A method for controlling the stability of a vehicle with distributed drive

Abstract

The application discloses a kind of distributed drive vehicle driving stability control method, comprising: establishing single-point preview deviation model, the front wheel angle of vehicle is calculated;Design rear wheel angle sliding mode controller, eliminate the deviation of actual vehicle mass center side slip angle and ideal mass center side slip angle;Design additional yaw moment sliding mode controller, eliminate the deviation of actual vehicle yaw rate and ideal yaw rate;According to the deviation of actual vehicle speed and ideal vehicle speed, design longitudinal vehicle speed PID controller;Determine the improved whale optimization algorithm, parameter self-tuning is carried out to sliding mode controller;According to Ackerman steering principle, vehicle front and rear axle dynamic load is respectively redistributed to vehicle angle, torque.The application is aimed at the problem of poor control caused by multiple parameter setting of sliding mode controller, by introducing improved whale optimization algorithm, for sliding mode controller parameter setting, can improve the stability performance of general sliding mode control method and its adaptability to different driving conditions.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving, specifically a method for controlling the driving stability of distributed drive vehicles. Background Technology

[0002] The independent controllability of each wheel in a distributed drive vehicle gives it low-speed maneuverability and high-speed stability, making it an ideal platform for intelligent vehicles. However, excessive steering angle / torque control inputs greatly increase the difficulty of control and make it difficult to guarantee driving stability.

[0003] Currently, stability control algorithms include linear control and nonlinear control. The main method is PID control, but it cannot handle external disturbances and cannot guarantee stability over a wide range. MPC control can effectively handle multiple system constraints, but its real-time performance is poor and its computation speed needs improvement. Sliding mode control system has strong anti-interference ability and fast response speed, and is a highly reliable and effective control method. Its unique design enables the system to stably control the target variable in complex environments without being affected by external disturbances. However, its performance is affected by multiple parameters. Therefore, parameter tuning is a major challenge for sliding mode control. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a distributed drive vehicle driving stability control method to address the shortcomings of the prior art. This control method addresses the problem of poor control caused by multi-parameter tuning of sliding mode controllers by introducing an improved whale optimization algorithm for sliding mode controller parameter tuning, which can improve the stability performance of general sliding mode control methods and their adaptability to different driving conditions.

[0005] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for controlling the driving stability of a distributed drive vehicle includes the following steps:

[0007] Step S1: Based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, establish a single-point anti-alignment deviation model and calculate the front wheel steering angle of the vehicle;

[0008] Step S2: Design a rear wheel steering angle sliding mode controller to eliminate the deviation between the actual vehicle center of gravity sideslip angle and the ideal center of gravity sideslip angle;

[0009] Step S3: Design an additional yaw moment sliding mode controller to eliminate the deviation between the actual vehicle yaw rate and the ideal yaw rate;

[0010] Step S4: Design a longitudinal speed PID controller based on the deviation between the actual vehicle speed and the ideal vehicle speed;

[0011] Step S5: An improved whale optimization algorithm is determined by initializing the chaotic map, updating the convergence factor nonlinearly, and adjusting the position of the adaptive inertia weight. The sliding mode controllers in steps S2 and S3 are then self-tuned according to the requirements of the fitness function.

[0012] Step S6: Based on the Ackermann steering principle and the dynamic load on the front and rear axles of the vehicle, the vehicle's steering angle and torque are redistributed respectively.

[0013] As a further improvement to the present invention, in step S1, based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, a single-point anti-aiming deviation model is established, specifically as follows:

[0014] Assume the vehicle maintains a constant yaw rate over a future period of time, and the vehicle's lateral speed v y Much smaller than the longitudinal vehicle speed v x Determine the vehicle's appropriate speed The value remains unchanged;

[0015] Based on this constant yaw rate assumption, the ideal yaw rate can be obtained:

[0016]

[0017] Where ΔY is the lateral deviation between the aiming point and the vehicle, and t p The second represents the aiming time, and β represents the vehicle's center of gravity sideslip angle.

[0018] Combining this with the mathematical relationships governing vehicle steering, the front wheel steering angle can be obtained as follows:

[0019]

[0020] Where h is the wheelbase between the front and rear axles of the vehicle, and K is the stability factor.

[0021] As a further improvement of the present invention, step S2 specifically includes:

[0022] S201: The deviation of the centroid side slip angle e β e1, as the sliding mode control variable, is the rear wheel rotation angle δ r As the output variable u1, i.e., e1 = e β =β-β d u1=δ r The integral sliding surface is designed as follows:

[0023] Where β is the actual sideslip angle of the vehicle's center of gravity; β d The ideal centroid sideslip angle; These are the proportional and integral coefficients, respectively, both of which are non-zero real numbers;

[0024] S202: Using the traditional exponential reaching law Based on this, a continuous curve is used to overcome the hysteresis of sgn(s). The continuous curve is designed as follows:

[0025]

[0026] Where θ is a non-zero real number;

[0027] The approach rate is designed as

[0028] Where s1 is the sliding surface; ε1 and K1 are non-zero coefficients;

[0029] S203: Combining the two-degree-of-freedom vehicle model, for equation e1=e β =β-β d Differentiating both sides, we get:

[0030]

[0031] Where m is the total vehicle mass; l f l is the distance from the front axle to the center of mass. r C is the distance from the rear axle to the center of mass. f C represents the front wheel lateral stiffness. r γ is the rear wheel lateral stiffness; γ is the actual yaw rate.

[0032] Substituting the convergence ratio into the sliding surface, we get:

[0033]

[0034] Combining equations (5) and (6), we can obtain the formula for the rear wheel steering angle:

[0035]

[0036] Where, δ r This refers to the rear wheel steering angle.

[0037] As a further improvement to the present invention, step S3 specifically includes:

[0038] S301: Using the actual yaw rate γ and the ideal yaw rate γ d deviation e γ e2 is the sliding mode control variable, and the additional yaw moment ΔM is the output variable u2, i.e., e2 = e γ =γ-γ d u2 = ΔM, the integral sliding surface is designed as follows: These are the proportional coefficient and the integral coefficient, respectively, both of which are non-zero real numbers;

[0039] S302: The approach rate is designed as follows

[0040] Where s2 is the sliding surface; ε2 and K2 are non-zero coefficients;

[0041] S303: Combining a two-degree-of-freedom vehicle model, for equation e2=e γ =γ-γ d Differentiating both sides, we get:

[0042]

[0043] Among them, I z Let be the moment of inertia of the vehicle's center of mass about an axis perpendicular to the ground.

[0044] Substituting the convergence ratio into the sliding surface, we get:

[0045]

[0046] Combining equations (9) and (10), we can obtain the formula for the additional yaw moment:

[0047]

[0048] As a further improvement of the present invention, step S4 specifically includes:

[0049] S401, Calculate speed deviation e v :

[0050] e v =v x_t -v x (12);

[0051] In the formula, v x_t For the target vehicle speed, v x This refers to the actual vehicle speed;

[0052] S402, Calculate the longitudinal moment of the vehicle:

[0053]

[0054] In the formula, k p k i k d These represent proportional, integral, and differential constants, respectively.

[0055] As a further improvement to the present invention, step S5 specifically includes:

[0056] S501: Input, Output, and Fitness Function: The inputs of an improved whale optimization algorithm include: lateral deviation index J1, velocity deviation index J2, yaw rate index J3, and center of mass sideslip angle index J4, respectively; population size; and number of iterations. The outputs include rear wheel steering angle sliding mode controller parameters. ε1, K1, and additional yaw moment sliding mode controller parameters ε2, K2; the fitness function is:

[0057]

[0058] In the formula, c1, c2, c3, and c4 are non-zero weighting coefficients, and e d (t′) represents the lateral deviation at time t′ within a period T, e v (t′) represents the velocity deviation at time t′ within one T cycle, γ(t′) is the yaw rate at time t′ within one T cycle, and β(t′) is the centroid sideslip angle at time t′ within one T cycle;

[0059] S502: Population Chaotic Initialization Design: The position of the whale population is initialized using a Chebyshev chaotic sequence, specifically:

[0060] Let the chaotic sequence generated in the D-dimensional Euclidean search space be expressed as:

[0061]

[0062] The mathematical expression for generating the Chebyshev chaotic map is:

[0063] y i+1,d =cos(χ·cos -1 y id ),y id ∈[-1,1](16);

[0064] In the formula, χ is the order of the Chebyshev mapping; when using it, a set of pseudo-random numbers y is arbitrarily generated within a given range. 1d ={y 1d After d = 1, 2, ..., D, the uniformly distributed sequence of random numbers y is obtained according to formula (15). N×D This is the Chebyshev chaotic sequence;

[0065] Mapping the chaotic sequence obtained from formula (16) into the algorithm solution space, we obtain the initialized whale population as follows:

[0066]

[0067] Among them, X che For the entire whale population; X i che For the i-th individual whale; X id che X represents the d-th dimension where the i-th individual whale is located; d max Let X represent the i-th individual whale.i che The upper limit of the value of the d-th dimension, X d min Let X represent the i-th individual whale. i che The lower bound of the value of the d-th dimension;

[0068] S503: Nonlinear convergence factor design:

[0069] The formula is updated using a nonlinear time-varying update strategy for the convergence factor a:

[0070]

[0071] In the formula, n and m are constants; t is the number of iterations; t max This represents the maximum number of iterations.

[0072] S504: Adaptive Inertia Weight Position Adjustment Design

[0073] Set individual whales The fitness is f i (t), the average fitness of the population is f min (t) represents the optimal fitness of the current population. Based on the numerical relationship of fitness, the population is divided into two subpopulations, and different inertia weights are used to adjust the position of individuals. The inertia weight design is as follows:

[0074]

[0075] Among them, w i (t) represents the inertia weight of the i-th individual whale in the t-th iteration; w max w represents the upper limit of the inertia weight. min This indicates the lower limit of the inertia weight value;

[0076] S505: Main Loop: An improved whale optimization algorithm performs N iterations, each iteration including the following steps:

[0077] a) Population chaos initialization;

[0078] b. Call the simulation module;

[0079] c. Record and calculate the fitness value and position of the best individual in the current population;

[0080] Where, optimal fitness represents the minimum fitness value, denoted by f. min (t) represents;

[0081] d. Update the nonlinear convergence factor;

[0082] e. Update parameters A, C, and l;

[0083] Where A and C are the coefficient vectors in the whale optimization algorithm; l is a random number between [-1, 1];

[0084] f. Calculate the average fitness value of the population and the corresponding inertia weight;

[0085] The average fitness of the population is

[0086] g. Update the individual's position based on the probability value p and the position parameter A;

[0087] Introducing the above improved method (i.e., inertia weight) yields the individual position update formula:

[0088]

[0089] Among them, X best (t) represents the optimal position vector at the current iteration number; D represents the distance vector between the optimal position and the individual position; A represents the coefficient variable; X rand (t) represents the whale position randomly selected in the t-th iteration; b is the logarithmic spiral shape constant; l is a random number between [-1, 1];

[0090] S506: After N iterations, output the optimal rear wheel steering angle sliding mode controller parameters. K1, ε1, and additional yaw moment sliding mode controller parameters K2, ε2.

[0091] As a further improvement of the present invention, step S6 specifically includes:

[0092] S601: Angle distributor receives the front wheel steering angle δ from the single-point aiming controller. f And the rear wheel steering angle δ of the sliding mode controller r And redistribute according to the following Ackermann angle relationship:

[0093]

[0094] In the formula, δ fl The left front wheel steering angle; δ fr The steering angle of the right front wheel; δ rl The left rear wheel steering angle; δ rr The steering angle of the right rear wheel; B f B is the distance between the front wheels; r This refers to the distance between the rear wheels;

[0095] S602: Angle distributor receives additional yaw moment ΔM and longitudinal force F. tThe loads are then redistributed according to the following dynamic load estimation relationship between the front and rear axles:

[0096]

[0097] In the formula, T fl T represents the torque of the left front wheel. fr T represents the torque of the right front wheel. rl T represents the torque of the left rear wheel. rr R is the torque of the right rear wheel; R is the wheel radius; F z_fl For the vertical load on the left front wheel; F z_fr The vertical load on the right front wheel; F z_rl For the vertical load on the left rear wheel; F z_rr The load is the vertical load on the right rear wheel.

[0098] As a further improvement to the present invention, the formula for the two-degree-of-freedom vehicle model in step S203 is:

[0099]

[0100] In the formula, m is the total vehicle mass; l f The distance from the front axle to the center of gravity; l r is the distance from the rear axle to the center of mass; v is the velocity at the center of mass; I z C is the moment of inertia of the vehicle's center of mass about an axis perpendicular to the ground; f C represents the front wheel lateral stiffness. r This refers to the rear wheel lateral stiffness.

[0101] The beneficial effects of this invention are as follows:

[0102] This invention addresses the problem of poor control caused by multi-parameter tuning of sliding mode controllers. By introducing an improved whale optimization algorithm, a fitness function that can maintain vehicle driving stability is designed and used for sliding mode controller parameter tuning, effectively improving vehicle stability under different driving conditions. Attached Figure Description

[0103] Figure 1 This is a schematic diagram of a distributed drive vehicle driving stability control method according to the present invention.

[0104] Figure 2 Iterative graph for the improved whale optimization algorithm.

[0105] Figure 3 This is a schematic diagram of Ackermann steering.

[0106] Figure 4 The figure shows a comparison of the results of low-adhesion pavement conditions under different control methods.

[0107] Figure 5The figure shows a comparison of the results of high-adhesion pavement conditions under different control methods. Detailed Implementation

[0108] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0109] The independent controllability of each wheel in a 4WIS-4WID vehicle enables it to achieve both low-speed maneuverability and high-speed stability, making it an ideal platform for intelligent vehicles. However, excessive steering angle / torque control inputs significantly increase the difficulty of control, making it hard to balance driving stability and path tracking accuracy. To address this, this embodiment proposes a hierarchical control strategy based on an improved whale algorithm, including a path tracking layer, a stability control layer, and a torque-steering angle allocation layer. Based on a single-point anticipation deviation model, a front-wheel steering controller was designed to achieve path tracking control. Combining a two-degree-of-freedom vehicle model and sliding mode control theory, a sliding mode controller integrating rear-wheel steering and direct yaw moment control was designed to eliminate the influence of the center-of-gravity sideslip angle and yaw rate on vehicle stability. A novel approach rate was introduced to eliminate chattering in sliding mode control. Dynamic load transfer and Ackermann steering theorem were used to distribute wheel torque and steering angle. To address the control issues caused by multi-parameter tuning of the sliding mode controller, an improved whale optimization algorithm (IWOA) was derived through three methods: chaotic mapping initialization, nonlinear update of convergence factors, and adaptive inertia weight position adjustment. A fitness function that balances vehicle driving stability and path tracking accuracy was designed and used for sliding mode controller parameter tuning. Simulation results show that compared to the front-wheel steering vehicle and the 4WIS-4WID vehicle, the 4WIS-4WID intelligent vehicle optimized based on the improved whale algorithm exhibits the best overall performance, significantly improving driving stability while maintaining path tracking accuracy.

[0110] like Figure 1 As shown, a distributed drive vehicle driving stability control method integrating an improved whale optimization algorithm includes the following steps:

[0111] Step S1: Based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, establish a single-point anti-alignment deviation model and calculate the front wheel steering angle of the vehicle.

[0112] Among them, based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, a single-point anti-aiming deviation model is established, specifically as follows:

[0113] Assume the vehicle maintains a constant yaw rate over a future period of time, and the vehicle's lateral speed v y Much smaller than the longitudinal vehicle speed v x Determine the vehicle's appropriate speed The value remains unchanged.

[0114] Based on this constant yaw rate assumption, the ideal yaw rate can be obtained:

[0115]

[0116] Where ΔY is the lateral deviation between the aiming point and the vehicle, and t p The second represents the aiming time, and β represents the vehicle's center of gravity sideslip angle.

[0117] Combining this with the mathematical relationships governing vehicle steering, the front wheel steering angle can be obtained as follows:

[0118]

[0119] Where h is the wheelbase between the front and rear axles of the vehicle, and K is the stability factor.

[0120] Step S2: Design a rear wheel steering angle sliding mode controller to eliminate the deviation between the actual vehicle center of gravity sideslip angle and the ideal center of gravity sideslip angle.

[0121] Step S2 is as follows:

[0122] S201: The deviation of the centroid side slip angle e β e1, as the sliding mode control variable, is the rear wheel rotation angle δ r As the output variable u1, i.e., e1 = e β =β-β d u1=δ r The integral sliding surface is designed as follows: Where β is the actual sideslip angle of the vehicle's center of gravity; β d The ideal centroid sideslip angle; These are the proportional and integral coefficients, respectively, both of which are non-zero real numbers.

[0123] S202: Using the traditional exponential reaching law Based on this, a continuous curve is used to overcome the hysteresis of sgn(s). The continuous curve is designed as follows:

[0124]

[0125] Where θ is a non-zero real number and s is the sliding surface;

[0126] The approach rate is designed as

[0127] Where s1 is the sliding surface; ε1 and K1 are non-zero coefficients.

[0128] S203: The formula for a two-degree-of-freedom vehicle model is:

[0129]

[0130] In the formula, m is the total vehicle mass; l fThe distance from the front axle to the center of gravity; l r is the distance from the rear axle to the center of mass; v is the velocity at the center of mass; I z C is the moment of inertia of the vehicle's center of mass about an axis perpendicular to the ground; f C represents the front wheel lateral stiffness. r This refers to the rear wheel lateral stiffness.

[0131] Based on the above two-degree-of-freedom vehicle model, for equation e1=e β =β-β d Differentiating both sides, we get:

[0132]

[0133] Where γ is the actual yaw rate.

[0134] Substituting the convergence ratio into the sliding surface, we get:

[0135]

[0136] Combining equations (5) and (6), we can obtain the formula for the rear wheel steering angle:

[0137]

[0138] Where, δ r This refers to the rear wheel steering angle.

[0139] Step S3: Design an additional yaw moment sliding mode controller to eliminate the deviation between the actual vehicle yaw rate and the ideal yaw rate; specifically including:

[0140] S301: Using the actual yaw rate γ and the ideal yaw rate γ d deviation e γ e2 is the sliding mode control variable, and the additional yaw moment ΔM is the output variable u2, i.e., e2 = e γ =γ-γ d u2 = ΔM, the integral sliding surface is designed as follows: These are the proportional coefficient and the integral coefficient, respectively, both of which are non-zero real numbers.

[0141] S302: The approach rate is designed as follows

[0142] Where s2 is the sliding surface; ε2 and K2 are non-zero coefficients.

[0143] S303: Combining a two-degree-of-freedom vehicle model, for equation e2=e γ =γ-γ d Differentiating both sides, we get:

[0144]

[0145] Among them, Iz Let be the moment of inertia of the vehicle's center of mass about an axis perpendicular to the ground.

[0146] Substituting the convergence ratio into the sliding surface, we get:

[0147]

[0148] Combining equations (9) and (10), we can obtain the formula for the additional yaw moment:

[0149]

[0150] Step S4: Based on the deviation e between the actual vehicle speed and the ideal vehicle speed v Design a longitudinal vehicle speed PID controller; specifically including:

[0151] S401, Calculate speed deviation e v :

[0152] e v =v x_t -v x (12);

[0153] In the formula, v x_t Let v be the target vehicle speed, i.e., the ideal vehicle speed. x This refers to the actual vehicle speed;

[0154] S402, Calculate the longitudinal moment of the vehicle:

[0155]

[0156] In the formula, k p k i k d These represent proportional, integral, and differential constants, respectively.

[0157] Step S5: An improved whale optimization algorithm is determined through chaotic mapping initialization, nonlinear update of convergence factor, and adjustment of adaptive inertia weight position. The sliding mode controller from steps S2 and S3 is then self-tuned according to the fitness function requirements. Step S5 specifically includes:

[0158] S501: Input, Output, and Fitness Function: The inputs of an improved whale optimization algorithm include: lateral deviation index J1, velocity deviation index J2, yaw rate index J3, and center of mass sideslip angle index J4, respectively; population size and number of iterations; the outputs include rear wheel steering angle sliding mode controller parameters (e.g., ... ε1, K1) and additional yaw moment sliding mode controller parameters (such as ε2, K2); the fitness function is:

[0159]

[0160] In the formula, c1, c2, c3, and c4 are non-zero weighting coefficients, and e d (t′) represents the lateral deviation at time t′ within a period T, e v (t′) represents the velocity deviation at time t′ within one T cycle, γ(t′) is the yaw rate at time t′ within one T cycle, and β(t′) is the centroid sideslip angle at time t′ within one T cycle.

[0161] S502: Population Chaotic Initialization Design: Chebyshev chaotic sequences are used to initialize the position of the whale population.

[0162] This embodiment uses the Chebyshev chaotic sequence, which has good autocorrelation and balance, to initialize the location of the whale population. The specific operation is as follows:

[0163] Let the chaotic sequence generated in the D-dimensional Euclidean search space be expressed as:

[0164]

[0165] The mathematical expression for generating the Chebyshev chaotic map is:

[0166] y i+1,d =cos(χ·cos -1 y id ),y id ∈[-1,1](16);

[0167] In the formula, χ is the order of the Chebyshev mapping; when using it, a set of pseudo-random numbers y is arbitrarily generated within a given range. 1d ={y 1d After d = 1, 2, ..., D, the uniformly distributed sequence of random numbers y is obtained according to formula (15). N×D This is the Chebyshev chaotic sequence;

[0168] Mapping the chaotic sequence obtained from formula (16) into the algorithm solution space, we obtain the initialized whale population as follows:

[0169]

[0170] Among them, X che For the entire whale population; X i che For the i-th individual whale; X id che X represents the d-th dimension where the i-th individual whale is located; d max X represents the upper limit of the value of the d-th dimension in which the i-th individual whale resides.d min This represents the lower bound of the value of the d-th dimension where the i-th individual whale is located.

[0171] S503: Nonlinear convergence factor design:

[0172] The formula is updated using a nonlinear time-varying update strategy for the convergence factor a:

[0173]

[0174] In the formula, n and m are constants; t is the current iteration number; t max This represents the maximum number of iterations.

[0175] In the early stages of iteration, to enhance the algorithm's ability to escape local optima, the convergence factor 'a' should be relatively large. The formula combines a sine function and an exponentially decaying function, causing 'a' to increase rapidly in the initial phase. The sine term provides the characteristic of rapid change, while the exponential term ensures the continuity and smoothness of the change. In the later stages of iteration, to improve the final convergence accuracy of the algorithm, the convergence factor 'a' should be smaller, and its decreasing rate should be slower.

[0176] S504: Adaptive Inertia Weight Position Adjustment Design

[0177] For some whale individuals X with good fitness i che (t), where there is a high probability that an individual X exists in its local region that can update the global optimal position. best (t), meaning that the solutions of these individuals are better than the current global optimum. To quickly find these individuals and update the global optimum position, the inertia weight w of whale individuals with lower fitness should be reduced. i (t), thereby enhancing its local optimization ability. Conversely, for individuals X with poor fitness... i che (t) represents an individual whose current position has poor quality, and the probability that its region contains a solution better than the current global optimum is low. To escape the current region and explore better solutions, the inertia weight w of these individuals should be increased. i (t), thereby enhancing its global optimization capability.

[0178] Set individual whale X i che The fitness of (t) is f i (t), the average fitness of the population is f min (t) represents the optimal fitness of the current population. Based on the numerical relationship of fitness, the population is divided into two subpopulations, and different inertia weights are used to adjust the position of individuals. The inertia weight design is as follows:

[0179]

[0180] Among them, w i (t) represents the inertia weight of the i-th individual whale in the t-th iteration; w max w represents the upper limit of the inertia weight. min This represents the lower limit of the inertia weight value.

[0181] S505: Main Loop: An improved whale optimization algorithm performs N iterations. Each iteration includes the following steps, see details below. Figure 2 :

[0182] a. Perform population chaos initialization according to the S502 method;

[0183] b. Call the simulation module (Carsim-simulink);

[0184] c. Record and calculate the fitness value f of the best individual in the current population. min (t) and the optimal position X best (t);

[0185] Where, optimal fitness represents the minimum fitness value, denoted by f. min (t) represents;

[0186] d. Update the nonlinear convergence factor a according to formula (18);

[0187] e. Update parameters A, C, and l;

[0188] A = 2ar1 - a; C = 2r2; where r1 and r2 are random vectors in the range [0,1]; a is the convergence factor, which gradually decreases to 0 as the number of iterations increases, expressed as: In the formula, t max This represents the maximum number of iterations. A and C are the coefficient vectors in the whale optimization algorithm; l is a random number between [-1, 1].

[0189] f. Calculate the average fitness value of the population. Calculate the corresponding inertia weight w i (t);

[0190] The average fitness of the population is

[0191] g. Update the individual's position based on the probability value p and the position parameter A (coefficient variable);

[0192] Introducing the inertia weight formula, we obtain the individual position update formula:

[0193]

[0194] Among them, X i (t+1) is the position vector of the i-th individual corresponding to the (t+1)-th iteration, X best (t) is the optimal position vector among all individuals corresponding to the current iteration number; D is the distance vector between the optimal position and the individual position; X rand (t) represents the whale position randomly selected in the t-th iteration; b is the logarithmic spiral shape constant; l is a random number between [-1, 1];

[0195] S506: After N iterations, output the optimal rear wheel steering angle sliding mode controller parameters. K1, ε1, and additional yaw moment sliding mode controller parameters K2, ε2.

[0196] The parameters that need to be determined in sliding mode control are: K1, ε1, K2, ε2. These parameter variables are used as the position components of the whale, with values ​​ranging from [0.01, 500]. The design of the fitness function in the improved whale optimization algorithm (IWOA) is very important. The fitness function characterizes the adaptability of an individual in the population to the current environment and the degree of superiority or inferiority of the individual. To ensure the accuracy of vehicle path tracking, driving stability, and good speed maintenance, a fitness function as shown in formula (14) is designed. Figure 2 The algorithm iterative process obtains the optimal whale position that minimizes the fitness function J through iteration, thus yielding the optimal simulation parameters. K1, ε1, K2, ε2.

[0197] Step S6: According to... Figure 3 The Ackermann steering principle and the dynamic loads on the front and rear axles of the vehicle, as shown, redistribute the vehicle's steering angle and torque, respectively. Step S6 specifically includes:

[0198] S601: Angle distributor receives the front wheel steering angle δ from the single-point aiming controller. f And the rear wheel steering angle δ of the sliding mode controller r And redistribute according to the following Ackermann angle relationship:

[0199]

[0200] In the formula, δ fl The left front wheel steering angle; δ fr The steering angle of the right front wheel; δ rl The left rear wheel steering angle; δ rr The steering angle of the right rear wheel; B f B is the distance between the front wheels;r This refers to the distance between the rear wheels;

[0201] S602: Angle distributor receives additional yaw moment ΔM and longitudinal force F. t The loads are then redistributed according to the following dynamic load estimation relationship between the front and rear axles:

[0202]

[0203] In the formula, T fl T represents the torque of the left front wheel. fr T represents the torque of the right front wheel. rl T represents the torque of the left rear wheel. rr R is the torque of the right rear wheel; R is the wheel radius; F z_fl For the vertical load on the left front wheel; F z_fr The vertical load on the right front wheel; F z_rl For the vertical load on the left rear wheel; F z_rr The load is the vertical load on the right rear wheel.

[0204] Figure 3 This is a schematic diagram of the Ackermann steering principle, where O is the instantaneous center of the vehicle's steering.

[0205] Figure 4 The comparison results for low-adhesion road surface conditions under different control strategies are shown, specifically the changes in vehicle yaw rate and sideslip angle. Low-adhesion road surface condition: set as a double lane change route with a road surface adhesion coefficient of 0.4 and a vehicle speed of 54 km / h to simulate the scenario of a vehicle changing lanes and overtaking on a low-speed, low-adhesion road surface. In this system, Controller 1 controls the Carsim driver model; Controller 2 controls the aiming control, only controlling the front wheel steering for path tracking; Controller 3 controls 4WIS-4WID; and Controller 4 is the method provided in this invention (i.e., IWOA-4WIS-4WID). As shown in the figure, without stability control, the yaw rate and sideslip angle of the vehicle controlled by Controllers 1 and 2 vary significantly. The IWOA-optimized 4WIS-4WID controller proposed in this invention achieves a maximum yaw rate of 13.63° / s, which is reduced by approximately 19.49%, 16.99%, and 4.62% compared to Controllers 1, 2, and 3, respectively. The maximum sideslip angle is 0.0235°, far lower than other control strategies, allowing the wheels to maintain good stability. In other words, the control method provided in this invention can ensure that the maximum yaw rate of the vehicle does not exceed 13.63° / s and the maximum sideslip angle does not exceed 0.0235°.

[0206] Figure 5The comparison results for high-adhesion road surface conditions under different control strategies show the changes in vehicle yaw rate and center of gravity sideslip angle. High-adhesion road surface condition: set as a double lane change route with a road adhesion coefficient of 0.85 and a vehicle speed of 120km / h to simulate the scenario of a vehicle changing lanes and overtaking at high speed on a good road surface. In this design, Controller 1 controls the Carsim driver model; Controller 2 controls the aiming, only controlling the front wheel steering for path tracking; Controller 3 controls 4WIS-4WID; and Controller 4 is the method provided in this invention (i.e., IWOA-4WIS-4WID). As shown in the figure, without stability control, the yaw rate and sideslip angle of the vehicle controlled by Controllers 1 and 2 vary significantly. The IWOA-optimized 4WIS-4WID controller proposed in this paper achieves a maximum yaw rate of 11.68° / s, which is reduced by approximately 38.07%, 28.56%, and 6.78% compared to Controllers 1, 2, and 3, respectively. The maximum sideslip angle is 0.04119°, far lower than other control strategies, indicating that the wheels maintain good stability under this control strategy. In other words, the control method proposed in this invention can ensure that the maximum yaw rate of the vehicle does not exceed 11.68° / s and the maximum sideslip angle does not exceed 0.04119°.

[0207] Summarize:

[0208] (1) For intelligent vehicles with four-wheel independent drive / independent steering systems, a 4WIS+DYC coordinated control method is designed. Compared with AFS and 4WIS vehicles, the 4WIS intelligent electric vehicle based on IWOA optimization can significantly improve the vehicle's path tracking performance and handling stability by reasonably adjusting the steering angle and driving torque of each wheel, thereby greatly improving active safety.

[0209] (2) Under low-speed, low-adhesion road conditions, the path tracking accuracy of vehicles using IWOA-4WIS-4WID is significantly improved, and they also have better driving stability. Under high-speed, high-curvature driving conditions, vehicles using IWOA-4WIS-4WID still have better driving stability, but the path tracking accuracy is still lower than that of AFS vehicles due to the control complexity caused by its high degree of freedom and the inconsistency of the dynamic response of each wheel.

[0210] In addition to the embodiments described above, the present invention may have other implementations. Any changes, modifications, substitutions, combinations, or simplifications made that depart from the spirit and principle of the present invention shall be considered equivalent substitutions and are included within the scope of protection claimed by the present invention.

Claims

1. A method for controlling the driving stability of a distributed drive vehicle, characterized in that, Includes the following steps: Step S1: Based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, establish a single-point anti-alignment deviation model and calculate the front wheel steering angle of the vehicle; Step S2: Design a rear wheel steering angle sliding mode controller to eliminate the deviation between the actual vehicle center of gravity sideslip angle and the ideal center of gravity sideslip angle; Step S3: Design an additional yaw moment sliding mode controller to eliminate the deviation between the actual vehicle yaw rate and the ideal yaw rate; Step S4: Design a longitudinal speed PID controller based on the deviation between the actual vehicle speed and the ideal vehicle speed; Step S5: An improved whale optimization algorithm is determined by initializing the chaotic map, updating the convergence factor nonlinearly, and adjusting the position of the adaptive inertia weight. The sliding mode controllers in steps S2 and S3 are then self-tuned according to the requirements of the fitness function. Step S6: Based on the Ackermann steering principle and the dynamic load on the front and rear axles of the vehicle, the vehicle's steering angle and torque are redistributed respectively.

2. The distributed drive vehicle driving stability control method according to claim 1, characterized in that: In step S1, based on the assumption of constant yaw rate and combined with the mathematical relationship of vehicle steering, a single-point anti-aiming deviation model is established, specifically as follows: Assume the vehicle maintains a constant yaw rate over a future period of time, and the vehicle's lateral speed v y Much smaller than the longitudinal vehicle speed v x Determine the vehicle's appropriate speed The value remains unchanged; Based on this constant yaw rate assumption, the ideal yaw rate can be obtained: Where ΔY is the lateral deviation between the aiming point and the vehicle, and t p The second represents the aiming time, and β represents the vehicle's center of gravity sideslip angle. Combining this with the mathematical relationships governing vehicle steering, the front wheel steering angle can be obtained as follows: Where h is the wheelbase between the front and rear axles of the vehicle, and K is the stability factor.

3. The distributed drive vehicle driving stability control method according to claim 1, characterized in that: Step S2 specifically includes: S201: The deviation of the centroid side slip angle e β e1, as the sliding mode control variable, is the rear wheel rotation angle δ r As the output variable u1, i.e., e1 = e β =β-β d u1=δ r The integral sliding surface is designed as follows: Where β is the actual sideslip angle of the vehicle's center of gravity; β d The ideal centroid sideslip angle; These are the proportional and integral coefficients, respectively, both of which are non-zero real numbers; S202: Using the traditional exponential reaching law Based on this, a continuous curve is used to overcome the hysteresis of sgn(s). The continuous curve is designed as follows: Where θ is a non-zero real number; The approach rate is designed as Where s1 is the sliding surface; ε1 and K1 are non-zero coefficients; S203: Combining the two-degree-of-freedom vehicle model, for equation e1=e β =β-β d Differentiating both sides, we get: Where m is the total vehicle mass; l f l is the distance from the front axle to the center of mass. r C is the distance from the rear axle to the center of mass. f C represents the front wheel lateral stiffness. r γ is the rear wheel lateral stiffness; γ is the actual yaw rate. Substituting the convergence ratio into the sliding surface, we get: Combining equations (5) and (6), we can obtain the formula for the rear wheel steering angle: Where, δ r This refers to the rear wheel steering angle.

4. The distributed drive vehicle driving stability control method according to claim 1, characterized in that: Step S3 specifically includes: S301: Using the actual yaw rate γ and the ideal yaw rate γ d deviation e γ e2 is the sliding mode control variable, and the additional yaw moment ΔM is the output variable u2, i.e., e2 = e γ =γ-γ d u2 = ΔM, the integral sliding surface is designed as follows: These are the proportional coefficient and the integral coefficient, respectively, both of which are non-zero real numbers; S302: The approach rate is designed as follows Where s2 is the sliding surface; ε2 and K2 are non-zero coefficients; S303: Combining a two-degree-of-freedom vehicle model, for equation e2=e γ =γ-γ d Differentiating both sides, we get: Among them, I z Let be the moment of inertia of the vehicle's center of mass about an axis perpendicular to the ground. Substituting the convergence ratio into the sliding surface, we get: Combining equations (9) and (10), we can obtain the formula for the additional yaw moment:

5. The distributed drive vehicle driving stability control method according to claim 1, characterized in that: Step S4 specifically includes: S401, Calculate speed deviation e v : yes v =v x_t -v x (12); In the formula, v x_t For the target vehicle speed, v x This refers to the actual vehicle speed; S402, Calculate the longitudinal moment of the vehicle: In the formula, k p k i k d These represent proportional, integral, and differential constants, respectively.

6. The distributed drive vehicle driving stability control method according to claim 1, characterized in that: Step S5 specifically includes: S501: Input, Output, and Fitness Function: The inputs of an improved whale optimization algorithm include: lateral deviation index J1, velocity deviation index J2, yaw rate index J3, and center of mass sideslip angle index J4, respectively; population size; and number of iterations. The outputs include rear wheel steering angle sliding mode controller parameters. ε1, K1, and additional yaw moment sliding mode controller parameters ε2, K2; Fitness function: In the formula, c1, c2, c3, and c4 are non-zero weighting coefficients, and e d (t′) represents the lateral deviation at time t′ within a period T, e v (t′) represents the velocity deviation at time t′ within one T cycle, γ(t′) is the yaw rate at time t′ within one T cycle, and β(t′) is the centroid sideslip angle at time t′ within one T cycle; S502: Population Chaotic Initialization Design: The position of the whale population is initialized using a Chebyshev chaotic sequence, specifically: Let the chaotic sequence generated in the D-dimensional Euclidean search space be expressed as: The mathematical expression for generating the Chebyshev chaotic map is: and i+1,d =cos(χ·cos -1 and id ),and id ∈[-1,1](16); In the formula, χ is the order of the Chebyshev mapping; when using it, a set of pseudo-random numbers y is arbitrarily generated within a given range. 1d ={y 1d After d = 1, 2, ..., D, the uniformly distributed sequence of random numbers y is obtained according to formula (15). N×D This is the Chebyshev chaotic sequence; Mapping the chaotic sequence obtained from formula (16) into the algorithm solution space, we obtain the initialized whale population as follows: Among them, X che For the entire whale population; X i che For the i-th individual whale; X id che X represents the d-th dimension where the i-th individual whale is located; d max Let X represent the i-th individual whale. i che The upper limit of the value of the d-th dimension, X d min Let X represent the i-th individual whale. i che The lower bound of the value of the d-th dimension; S503: Nonlinear convergence factor design: The formula is updated using a nonlinear time-varying update strategy for the convergence factor a: In the formula, n and m are constants; t is the number of iterations; t max This represents the maximum number of iterations. S504: Adaptive Inertia Weight Position Adjustment Design Set individual whale X i che The fitness of (t) is f i (t), the average fitness of the population is f min (t) represents the optimal fitness of the current population. Based on the numerical relationship of fitness, the population is divided into two subpopulations, and different inertia weights are used to adjust the position of individuals. The inertia weight design is as follows: Among them, w i (t) represents the inertia weight of the i-th individual whale in the t-th iteration; w max w represents the upper limit of the inertia weight. min This indicates the lower limit of the inertia weight value; S505: Main Loop: An improved whale optimization algorithm performs N iterations, each iteration including the following steps: a) Population chaos initialization; b. Call the simulation module; c. Record and calculate the fitness value and position of the best individual in the current population; Where, optimal fitness represents the minimum fitness value, denoted by f. min (t) represents; d. Update the nonlinear convergence factor; e. Update parameters A, C, and l; Where A and C are the coefficient vectors in the whale optimization algorithm; l is a random number between [-1, 1]; f. Calculate the average fitness value of the population and the corresponding inertia weight; The average fitness of the population is g. Update the individual's position based on the probability value p and the position parameter A; S506: After N iterations, output the optimal rear wheel steering angle sliding mode controller parameters. K1, ε1, Additional yaw moment sliding mode controller parameters K2, ε2.

7. The method for controlling the driving stability of a distributed drive vehicle according to claim 1, characterized in that: Step S6 specifically includes: S601: Angle distributor receives the front wheel steering angle δ from the single-point aiming controller. f And the rear wheel steering angle δ of the sliding mode controller r And redistribute according to the following Ackermann angle relationship: In the formula, δ fl The left front wheel steering angle; δ fr The steering angle of the right front wheel; δ rl The left rear wheel steering angle; δ rr The steering angle of the right rear wheel; B f B is the distance between the front wheels; r This refers to the distance between the rear wheels; S602: Angle distributor receives additional yaw moment ΔM and longitudinal force F. t The loads are then redistributed according to the following dynamic load estimation relationship between the front and rear axles: In the formula, T fl T represents the torque of the left front wheel. fr T represents the torque of the right front wheel. rl T represents the torque of the left rear wheel. rr R is the torque of the right rear wheel; R is the wheel radius; F z_fl For the vertical load on the left front wheel; F z_fr The vertical load on the right front wheel; F z_rl For the vertical load on the left rear wheel; F z_rr The load is the vertical load on the right rear wheel.

Images