(t / Max iter ) 2 / Maxiter ·(X i (t) - X best ) represents the process of moving away from strangers.

[0102] Step 3: Improve the parrot search algorithm to obtain an improved parrot search algorithm.

[0103] First, initialize the parrot population using the monogamous strategy. The formula is as follows:

[0104]

[0105] a1(i,j) = lb + ub - a(i,j)

[0106] In the formula, α(i, j) is the initial position of the population, α1(i, j) is the updated initial position after the paired parrots learn from each other, lb is the lower bound, and ub is the upper bound.

[0107] In the initialization stage, the Parrot Optimization Algorithm (POA) usually randomly generates a set of candidate solutions as the initial population. However, this random initialization method may lead to a lack of diversity in the population, especially when the solution space is large or the problem complexity is high. A lack of diversity in the population means that the initial population may be concentrated in a local area of the solution space, thus limiting the global search ability of the algorithm. This may cause the algorithm to be difficult to jump out of the local optimal solution in the subsequent iterative process, affecting the accuracy and reliability of the final optimization results. Therefore, a paired initialization strategy is designed. This strategy combines the habit of parrots often acting in pairs. Two individuals learn from each other and exchange information, expanding the positions of the initialized population, making the population evenly distributed, and improving the population quality at the same time.

[0108] Secondly, the specific formula for designing the stay factor is as follows:

[0109]

[0110] In the formula, μ min and μ max are the minimum and maximum values of the stay factor respectively. The mathematical model corresponding to the updated stay behavior is:

[0111] X i (t + 1) = s·X i (t) + X best ·Levy(dim) + rand(0, 1)·ones(1, dim)

[0112] Similar to most swarm intelligence optimization algorithms, the PO algorithm also faces the problem of being easily trapped in local optimal solutions, converging prematurely to a solution that is not the global optimum during the search process, resulting in stagnation of the algorithm performance and the quality of the solution not reaching the expected global optimum level. Therefore, a stay factor is designed to improve the stay duration of parrots in the stay behavior, thus helping the algorithm jump out of the local optimal solution.

[0113] Then, the leader role in the Salp Swarm Algorithm is introduced into the communication behavior of the Parrot Optimization Algorithm to update the parrot positions. The specific formula is:

[0114]

[0115] In the formula, rand is a random number in the range of [0, 1], and k1 is the convergence factor, and its expression is:

[0116]

[0117] Since individual parrots tend to imitate and follow other individuals in the group during the communication process, this may lead to a reduction in population diversity. When most individuals in the population gather in a local optimal area, other individuals may find it difficult to jump out of this area due to imitation behavior, thus limiting the global search ability of the algorithm. Secondly, the information dissemination mechanism in the communication behavior may not be flexible and efficient enough to make full use of the effective information in the population to guide the search direction of individuals. This may result in a slow convergence rate during the iteration process of the algorithm and it is difficult to reach the global optimal solution. Therefore, introducing the leader role in the salp swarm optimization algorithm into the communication behavior stage of the parrot optimization algorithm can optimize the information dissemination mechanism. The leader position update and search strategy can become the object of imitation and learning for other individuals, which helps to accelerate the information dissemination speed in the population and improve the overall performance of the algorithm.

[0118] Finally, the position is updated through the fear of strangers behavior of the golden sine algorithm, and the specific formula is:

[0119]

[0120]

[0121]

[0122] In the formula, b1 and b2 are the golden section coefficients, r1 is a random number in the range of [0, 2π], and r2 is a random number in the range of [0, π].

[0123] Combining the golden sine algorithm with the fear of strangers behavior of parrots, making use of the rapid convergence of the golden sine algorithm and the fear of strangers behavior in the parrot optimization algorithm can increase the diversity of the algorithm and avoid premature convergence. Therefore, combining these two algorithms can balance the convergence speed and diversity, enabling the algorithm to find higher-quality optimal solutions while maintaining efficient convergence.

[0124] Figure 3 It is the vehicle simulation result diagram under the double lane change condition. At a vehicle speed of 50 km / h, the lateral displacement, yaw rate, and sideslip angle of the center of mass obtained by different control methods are compared and analyzed. Among them Figure 3 (a) shows the comparison of lateral displacements under high road adhesion, Figure 3 (b) shows the comparison of yaw rates under high road adhesion, Figure 3 (c) shows the comparison of sideslip angles of the center of mass under high road adhesion, Figure 3 (d) shows the comparison of lateral displacements under the wet and slippery uphill condition, Figure 3 (e) shows the comparison of yaw rates under the wet and slippery uphill condition, Figure 3 (f) shows the comparison of sideslip angles of the center of mass under the wet and slippery uphill condition.Figure 3 It can be seen that after the improved parrot search algorithm is applied to the steer-by-wire system of the vehicle, under two different working conditions, the lateral displacement of the vehicle is closer to the ideal value, and the changes in yaw rate and sideslip angle of the center of mass are smoother, and the maximum values are also smaller compared to other control methods.

[0125] Figure 4 Figure 5 shows the vehicle simulation results under the condition of sinusoidal input of the steering wheel. Under the condition of a vehicle speed of 50 km / h and a low road surface adhesion coefficient, the front wheel angles, yaw rates, and sideslip angles of the center of mass obtained by different control methods are compared and analyzed. Among them Figure 4 (a) is the comparison curve of the front wheel angles of different control methods on a wet and slippery road surface, Figure 4 (b) is the comparison curve of the yaw rates of different control methods on a wet and slippery road surface, Figure 4 (c) is the comparison curve of the sideslip angles of the center of mass of different control methods on a wet and slippery road surface, Figure 4 (d) is the comparison curve of the front wheel angles of different control methods on an ice and snow road surface, Figure 4 (e) is the comparison curve of the yaw rates of different control methods on an ice and snow road surface, Figure 4 (f) is the comparison curve of the sideslip angles of the center of mass of different control methods on an ice and snow road surface. It can be seen that Figure 4 after the improved parrot search algorithm is applied to the steer-by-wire system of the vehicle, under the two working conditions of low road surface adhesion, the peak values of the sideslip angle of the center of mass and the yaw rate are smaller, and the driving state of the vehicle is more stable.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the scope of its protection. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: after reading the present invention, those skilled in the art can still make various changes, modifications, or equivalent replacements to the specific embodiments of the invention, but these changes, modifications, or equivalent replacements are all within the scope of the protection of the pending claims of the invention.

Claims

1. A linear active disturbance rejection control method for a steer-by-wire system of an automobile, characterized in that, It includes the following steps: Step 1: Establish a steer-by-wire system model for the vehicle; Step 2: Obtain the parrot flock intelligent search algorithm; Step 3: According to the control effect of the steer-by-wire system of the vehicle under different road conditions, improve the parrot flock intelligent search algorithm obtained in Step 2 to obtain an improved parrot flock intelligent search algorithm, and calculate the fitness value of the improved parrot flock intelligent search algorithm; Step 4: According to the improved parrot flock intelligent search algorithm in Step 3, after iterative processing, the optimal fitness is selected, and the best position coordinates corresponding to the optimal fitness are assigned to the parameters of the linear active disturbance rejection controller. The parameters of the linear active disturbance rejection controller include the observer bandwidth w0, the controller bandwidth w c and the system gain estimation value b0; Step 5: Substitute the linear active disturbance rejection controller parameters obtained in Step 4 into the linear active disturbance rejection controller, and apply it to the steer-by-wire system model established in Step 1 to control the vehicle to complete various different working conditions.

2. The linear active disturbance rejection control method for the steer-by-wire system of an automobile according to claim 1, wherein The vehicle steer-by-wire system model in Step 1 is specifically as follows: T tbs = K tbs (θ s - θ rim / G)(2) T m = C m i m (5) Among them, Formula (1) and Formula (2) are the dynamic models of the steering wheel assembly. In the formulas, T s , J s , θ s , , and T fs are respectively the torque, moment of inertia, angle of rotation, angular velocity, angular acceleration, and frictional torque of the steering wheel; B s is the damping coefficient of the steering column; T tbs , K tbs are the torque value and torsional stiffness of the torsion bar sensor; θ rim is the road feel motor angle of rotation; G is the reduction ratio of the reducer; Formula (3) is the dynamic model of the steering actuator motor and its reducer. In the formula, T m is the steering motor torque; J m is the moment of inertia of the steering motor shaft; B m is the damping coefficient of the steering motor shaft; K m is the torsional stiffness of the steering motor shaft; θ m , are respectively the angle of rotation, angular velocity, and angular acceleration of the steering motor shaft; G m is the reduction ratio of the steering motor reducer; R p is the radius of the pinion of the pitch circle of the steering gear; X r is the rack displacement; T fm is the equivalent frictional torque of the steering motor; Formulas (4) and (5) are the electromotive force balance equations in the armature winding of the steering actuator motor. In the formulas, u m is the armature voltage of the steering actuator motor; i m , are respectively the armature current and current change rate of the steering actuator motor; R m is the armature winding resistance of the steering actuator motor; L m is the armature winding inductance of the steering actuator motor; K b is the back electromotive force proportional coefficient of the steering actuator motor; C m is the electromagnetic torque coefficient of the steering actuator motor; Formula (6) is the dynamic model of the rack and pinion steering gear. In the formula, M r , B r are the rack mass and damping coefficient; K kt is the kingpin torsional stiffness coefficient; G w1 and G w2 are the transmission ratios from the rack to the left and right front wheels; and are respectively the rack translation speed and acceleration; θ w1 and θ w2 are the angles of rotation of the left and right steering wheels; F s is the frictional force of the steering gear; Formulas (7) and (8) are the dynamic models of the steering wheels. In the formulas, J w , B w is the moment of inertia and damping coefficient of the front wheel rotating around the kingpin; T w1 and T w2 are the self-aligning torques of the left and right front wheels; are the angular velocity and angular acceleration of the left steering wheel respectively; are the angular velocity and angular acceleration of the right steering wheel respectively.

3. The linear active disturbance rejection control method for the steer-by-wire system of an automobile according to claim 1, characterized in that, The parrot flock intelligent search algorithm in Step 2 includes the following steps: Step 2.1: Randomly initialize the parrot population; Step 2.2: Randomly select a behavior for position update. The behaviors include four types, namely foraging behavior, staying behavior, communication behavior, and fear behavior towards strangers; Step 2.3: Calculate the fitness value after position update of all parrot individuals, and select the optimal fitness value; Step 2.4: Determine whether the maximum number of iterations is reached. If so, output the position coordinates and fitness of the optimal parrot individual; if not, return to Step 2.

2.

4. The linear active disturbance rejection control method for the steer-by-wire system of an automobile according to claim 3, characterized in that, In Step 2.2, the parrot estimates the position of the food by observing the position of the food or considering the position of the owner, and then flies towards its respective position. The position update of the foraging behavior is as follows: Where, X i (t) is the current position of the parrot, X i (t + 1) is the updated position, X M (t) is the average position of the current population, Levy(dim) is the Levy flight, which is used to describe the flight of the parrot, dim is the dimension of the problem space, X best is the best position from initialization to the current search and the current position of the master, t is the current iteration number, rand(0, 1) is a random number within [0, 1], Max iter is the maximum number of iterations; The staying behavior means that the parrot suddenly flies towards the owner and stays at a certain part of the body for a preset time. Its position update is as follows: X i (t + 1) = X i (t) + X best ·Levy(dim) + rand(0, 1)·ones(1, dim) where ones(1, dim) is a vector of all 1s with dimension dim, and X best ·Levy(dim) represents the process of flying towards the host, and rand(0, 1)·ones(1, dim) represents randomly staying at a certain body part of the host; The communication behavior includes the communication of flying towards the flock and not flying towards the flock. Its mathematical model is as follows: where 0.2·rand(0,1)·(1 - t / Max iter )·(X i (t) - X M (t)) represents the process of a parrot individual joining the group for communication, and 0.2·rand(0,1)·exp(-t / rand(0,1)·Max iter ) represents the process of the parrot flying away immediately after communication. P is a random number within the range of [0,1], which determines which of these behaviors occurs; The position update of the fear behavior towards strangers is as follows: where rand(0,1)·cos(0.5πt / Max iter )·(X best -X i (t)) represents the process of redirecting to fly towards the owner, and cos(rand(0,1)·π)·(t / Max iter ) 2 / Maxiter ·(X i (t)-X best ) represents the process of moving away from strangers.

5. The linear active disturbance rejection control method for a steer-by-wire system of an automobile according to claim 4, wherein In Step 3, the parrot flock intelligent search algorithm is improved to obtain an improved parrot flock intelligent search algorithm, which specifically includes the following steps: Step 3.1: Initialize the parrot population using the cohabitation strategy. The three-dimensional coordinates of the parrots are used as the three parameters w0, w, and b0 of the linear active disturbance rejection controller; c and Step 3.2: Randomly select a behavior for each parrot individual and start iterative update; Step 3.3: Select corresponding strategies to optimize different behaviors, namely update the position of the staying behavior by introducing a staying factor; update the position of the communication behavior by fusing the salp swarm algorithm; update the position of the fear behavior towards strangers by the golden sine algorithm; Step 3.4: After each parrot individual updates its position, calculate the fitness of the parrot population, and select the optimal parrot individual according to the size of the fitness; Step 3.5: Determine whether the maximum number of iterations is reached. If so, output the position coordinates and fitness value of the optimal parrot individual. If not, return to Step 3.

2.

6. The linear active disturbance rejection control method for the steer-by-wire system of an automobile according to claim 5, wherein In Step 3.1, the parrot population is initialized using the monogamous strategy. The formula is as follows: a1(i,j)=lb+ub-a(i,j) In the formula, α(i, j) is the initial position of the population, i is the individual index, that is, the i-th candidate solution in the population, j is the dimension index, that is, the j-th dimension, α1(i, j) is the updated initial position after mutual learning of paired parrots, lb is the lower bound, and ub is the upper bound.

7. The linear active disturbance rejection control method for the steer-by-wire system of an automobile according to claim 6, characterized in that, The specific formula of the staying factor in Step 3.3 is as follows: where s is the residence factor; μ min and μ max are the minimum and maximum values of the residence factor, respectively; the mathematical model corresponding to the updated residence behavior is: X i (t + 1)= s·X i (t)+X best ·Levy(dim)+rand(0,1)·ones(1,dim) where X i (t) is the current position of the parrot, X i (t + 1) is the updated position, Levy(dim) is the Levy flight, rand(0, 1) is a random number within [0, 1], X best is the best position from initialization to the current search; Introduce the leader role in the salp swarm algorithm into the communication behavior of the parrot optimization algorithm to update the parrot position. The specific formula is: where rand(0,1) is a random number within the range of [0,1], and k1 is a convergence factor, and its expression is: Update the position through the fear behavior of strangers by the golden sine algorithm, and the specific formula is: X i (t + 1)=X i (t)·|sin(r1)| - r2·sin(r1)·|b1·X best (t)-b2·X i (t)| where b1 and b2 are golden section coefficients, r1 is a random number within the range of [0, 2π], and r2 is a random number within the range of [0, π].

8. Linear auto-disturbance rejection control system for vehicle-by-wire steering system, characterized in that, Including: Model establishment module: used to establish a steer-by-wire system model of the vehicle; Algorithm acquisition module: used to acquire the intelligent search algorithm of the parrot flock; Algorithm improvement module: used to improve the obtained intelligent search algorithm of the parrot flock according to the control effect of the steer-by-wire system of the vehicle under different road conditions, obtain the improved intelligent search algorithm of the parrot flock, and calculate the fitness value of the improved intelligent search algorithm of the parrot flock; Iterative processing module: It is used to select the optimal fitness after iterative processing according to the improved parrot flock intelligent search algorithm, and assign the best position coordinates corresponding to the optimal fitness to the parameters of the linear active disturbance rejection controller. The parameters of the linear active disturbance rejection controller include the observer bandwidth w0, the controller bandwidth w c and the system gain estimation value b0; Substitution module: used to substitute the obtained linear active disturbance rejection controller parameters into the linear active disturbance rejection controller, and apply them to the steer-by-wire system model established in step 1 to control the vehicle to complete various different working conditions.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the linear active disturbance rejection control method for the steer-by-wire system of the vehicle according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the linear active disturbance rejection control method for the steer-by-wire system of the vehicle according to any one of claims 1 to 7.

Images