Multi-strategy optimized automobile tire burst stability hybrid control method
By combining the improved zebra optimization algorithm with the coordinated sliding mode and active steering fuzzy controller, the problem of vehicle body instability after a tire blowout is solved, rapid optimization and global optimal control effects are achieved, and vehicle body stability and safety are improved.
Patent Information
- Application Number
- CN202511125590.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-12
- Publication Date
- 2025-10-28
AI Technical Summary
Existing technologies cannot effectively reduce lateral displacement after a tire blowout, resulting in vehicle body instability. Traditional sliding mode controller parameter tuning is difficult, and swarm intelligence algorithms such as the zebra optimization algorithm are prone to falling into local optimality and cannot achieve better control performance.
A hybrid control method with multi-strategy optimization is adopted, combined with an improved version of the zebra optimization algorithm. The parameters of the sliding mode controller are optimized through the sine-cosine optimization strategy, the adaptive dimension-by-dimension pinhole imaging reverse learning strategy and the Gaussian mutation strategy. A coordinated sliding mode controller and an active steering fuzzy controller model are established to form a hybrid controller, which can achieve rapid optimization and global optimization.
After a tire blowout, the vehicle body stability can be quickly adjusted to reduce lateral displacement and yaw rate, thereby improving body control stability and handling safety.
Smart Images
Figure CN120840587A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle stability control in vehicles with tire blowouts using sliding mode control, and specifically to a hybrid control method for vehicle tire blowout stability optimized by multiple strategies. Background Technology
[0002] Currently, tire blowouts account for a significant proportion of fatal accidents. To reduce such accidents, tire quality monitoring standards have been strengthened in recent years, and vehicles are required to be equipped with tire pressure monitoring systems. However, this system can only proactively adjust the vehicle's stability after a blowout based on the current driving conditions, and driver safety remains seriously threatened. Traditional sliding mode controllers are commonly used in automotive control systems, but their parameter tuning has always been a challenge. Swarm intelligence algorithms, a heuristic algorithm, are introduced to optimally find the fixed parameters in the sliding mode controller. Compared with traditional parameter tuning methods, swarm intelligence algorithms have global search and high parallelism characteristics, enabling more effective searching of the parameter space and improving controller performance. However, this method cannot directly use slip ratio as the control objective, thus failing to achieve better control performance. To improve the control strategy of tire blowout vehicle stability systems, many scholars have conducted research. Among them, some scholars, such as Liu Wei et al., established a sliding mode controller based on the center of gravity sideslip angle and yaw rate, verifying the effectiveness of the sliding mode controller in vehicle stability systems. Li Z et al. combined swarm intelligence algorithms with classical controllers, solving the deficiency of traditional controllers in providing effective data to the system due to parameter issues. The Zebra Optimization Algorithm (ZOA) is a novel swarm intelligence algorithm proposed by Eva Trojovskád et al. in 2022. It seeks optimization by simulating zebra behavior and has strong optimization capabilities and fast convergence speed, making it very suitable for optimizing sliding mode control parameters. However, ZOA is prone to getting trapped in local optima during the optimization process and needs further improvement. Summary of the Invention
[0003] The purpose of this invention is to provide a multi-strategy optimized hybrid control method for vehicle tire blowout stability to overcome the problems existing in the prior art. This invention can effectively reduce the lateral displacement of the vehicle and maintain the vehicle's driving stability when a tire blowout accident occurs.
[0004] To achieve the above objectives, the present invention adopts the following technical solution:
[0005] A multi-strategy optimized hybrid control method for vehicle tire blowout stability includes the following steps:
[0006] Step 1: Establish the car magic tire model and the ideal two-degree-of-freedom model;
[0007] Step 2: Correct the parameters in the car magic tire model according to the changes in the tire blowout parameters to obtain the corrected car magic tire model. In addition, establish a coordinated sliding mode controller model based on yaw rate and center of gravity sideslip angle. At the same time, considering that the direct manifestation of a tire blowout is that the car has a large front wheel steering angle, establish an active steering fuzzy controller model. The coordinated sliding mode controller model and the active steering fuzzy controller model together form a hybrid controller model.
[0008] Step 3: The car tire blowout system model is formed by combining the car magic tire model and ideal two-degree-of-freedom model in Step 1, as well as the modified car magic tire model and hybrid controller model in Step 2.
[0009] Step 4: Based on the factors that greatly affect the inherent parameters of the coordinated sliding mode controller model, the Zebra optimization algorithm is used to optimize the inherent parameters of the coordinated sliding mode controller model;
[0010] Step 5: Based on the requirement that the controller needs to maintain vehicle stability during a tire blowout, the Zebra optimization algorithm in Step 4 is improved to obtain the improved Zebra optimization algorithm, and the Zebra fitness is calculated.
[0011] Step 6: Based on the improved zebra optimization algorithm in Step 5, the optimal fitness is selected after iterative processing, and the corresponding best position coordinates of the optimal fitness are assigned to the inherent parameters in the coordinated sliding mode controller model.
[0012] Step 7: Substitute the inherent parameters obtained in Step 6 into the coordinated sliding mode controller model and apply them to the vehicle tire blowout system model established in Step 3 to control the timely adjustment and stabilization of the vehicle body after a tire blowout.
[0013] Furthermore, the specific formula for the car magic tire model established in step 1 is as follows:
[0014]
[0015] In the formula, y represents the tire force output, which includes the tire longitudinal force, lateral force, and self-aligning torque; x represents the input variables, which include the longitudinal slip ratio k and the wheel slip angle α; B represents the stiffness factor; C represents the shape factor; D represents the peak factor; and E represents the curvature factor.
[0016] The specific formula for the longitudinal force of the tire is as follows:
[0017] df z =(F z -F z0 ) / F z0
[0018] In the formula, F z For the vertical load of the tire; Fz0 Assuming the nominal vertical load, substituting the above formula into the formula for the automobile magic tire model yields the longitudinal force under pure longitudinal slip conditions:
[0019]
[0020] In the calculation of longitudinal forces, the symbol subscript is represented by x; similarly, in the calculation of transverse forces, the symbol subscript is represented by y.
[0021] The parameters in the formula are functions of the vertical load and the camber angle, obtained by fitting parameters from the car's magic tire model. The specific formula is as follows:
[0022]
[0023] In the formula, p cxi p dxi p exi p kxi p vxi p hxi All are longitudinal force weighting factors under pure longitudinal sliding conditions, i = 1, 2, 3, ...; γ is the outward tilt angle; u x K is the longitudinal friction coefficient; x For longitudinal stiffness; S Vx and S Hx The vertical and horizontal drift of the curve relative to the origin are considered. Based on the coupling relationship, the longitudinal force of the vehicle is calculated using a weighted function, as follows:
[0024]
[0025] Where r cxi 、r bxi 、r exi 、r hxi All are longitudinal force weighting factors under combined working conditions, i = 1, 2, ...; G xa (a,k,F z ) is a weighted function of the longitudinal force under combined working conditions, which is the sideslip angle α, slip ratio k, and vertical load F. z The function;
[0026] Similarly, the formula for lateral force is obtained as follows:
[0027]
[0028] Among them, B yk For lateral stiffness factor, C yk For the lateral force shape factor, E yk Lateral force curvature factor, S HYk The lateral force curve shows horizontal drift, and the difference between its expression and the longitudinal force expression under combined loading conditions is that its weighting factor is r. _yi, i = 1, 2, ...;
[0029] An ideal two-degree-of-freedom model is established as follows:
[0030]
[0031] In the formula, β is the sideslip angle of the vehicle's center of gravity; W z I is the yaw rate of the car. z C is the moment of inertia of the car about the Z-axis; f C r V is the lateral stiffness coefficient of the front and rear wheels of a car. x V is the longitudinal speed of the car. y L is the lateral speed of the car. f and L r The distance from the center of gravity to the front and rear axles is given by δ, where δ is the front wheel steering angle and B is the distance from the center of gravity to the front and rear axles, respectively. f and B r The track width is the distance between the front and rear axles;
[0032] The ideal values of the center of mass sideslip angle and yaw rate are calculated based on the ideal two degrees of freedom. These two parameters are the inputs for the subsequent coordinated sliding mode controller model.
[0033] Furthermore, the establishment of the coordinated sliding mode controller model in step 2 is as follows:
[0034] The ideal two-degree-of-freedom formula can be rewritten in the following format:
[0035]
[0036] In the formula, ΔM z To add yaw moment, m is the total vehicle mass, and the left side of the equation represents the first derivative of the yaw rate and the sideslip angle. Taking the yaw rate sliding mode controller as an example, the yaw rate tracking error and its derivative are defined as follows:
[0037]
[0038] In the formula, W z * e represents the actual output value of the ideal two-degree-of-freedom model. wz Input parameters for the sliding mode controller and define the sliding surface as follows:
[0039]
[0040] In the formula, C wz S is a weighting coefficient that is greater than 0 between the deviation and the rate of change of the deviation. wz For the sliding surface, the sliding approach method is chosen to be the constant velocity approach law, that is:
[0041]
[0042] In the formula, K w These are inherent parameters for sliding mode control;
[0043] Finally, substituting the rewritten two-degree-of-freedom formula, we obtain the additional yaw moment formula, as follows:
[0044]
[0045] The tracking error, derivative, sliding surface, and approach method of the center-of-gravity sideslip angle sliding mode controller are the same as those of the yaw rate sliding mode control, i.e.:
[0046]
[0047] In the formula, K β This is another inherent parameter in sliding mode control;
[0048] An active steering fuzzy controller model is established. The membership function of the active steering fuzzy controller model is a Gaussian curve. The input e is the ideal lateral displacement y of the vehicle. * The deviation of the actual lateral displacement y, with a universe of discourse of [-2,2] and a fuzzy subset of [S1,S2,M1,M2,L1,L2], is the second input. The universe of discourse is [-2,2], and the fuzzy subset is [s1,s2,m1,m2,l1,l2]. The system outputs the front wheel steering angle δ, with a universe of discourse of [-6,6] and a fuzzy subset of [X1,X2,X3,Y1,Y2,Y3,Z1,Z2,Z3]. Fuzzy rules are established through membership functions. This active steering fuzzy controller uses lateral displacement as the control parameter and dynamically compensates for the front wheel steering angle of the car by defuzzifying the rules and considering the driving conditions of the car with a tire blowout.
[0049] Furthermore, step 4 of the Zebra optimization algorithm includes the following steps:
[0050] Step 4.1: Randomly initialize the zebra population;
[0051] Step 4.2: After initializing the zebra population, each zebra in the population begins to forage and updates its behavior and location accordingly;
[0052] Step 4.3: Considering the predators that zebra populations may encounter during foraging, and that zebra populations will adopt different defense strategies against different predators, update the location based on the predator defense strategies.
[0053] Furthermore, in step 4.2, regarding the zebra foraging behavior, the zebra with the optimal position in the population is considered the vanguard zebra, leading other members to a better position. The specific formula is as follows:
[0054]
[0055] In the formula, x i,j new,P1 x is the position parameter of the i-th zebra after updating in the j-th dimension. i,j The zebra's position parameter before this stage; r is a random number in the range [0,1]; I is a random number in the range [1,2]; PZ j The optimal position parameters for the zebra population during the entire optimization process are calculated using the fitness function. The fitness of the current position and the position before the update is then used to update the position of the i-th zebra according to the above formula.
[0056] In step 4.3, when the zebra herd encounters a predator, it begins to defend itself, choosing opportune moments to attack weaker predators and fleeing from larger ones. During this process, the zebra herd's position is updated as follows:
[0057]
[0058] In the formula, t is the current iteration number, T is the set iteration number, and S1 and S2 are the objective functions for Zebra to select defense and attack strategies, respectively; x i,j Let R be the position of the i-th zebra individual; R is a constant of 0.01; T is the maximum number of iterations; P s Probabilities for two types of responses; AZ j This refers to the state of the zebra being attacked.
[0059] Furthermore, step 5 involves improving the Zebra optimization algorithm to obtain an improved Zebra optimization algorithm, specifically including the following steps:
[0060] Step 5.1: Randomly initialize the zebra population;
[0061] Step 5.2: After determining the initial location of the zebra population, begin preparing for algorithm iteration and updates;
[0062] Step 5.3: The zebra population prepares for foraging behavior. Before this, a sine and cosine optimization strategy is used to enable the zebra population to search in a larger spatial range, increase the foraging area, and enhance the flexibility of the algorithm.
[0063] Step 5.4: After the zebra population completes the search for foraging area, it prepares to forage. After that, the zebra population is likely to encounter predators. An adaptive dimensional pinhole imaging reverse learning strategy is adopted to make corresponding adjustments when zebras face large predators. Therefore, the position update formula for the defense phase of large predators in the algorithm calculation formula is revised.
[0064] Step 5.5: Considering the diversity of zebra predators, zebra populations need to adopt different coping strategies when encountering small predators. At this time, the Gaussian mutation strategy is used to revise the update formula of the zebra's defensive position when facing small predators, thereby improving population diversity.
[0065] Step 5.6: Calculate the fitness of the zebra population and select the pioneer zebra based on the fitness level;
[0066] Step 5.7: Determine if the maximum number of iterations has been reached. If yes, output the pioneer zebra's position coordinates and fitness. If no, return to step 5.3.
[0067] Furthermore, in step 5.1, the zebra population is randomly initialized using the following formula:
[0068] x i,j =lb j +r·(ub j -lb j )
[0069] In the formula, x i,j The initial position of a zebra individual in the population, lb j To find the optimal lower boundary, ub j To optimize the upper boundary, r is a random number in the range [0,1].
[0070] Furthermore, step 5.3 employs a sine and cosine strategy to make the zebras move more flexibly. The specific formula is as follows:
[0071]
[0072] In the formula, X new It is the zebra jumping out of the local optimum to a new position; X new,P1 r1 is the best location the zebra finds while foraging; a is a natural number greater than 0; r2 is a random number in [2, 2π]; r3 is a random number in [0, 2] used to control the distance between the individual and the optimal solution; r4 is the transition probability, a random number in [0, 1].
[0073] Furthermore, in step 5.4, the position update formula for the defense phase of large predators is revised by combining the adaptive dimensional pinhole imaging back-learning strategy, specifically as follows:
[0074]
[0075] In the formula, a j and b j These are the upper and lower bounds of the j-th dimension, respectively, where n is the adjustment parameter, and X' best (t) is the optimal solution obtained in the t-th iteration, representing the approximate position of the pioneer zebra, T is the maximum number of iterations, (1-t / T) is used to control the zebra's movement, and X M (t) is the average position of the current solution at the t-th iteration;
[0076] Step 5.5 uses a Gaussian mutation strategy to revise the zebra's defensive position update formula when facing small predators, specifically as follows:
[0077]
[0078] In the formula, σ is the variance of the Gaussian function, and X new1 This represents the optimal position for the current iteration of the population search.
[0079] Furthermore, in step 6, based on the pioneer zebra position coordinates obtained in step 5.7, the coordinate parameters are assigned to the inherent parameters in the coordinated sliding mode controller model. Specifically, the method is as follows:
[0080] Zebra populations exhibit three movement processes: foraging, defense against predators, and attack against predators. Each process involves position updates. By calculating the fitness value of individual zebras, the position coordinates of the zebra with the lowest fitness value are selected, and then iterative updates are performed. During the iteration process, if a smaller fitness value exists, the position coordinates of the zebra is replaced. After the zebra population iteration is completed, the three-dimensional position coordinates of the lead zebra are assigned to the inherent parameters in the coordinated sliding mode controller model.
[0081] Compared with the prior art, the present invention has the following beneficial technical effects:
[0082] This invention proposes a zebra optimization algorithm based on sine and cosine backward learning combined with Gaussian mutation. The sine and cosine optimization strategy enables the algorithm to find the optimal solution in a large solution space. An adaptive, dimensional pinhole imaging backward learning strategy is proposed to increase the zebra population's ability to formulate appropriate strategies and continue searching for the optimal solution when facing large predators. A Gaussian mutation strategy is proposed to increase population diversity and avoid the optimal solution being a local optimum, thus preventing the achievement of a global optimum. (See attached...) Figure 3 The fitness curve shows that, under the premise of ensuring rapid optimization, the optimal sliding mode control parameters are obtained. This invention has a better control effect after the car experiences a thinning.
[0083] Furthermore, this invention employs a sine and cosine optimization strategy during the zebra foraging behavior stage, which solves the problem of premature convergence of ZOA at this stage, preventing the complete search of the spatial region; and employs an adaptive small-hole dimension-by-dimensional backward learning strategy during the predator defense stage, which is prone to getting stuck in local extrema.
[0084] Compared to existing Zebra optimization algorithms, this invention achieves faster convergence, smaller lateral displacement of the vehicle, and lower yaw rate. Attached Figure Description
[0085] The accompanying drawings are provided to further understand the invention and constitute a part of this invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.
[0086] Figure 1This is a flowchart illustrating the present invention;
[0087] Figure 2 This is a flowchart of the improved zebra optimization algorithm involved in this invention;
[0088] Figure 3 The present invention relates to a hybrid controller structure diagram, wherein (a) is the test result of a high-dimensional function algorithm, (b) is the test result of a non-convex function algorithm, (c) is the test result of a multi-peak function algorithm, (d) is the test result of a nonlinear function algorithm, (e) is the test result of a function algorithm with interference, and (f) is the test result of a complex function.
[0089] Figure 4 The diagram shows the control effect of this invention after a tire blowout in a car. (a) shows the lateral displacement control result, (b) shows the lateral acceleration control result, (c) shows the yaw rate control result, and (d) shows the center of gravity sideslip angle control result. Detailed Implementation
[0090] The present invention will now be described in further detail with reference to the accompanying drawings:
[0091] See Figure 1 The flowchart of this invention shows that the basic zebra optimization algorithm has been optimized and improved, and a zebra optimization algorithm based on sine and cosine backward learning combined with Gaussian mutation is proposed, such as... Figure 2 This method is applied to optimize sliding mode controllers. First, a car magic tire model and an ideal two-degree-of-freedom model are established. Second, a sine and cosine optimization strategy is designed to allow the zebra population to escape local optima and search in a larger spatial region. Then, an adaptive, dimensional pinhole imaging-based back-learning strategy is designed to more effectively avoid local extrema. Finally, a Gaussian mutation strategy is used to revise the zebra's defensive position update formula when facing small predators. Figure 3 and Figure 4 It can be seen that ZSWA converges faster, has a shorter lateral displacement, and reduces all expected parameters to varying degrees compared to the Zebra algorithm.
[0092] The specific steps are as follows:
[0093] Step 1: Establish the car magic tire model and the ideal two-degree-of-freedom model, as follows:
[0094]
[0095] In the formula, y represents the tire force output, such as longitudinal force, lateral force, and self-aligning torque; x represents the input variables such as longitudinal slip ratio k or wheel slip angle α; B is the stiffness factor; C is the shape factor; D is the peak factor; and E is the curvature factor.
[0096] Tires primarily output longitudinal and lateral forces. A car generates longitudinal force under pure longitudinal slip conditions and lateral force under pure sideslip conditions. However, during normal driving, a car is not in a purely longitudinal slip or purely sideslip condition, but rather a combined condition with certain coupling relationships. Taking longitudinal force as an example, its specific formula is as follows:
[0097] df z =(F z -F z0 ) / F z0
[0098] In the formula, F z For the vertical load of the tire; F Z0 Let be the nominal vertical load. Substituting the above equation into the basic formula, we obtain the longitudinal force under pure longitudinal sliding conditions:
[0099]
[0100] The parameters in the formula are functions of the vertical load and the camber angle, obtained by fitting parameters from the car's magic tire model. The specific formula is as follows:
[0101]
[0102] In the formula, p cxi p dxi p exi p kxi p vxi p hxi (i = 1, 2, 3...) represents the longitudinal force weighting factor under pure longitudinal sliding conditions, which is a constant parameter; γ is the outward tilt angle; u x K is the longitudinal friction coefficient; x For longitudinal stiffness; S Vx and S Hx This represents the vertical and horizontal drift of the curve relative to the origin. Based on the coupling relationship, the longitudinal force of the vehicle is calculated using a weighted function, as follows:
[0103]
[0104] Where r cxi 、r bxi 、r exi 、r hxi (i = 1, 2, 3...) are all longitudinal force weighting factors under combined working conditions, and are constant parameters; G xa (a,k,F z ) is a weighted function of the longitudinal force under combined working conditions, which is the sideslip angle α, slip ratio k, and vertical load F. z The function.
[0105] Similarly, the formula for lateral force is obtained as follows:
[0106]
[0107] Among them, B yk For lateral stiffness factor, C yk For the lateral force shape factor, E yk Lateral force curvature factor, S HYk The lateral force curve shows horizontal drift, and the difference between its expression and the longitudinal force expression under combined loading conditions is that its weighting factor is r. _yi (i = 1, 2, ...).
[0108] An ideal two-degree-of-freedom model is established as follows:
[0109]
[0110] In the formula, β is the sideslip angle of the vehicle's center of gravity; W z I is the yaw rate of the car. z C is the moment of inertia of the car about the Z-axis; f C r V is the lateral stiffness coefficient of the front and rear wheels of a car. x V is the longitudinal speed of the car. y L is the lateral speed of the car. f and L r The distance from the center of gravity to the front and rear axles is given by δ, where δ is the front wheel steering angle and B is the distance from the center of gravity to the front and rear axles, respectively. f and B r The track width is the distance between the front and rear axles.
[0111] Step 2: Correct the parameters in the car magic tire model based on the changes in the tire blowout parameters to obtain the corrected car magic tire model. In addition, establish a coordinated sliding mode controller model based on yaw rate and center of gravity sideslip angle. At the same time, considering that the direct manifestation of a tire blowout is that the car has a large front wheel steering angle, establish an active steering fuzzy controller model. The coordinated sliding mode controller model and the active steering fuzzy controller model together form a hybrid controller model.
[0112] The details are as follows:
[0113] The ideal two-degree-of-freedom formula can be rewritten in the following format:
[0114]
[0115] In the formula, ΔM z To add yaw moment, m is the total vehicle mass, and the left side of the equation represents the first derivative of the yaw rate and the sideslip angle. Taking the yaw rate sliding mode controller as an example, the yaw rate tracking error and its derivative are defined as follows:
[0116]
[0117] In the formula, Wz * e represents the actual output value of the ideal two-degree-of-freedom model. wz Input parameters for the sliding mode controller and define the sliding surface as follows:
[0118]
[0119] In the formula, C wz S is a weighting coefficient that is greater than 0 between the deviation and the rate of change of the deviation. wz For the sliding surface, the sliding approach method is chosen to be the constant velocity approach law, that is:
[0120]
[0121] In the formula, K w These are inherent parameters for sliding mode control;
[0122] Finally, substituting the rewritten two-degree-of-freedom formula, we obtain the additional yaw moment formula, as follows:
[0123]
[0124] The tracking error, derivative, sliding surface, and approach method of the center-of-gravity sideslip angle sliding mode controller are the same as those of the yaw rate sliding mode control, i.e.:
[0125]
[0126] In the formula, K β This is another inherent parameter in sliding mode control;
[0127] An active steering fuzzy controller model is established. The membership function of the active steering fuzzy controller model is a Gaussian curve. The input e is the ideal lateral displacement y of the vehicle. * The deviation of the actual lateral displacement y, with a universe of discourse of [-2,2] and a fuzzy subset of [S1,S2,M1,M2,L1,L2], is the second input. The universe of discourse is [-2,2], with a fuzzy subset [s1,s2,m1,m2,l1,l2]. The system outputs the front wheel steering angle δ, with a universe of discourse [-6,6] and a fuzzy subset [X1,X2,X3,Y1,Y2,Y3,Z1,Z2,Z3]. Fuzzy rules are established through membership functions. The symbols in the fuzzy subsets represent ranges for convenience and have no actual meaning. This active steering fuzzy controller model uses lateral displacement as the control parameter and dynamically compensates for the front wheel steering angle by decomposing the fuzzy rules and considering the driving conditions of a vehicle with a tire blowout.
[0128] Step 3: The tire blowout system model is completed. The tire blowout system model is composed of the car magic tire model, the ideal two-degree-of-freedom model, the modified car magic tire model, and the hybrid controller model.
[0129] Step 4: To coordinate factors that significantly influence the inherent parameters of the sliding mode controller model, the Zebra Optimization Algorithm is used to optimize these inherent parameters. The basic Zebra Optimization Algorithm consists of the following steps:
[0130] The zebra's behavior of searching for pasture during foraging involves the zebra in the optimal position within the herd acting as the vanguard zebra, leading other members to a better location. The specific formula for this stage is as follows:
[0131]
[0132] In the formula, x i,j new,P1 x is the position parameter of the i-th zebra after updating in the j-th dimension. i,j The position parameter of the individual before this stage; r is a random number in the range [0,1]; I is a random number in the range [1,2]; PZ j The optimal position parameters for the zebra population are defined throughout the optimization process. The fitness of the current position compared to the previous position is calculated using a fitness function, and the position of the i-th zebra is updated according to the above formula.
[0133] When zebra herds encounter predators, they initially adopt a defensive posture, attacking weaker predators when the opportunity arises, and fleeing from larger predators. During this process, the zebra herd's location updates as follows:
[0134]
[0135] In the formula, t is the current iteration number, and T is the set iteration number. S1 and S2 are the objective functions for Zebra to select defense and attack strategies, respectively; x i,j Let R be the i-th zebra individual; R is a constant of 0.01; T is the maximum number of iterations; P s Probabilities for two types of responses; AZ j This refers to the state of the zebra being attacked.
[0136] Step 5: Based on the requirement that the controller needs to maintain vehicle stability during a tire blowout, the Zebra optimization algorithm in Step 4 is improved to obtain the improved Zebra optimization algorithm. The Zebra fitness is then calculated, as follows:
[0137] First, the zebra population is randomly initialized using the following formula:
[0138] x i,j =lb j +r·(ub j -lb j )
[0139] In the formula, x i,jLet represent the initial position of the zebra individual in the population, lbj represent the lower boundary of the optimization, ubj represent the upper boundary of the optimization, and r represent a random number in the range [0,1].
[0140] Biological evolutionary strategies have enabled the eagle population to evolve, and the specific steps are as follows:
[0141] The individual position is optimized using a sine and cosine strategy, as shown in the following formula:
[0142]
[0143] In the formula, X new It is the zebra jumping out of the local optimum to a new position; X new,P1 r1 is the best location the zebra finds while foraging; a is a natural number greater than 0; r2 is a random number in [2, 2π]; r3 is a random number in [0, 2] used to control the distance between the individual and the optimal solution; r4 is the transition probability, a random number in [0, 1].
[0144] By combining an adaptive dimensional pinhole imaging back-learning strategy, the defense formula against large predators is revised as follows:
[0145]
[0146] In the formula, a j and b j These are the upper and lower bounds of the j-th dimension, respectively, where n is the adjustment parameter, and X' best (t) is the optimal solution obtained in the t-th iteration, representing the approximate position of the pioneer zebra, T is the maximum number of iterations, (1-t / T) is used to control the zebra's movement, and X M (t) is the average position of the current solution at the t-th iteration.
[0147] The position update formula for zebras facing small predators is revised using a Gaussian mutation strategy, as follows:
[0148]
[0149] In the formula, σ is the variance of the Gaussian function, and X... new1 This represents the optimal position for the current iteration of the population search.
[0150] Step 6: Based on the improved zebra optimization algorithm in Step 5, the optimal fitness is selected after iterative processing, and the corresponding best position coordinates of the optimal fitness are assigned to the inherent parameters in the coordinated sliding mode controller model to optimize the performance of the hybrid controller and obtain the optimal parameters.
[0151] The improved Zebra optimization algorithm is used to optimize PID parameters, specifically as follows:
[0152] Zebra populations exhibit three movement phases: foraging, defense against predators, and attack against predators. Each phase involves position updates. By calculating the fitness value of individual zebras, the position coordinates of the zebra with the lowest fitness value are selected, and then iterative updates are performed. During iteration, if a smaller fitness value exists, the zebra's position coordinates are replaced. After the zebra population iteration is complete, the three-dimensional position coordinates of the lead zebra are assigned to K in the coordinated sliding mode controller model. w K β Inherent parameters.
[0153] Step 7: Substitute the inherent parameters obtained in Step 6 into the coordinated sliding mode controller model and apply it to the vehicle tire blowout system model established in Step 3. Control the vehicle to adjust and stabilize the body in a timely manner after a tire blowout, and observe the control effect. Figure 4 As shown.
[0154] Figure 3 To compare the fitness curves of the ZSWA algorithm of this invention with other algorithms in searching for the global optimum, (a) tests the algorithm's search ability by maximizing or minimizing the output of the function; (b), (c), and (d) evaluate the algorithm's search ability in high-dimensional, non-convex, multimodal, and nonlinear optimization problems; (e) evaluates the algorithm's performance and robustness under noisy conditions; and (f) simulates complex real-world problems to verify the algorithm's solution ability under complex conditions. Figure 3 It can be seen that there is a certain gap between the unoptimized algorithm and the optimized algorithm ZSWA, which verifies the effectiveness of the optimization strategy. Adding a controller can effectively improve the controller performance.
[0155] Figure 4 The control curves of the ZSWA-optimized hybrid controller proposed in this invention are compared with other single controllers. Figure (a) shows the lateral displacement curves of different controllers during a tire blowout. Figure 4 Figure (b) shows the lateral acceleration curves of different controllers during a tire blowout; Figure (c) shows the yaw rate curves of different controllers during a tire blowout; Figure (d) shows the sideslip angle curves of different controllers during a tire blowout. Figure 4 It can be seen that the ZSWA-optimized hybrid controller can effectively reduce the deviation of various parameters when a tire blowout occurs, which can enable the vehicle to achieve high control stability and handling safety.
[0156] 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 its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the invention, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims of the invention.
Claims
1. A multi-strategy optimized hybrid control method for vehicle tire blowout stability, characterized in that, Includes the following steps: Step 1: Establish the car magic tire model and the ideal two-degree-of-freedom model; Step 2: Correct the parameters in the car magic tire model according to the changes in the tire blowout parameters to obtain the corrected car magic tire model. In addition, establish a coordinated sliding mode controller model based on yaw rate and center of gravity sideslip angle. At the same time, considering that the direct manifestation of a tire blowout is that the car has a large front wheel steering angle, establish an active steering fuzzy controller model. The coordinated sliding mode controller model and the active steering fuzzy controller model together form a hybrid controller model. Step 3: The car tire blowout system model is formed by combining the car magic tire model and ideal two-degree-of-freedom model in Step 1, as well as the modified car magic tire model and hybrid controller model in Step 2. Step 4: Based on the factors that greatly affect the inherent parameters of the coordinated sliding mode controller model, the Zebra optimization algorithm is used to optimize the inherent parameters of the coordinated sliding mode controller model; Step 5: Based on the requirement that the controller needs to maintain vehicle stability during a tire blowout, the Zebra optimization algorithm in Step 4 is improved to obtain the improved Zebra optimization algorithm, and the Zebra fitness is calculated. Step 6: Based on the improved zebra optimization algorithm in Step 5, the optimal fitness is selected after iterative processing, and the corresponding best position coordinates of the optimal fitness are assigned to the inherent parameters in the coordinated sliding mode controller model. Step 7: Substitute the inherent parameters obtained in Step 6 into the coordinated sliding mode controller model and apply them to the vehicle tire blowout system model established in Step 3 to control the timely adjustment and stabilization of the vehicle body after a tire blowout.
2. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 1, characterized in that, The specific formula for the car magic tire model established in step 1 is as follows: In the formula, y represents the tire force output, which includes the tire longitudinal force, lateral force, and self-aligning torque; x represents the input variables, which include the longitudinal slip ratio k and the wheel slip angle α; B represents the stiffness factor; C represents the shape factor; D represents the peak factor; and E represents the curvature factor. The specific formula for the longitudinal force of the tire is as follows: df z =(F z -F z0 ) / F z0 In the formula, F z For the vertical load of the tire; F z0 Assuming the nominal vertical load, substituting the above formula into the formula for the automobile magic tire model yields the longitudinal force under pure longitudinal slip conditions: In the calculation of longitudinal forces, the symbol subscript is represented by x; similarly, in the calculation of transverse forces, the symbol subscript is represented by y. The parameters in the formula are functions of the vertical load and the camber angle, obtained by fitting parameters from the car's magic tire model. The specific formula is as follows: In the formula, p cxi p dxi p exi p kxi p vxi p hxi All are longitudinal force weighting factors under pure longitudinal sliding conditions, i = 1, 2, 3, ...; γ is the outward tilt angle; u x K is the longitudinal friction coefficient; x For longitudinal stiffness; S Vx and S Hx The vertical and horizontal drift of the curve relative to the origin are considered. Based on the coupling relationship, the longitudinal force of the vehicle is calculated using a weighted function, as follows: Where r cxi 、r bxi 、r exi 、r hxi All are longitudinal force weighting factors under combined working conditions, i = 1, 2, ...; G xa (a,k,F z ) is a weighted function of the longitudinal force under combined working conditions, which is the sideslip angle α, slip ratio k, and vertical load F. z The function; Similarly, the formula for lateral force is obtained as follows: Among them, B yk For lateral stiffness factor, C yk For the lateral force shape factor, E yk Lateral force curvature factor, S HYk The lateral force curve shows horizontal drift, and the difference between its expression and the longitudinal force expression under combined loading conditions is that its weighting factor is r. _yi , i = 1, 2, ...; An ideal two-degree-of-freedom model is established as follows: In the formula, β is the sideslip angle of the vehicle's center of gravity; W z I is the yaw rate of the car. z C is the moment of inertia of the car about the Z-axis; f C r V is the lateral stiffness coefficient of the front and rear wheels of a car. x V is the longitudinal speed of the car. y L is the lateral speed of the car. f and L r The distance from the center of gravity to the front and rear axles is given by δ, where δ is the front wheel steering angle and B is the distance from the center of gravity to the front and rear axles, respectively. f and B r The track width is the distance between the front and rear axles; The ideal values of the center of mass sideslip angle and yaw rate are calculated based on the ideal two degrees of freedom. These two parameters are the inputs for the subsequent coordinated sliding mode controller model.
3. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 2, characterized in that, Step 2 involves establishing the sliding mode controller model, as detailed below: The ideal two-degree-of-freedom formula can be rewritten in the following format: In the formula, ΔM z To add yaw moment, m is the total vehicle mass, and the left side of the equation represents the first derivative of the yaw rate and the sideslip angle. Taking the yaw rate sliding mode controller as an example, the yaw rate tracking error and its derivative are defined as follows: In the formula, W z * e represents the actual output value of the ideal two-degree-of-freedom model. wz Input parameters for the sliding mode controller and define the sliding surface as follows: In the formula, C wz S is a weighting coefficient that is greater than 0 between the deviation and the rate of change of the deviation. wz For the sliding surface, the sliding approach method is chosen to be the constant velocity approach law, that is: In the formula, K w These are inherent parameters for sliding mode control; Finally, substituting the rewritten two-degree-of-freedom formula, we obtain the additional yaw moment formula, as follows: The tracking error, derivative, sliding surface, and approach method of the center-of-gravity sideslip angle sliding mode controller are the same as those of the yaw rate sliding mode control, i.e.: In the formula, K β This is another inherent parameter in sliding mode control; An active steering fuzzy controller model is established. The membership function of the active steering fuzzy controller model is a Gaussian curve. The input e is the ideal lateral displacement y of the vehicle. * The deviation of the actual lateral displacement y, with a universe of discourse of [-2,2] and a fuzzy subset of [S1,S2,M1,M2,L1,L2], is the second input. The universe of discourse is [-2,2], and the fuzzy subset is [s1,s2,m1,m2,l1,l2]. The system outputs the front wheel steering angle δ, with a universe of discourse of [-6,6] and a fuzzy subset of [X1,X2,X3,Y1,Y2,Y3,Z1,Z2,Z3]. Fuzzy rules are established through membership functions. This active steering fuzzy controller uses lateral displacement as the control parameter and dynamically compensates for the front wheel steering angle of the car by defuzzifying the rules and considering the driving conditions of the car with a tire blowout.
4. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 3, characterized in that, Step 4 of the Zebra optimization algorithm includes the following steps: Step 4.1: Randomly initialize the zebra population; Step 4.2: After initializing the zebra population, each zebra in the population begins to forage and updates its behavior and location accordingly; Step 4.3: Considering the predators that zebra populations may encounter during foraging, and that zebra populations will adopt different defense strategies against different predators, update the location based on the predator defense strategies.
5. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 4, characterized in that, In step 4.2, regarding zebra foraging behavior, the zebra with the optimal position in the population is considered the lead zebra, guiding other members to a better position. The specific formula is as follows: In the formula, x i,j new,P1 x is the position parameter of the i-th zebra after updating in the j-th dimension. i,j The zebra's position parameter before this stage; r is a random number in the range [0,1]; I is a random number in the range [1,2]; PZ j The optimal position parameters for the zebra population during the entire optimization process are calculated using the fitness function. The fitness of the current position and the position before the update is then used to update the position of the i-th zebra according to the above formula. In step 4.3, when the zebra herd encounters a predator, it begins to defend itself, choosing opportune moments to attack weaker predators and fleeing from larger ones. During this process, the zebra herd's position is updated as follows: In the formula, t is the current iteration number, T is the set iteration number, and S1 and S2 are the objective functions for Zebra to select defense and attack strategies, respectively; x i,j Let R be the position of the i-th zebra individual; R is a constant of 0.01; T is the maximum number of iterations; P s Probabilities for two types of responses; AZ j This refers to the state of the zebra being attacked.
6. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 5, characterized in that, Step 5 involves improving the Zebra optimization algorithm to obtain the improved Zebra optimization algorithm, which specifically includes the following steps: Step 5.1: Randomly initialize the zebra population; Step 5.2: After determining the initial location of the zebra population, begin preparing for algorithm iteration and updates; Step 5.3: The zebra population prepares for foraging behavior. Before this, a sine and cosine optimization strategy is used to enable the zebra population to search in a larger spatial range, increase the foraging area, and enhance the flexibility of the algorithm. Step 5.4: After the zebra population completes the search for foraging area, it prepares to forage. After that, the zebra population is likely to encounter predators. An adaptive dimensional pinhole imaging reverse learning strategy is adopted to make corresponding adjustments when zebras face large predators. Therefore, the position update formula for the defense phase of large predators in the algorithm calculation formula is revised. Step 5.5: Considering the diversity of zebra predators, zebra populations need to adopt different coping strategies when encountering small predators. At this time, the Gaussian mutation strategy is used to revise the update formula of the zebra's defensive position when facing small predators, thereby improving population diversity. Step 5.6: Calculate the fitness of the zebra population and select the pioneer zebra based on the fitness level; Step 5.7: Determine if the maximum number of iterations has been reached. If yes, output the pioneer zebra's position coordinates and fitness. If no, return to step 5.
3.
7. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 6, characterized in that, In step 5.1, the zebra population is randomly initialized using the following formula: x i,j =lb j +r·(ub j -lb j ) In the formula, x i,j The initial position of a zebra individual in the population, lb j To find the optimal lower boundary, ub j To optimize the upper boundary, r is a random number in the range [0,1].
8. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 6, characterized in that, Step 5.3 employs a sine and cosine strategy to make the zebras move more flexibly. The specific formula is as follows: In the formula, X new It is the zebra jumping out of the local optimum to a new position; X new,P1 r1 is the best location the zebra finds while foraging; a is a natural number greater than 0; r2 is a random number in [2, 2π]; r3 is a random number in [0, 2] used to control the distance between the individual and the optimal solution; r4 is the transition probability, a random number in [0, 1].
9. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 6, characterized in that, In step 5.4, the position update formula for the defense phase of large predators is revised by combining the adaptive dimensional pinhole imaging back-learning strategy, as follows: In the formula, a j and b j These are the upper and lower bounds of the j-th dimension, respectively, where n is the adjustment parameter, and X' best (t) is the optimal solution obtained in the t-th iteration, representing the approximate position of the pioneer zebra, T is the maximum number of iterations, (1-t / T) is used to control the zebra's movement, and X M (t) is the average position of the current solution at the t-th iteration; Step 5.5 uses a Gaussian mutation strategy to revise the zebra's defensive position update formula when facing small predators, specifically as follows: In the formula, σ is the variance of the Gaussian function, and X new1 This represents the optimal position for the current iteration of the population search.
10. The multi-strategy optimized hybrid control method for vehicle tire blowout stability according to claim 6, characterized in that, In step 6, based on the pioneer zebra position coordinates obtained in step 5.7, the coordinate parameters are assigned to the inherent parameters in the coordinated sliding mode controller model. Specifically, the method is as follows: Zebra populations exhibit three movement processes: foraging, defense against predators, and attack against predators. Each process involves position updates. By calculating the fitness value of individual zebras, the position coordinates of the zebra with the lowest fitness value are selected, and then iterative updates are performed. During the iteration process, if a smaller fitness value exists, the position coordinates of the zebra is replaced. After the zebra population iteration is completed, the three-dimensional position coordinates of the lead zebra are assigned to the inherent parameters in the coordinated sliding mode controller model.