Torque distribution control method for distributed intelligent electric vehicles
Through the torque distribution control method of the layered structure, the expected yaw torque of the sliding mode is calculated and adaptive compensation is performed. Combined with PID control, the wheel torque is corrected, and the problem of distributed electric drive vehicles is solved and the problem of load uncertainty is improved. Safety and stability and motion tracking capabilities.
Patent Information
- Application Number
- CN202210738853.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-28
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2042-06-28
AI Technical Summary
Distributed electric-driven vehicles are sensitive to changes in load parameters. Existing research rarely considers the impact of load uncertainty on system control, resulting in insufficient safety and stability.
The torque distribution control method with a layered structure includes the upper layer calculating the expected yaw torque of the sliding mode and performing adaptive compensation, the lower layer calculating the output torque of the four wheels, and eliminating the impact of load uncertainty through the sliding mode function and neural network. PID control is used to correct wheel torque.
The control performance of distributed electric drive vehicles on load uncertainty is improved, the safety and stability of the system and motion tracking capabilities are enhanced, and the impact of load uncertainty on system control is eliminated.
Smart Images

Figure CN114889450B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent electric vehicles, and in particular relates to a torque distribution control method for a distributed intelligent electric vehicle. Background Art
[0002] Pure electric vehicles have the advantages of low environmental pollution, wide energy sources, and low vehicle costs, and have good development prospects. Distributed drive electric vehicles use motors to directly drive vehicle wheels, with outstanding features such as short transmission chains and high transmission efficiency. The motor is both an information unit and a fast-response execution unit of the vehicle. By reasonably distributing the torque of the motors of each drive wheel, direct yaw torque and drive anti-skid control can be achieved, thereby improving the active safety of the vehicle.
[0003] Distributed electric drive vehicles eliminate the bulky engines, mechanical transmissions and other mechanical equipment of traditional vehicles, which reduces the weight of the vehicle equipment. This also means that distributed electric drive vehicles are relatively sensitive to changes in load parameters. In existing research, the focus is mainly on model-based control strategies, while the impact of load uncertainty is less considered. Summary of the invention
[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned technology and to provide a torque distribution control method for a distributed intelligent electric vehicle, which adopts a hierarchical structure: the yaw moment solution method can improve the system's control performance under load uncertainty; the torque distribution method can improve the safety and stability of the distributed electric drive vehicle.
[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a torque distribution control method for a distributed intelligent electric vehicle, characterized in that: a hierarchical structure is adopted: the upper layer is a yaw moment decision layer, and the lower layer is a torque distribution layer; the sliding mode expected yaw moment is calculated in the upper layer of the control system and adaptively compensated; the output torque of the four wheels is calculated in the lower layer of the control system, so as to eliminate the influence of load uncertainty on system control and improve the safety and stability of the distributed electric drive vehicle. The specific steps are as follows:
[0006] Step 1): Establish vehicle state space equation;
[0007] Step 2): Calculate the desired center of mass sideslip angle and the desired yaw rate according to the vehicle longitudinal speed and the front wheel turning angle;
[0008] Step 3): Construct a new sliding mode function in which the sideslip angle and yaw rate of the center of mass converge to the expected value at the same time, and calculate the sliding mode expected yaw moment;
[0009] Step 4): adaptively compensating the sliding mode desired yaw moment to obtain the desired yaw moment by calculation;
[0010] Step 5): for the yaw moment, the desired torque of the four wheels is solved by using an optimal allocation method considering tire utilization;
[0011] Step 6): In order to solve the problem of wheel slippage when the vehicle starts, the desired wheel torque is corrected.
[0012] Furthermore, the vehicle state space equation described in step 1) is specifically:
[0013] S21: The vehicle state space equation is as follows:
[0014]
[0015] Among them, x1 = β; x2 = ω rc ; κ is the interference term caused by parameter uncertainty, and its upper bound is Δm is the unknown additional load mass; C f Indicates the cornering stiffness of the front tire; C r represents the cornering stiffness of the rear tire; β represents the sideslip angle of the center of mass; ω rc represents the yaw rate; M represents the yaw moment around the center of mass; m represents the vehicle mass; δ f Indicates the front wheel turning angle; I z represents the yaw moment of inertia; a and b represent the distance from the center of mass of the vehicle to the front and rear axles respectively; v x Indicates the longitudinal velocity of the vehicle.
[0016] Furthermore, in step 2), the desired center of mass sideslip angle and the desired yaw rate are specifically:
[0017] S31: The expected center of mass sideslip angle is calculated as follows:
[0018]
[0019] The meanings of the parameters in the formula are consistent with those in formulas S21 and S22. d represents the expected sideslip angle of the center of mass; K represents the stability factor; μ represents the wheelbase; μ represents the road adhesion coefficient; g represents the acceleration of gravity;
[0020] S32: The expected yaw rate is calculated as follows:
[0021]
[0022] The meanings of the parameters in the formula are consistent with those in the S21 and S31 formulas. rcd represents the desired yaw rate.
[0023] Furthermore, in step 3), the sliding mode expected yaw moment is specifically:
[0024] S41: The center of mass sideslip angle tracking error is calculated as follows:
[0025] e1=x1-x 1d
[0026] Among them, e1 represents the center of mass sideslip angle tracking error, x1 represents the vehicle center of mass sideslip angle, and x 1d Represents the expected value of the system state x1, that is, the expected center of mass sideslip angle β d ;
[0027] S42: The second tracking error is calculated as follows:
[0028]
[0029] The parameters in the formula have the same meaning as those in formulas S22 and S41. x2 represents the virtual input value; x 2d represents the expectation of the virtual input quantity; ζ1 represents the sliding mode parameter, which is a positive constant;
[0030] S43: The sliding mode function is as follows:
[0031] s=ζ2e1+e2
[0032] The meanings of the parameters in the formula are consistent with those in formulas S41 and S42. s represents the sliding mode function; ζ2 represents the sliding mode parameter, which is a positive constant;
[0033] S44: The reaching law is as follows:
[0034]
[0035] The parameters in the formula have the same meaning as those in the S43 formula. k1 and k2 represent controller parameters and satisfy
[0036] S45: The expected yaw moment of sliding mode is calculated as follows:
[0037]
[0038] The meanings of the parameters in the formula are consistent with those in formulas S22, S44, S43, S42, and S41;
[0039] Further, in step 4), the adaptive dynamic synovial desired yaw moment is specifically:
[0040] S51: The neural network input is as follows:
[0041]
[0042] The meanings of the parameters in the formula are consistent with those in formulas S43 and S44;
[0043] S52: The output matrix of the hidden layer neurons is as follows:
[0044]
[0045] The parameters in the formula have the same meaning as those in the S51 formula. i represents the number of network inputs, j represents the number of nodes in the hidden layer, and b represents the number of nodes in the hidden layer. ij represents the neuron width value of the jth hidden node corresponding to the i-th input, c ij Indicates the central parameter value of the jth neuron corresponding to the i-th input;
[0046] S53: The estimated value of the interference term is as follows:
[0047]
[0048] The meanings of the parameters in the formula are consistent with those in formulas S51 and S52. j is the network weight matrix;
[0049] S54: The adaptive dynamic sliding membrane desired yaw moment M is calculated as follows:
[0050]
[0051] The parameters in the formula have the same meaning as those in formulas S22, S44, S43, S42, S41, and S53. k1 and k2 represent controller parameters and satisfy
[0052] S55: The update rate of the network weight matrix is as follows:
[0053]
[0054] The parameters in the formula have the same meanings as those in formulas S21, S52, and S43. η represents the update rate coefficient, which is a normal number.
[0055] S56: b ij and c ij The update rule for is calculated as follows:
[0056]
[0057] The parameters in the formula have the same meanings as those in formulas S21, S51, S52, S53, S43, and S44. γ∈(0,1) represents the learning rate. represents the momentum factor.
[0058] Further, in step 5), the desired torque of the four wheels is specifically:
[0059] S61: Use quadratic programming to optimize torque distribution, take minimum tire utilization as the objective function, and obtain the output torque of each wheel as:
[0060]
[0061] The parameters in the formula have the same meanings as those in formulas S21 and S54. r represents the rolling radius of the tire, d represents the wheel track, and T i Indicates the motor output torque, F zi Represents the vertical force of the tire, i=1,2,3,4.
[0062] Furthermore, the PID control method described in step 6) is used to correct the desired wheel torque, specifically:
[0063] S71: The target slip ratio is calculated as follows:
[0064] S goal =(S min +S max ) / 2
[0065] Among them, S goal Indicates the target slip rate; S min With S max Respectively represent the maximum and minimum values of the target slip rate;
[0066] S72: The slip ratio difference is calculated as follows:
[0067] λ i =S goal -S i-real ,i=1,2,3,4
[0068] The meanings of the parameters in the formula are consistent with those in the S71 formula. i_real represents the actual slip rate of the i-th wheel;
[0069] S73: The torque correction amount based on PID is calculated as follows:
[0070]
[0071] The meanings of the parameters in the formula are consistent with those in the S72 formula. P represents the proportional gain, k I represents the integral gain, k D represents the differential gain;
[0072] S74: The expected torques of the four wheels are calculated as follows:
[0073] T i d =T i +ΔT i ,i=1,2,3,4
[0074] The meanings of the parameters in the formula are consistent with those in the S73 formula.
[0075] Beneficial effects: The invention aims at the influence of load parameters on distributed electric drive vehicles, and proposes a distributed drive electric vehicle torque distribution control method. The vehicle based on the invention has better adaptability to the uncertainty of the whole vehicle mass, and has the vehicle posture maintenance ability and motion tracking ability. It can eliminate the influence of load uncertainty on system control and has better stability and tracking performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0076] Figure 1 is a control strategy block diagram of the present invention;
[0077] Figure 2 It is a schematic diagram of a two-degree-of-freedom vehicle dynamics model;
[0078] Figure 3 It is a schematic diagram of RBF neural network;
[0079] Figure 4 It is a schematic diagram of the motion trajectory of the uncontrolled vehicle in test scenario 1;
[0080] Figure 5 is a schematic diagram of the vehicle speed tracking response of the uncontrolled vehicle in test scenario 1;
[0081] Figure 6 It is a comparative schematic diagram of the vehicle center of mass side slip angle in test scenario 1;
[0082] Figure 7 It is a comparative schematic diagram of the vehicle yaw rate in test scenario 1;
[0083] Figure 8 It is a schematic diagram of the motion trajectory of the uncontrolled vehicle in the test field 2;
[0084] Fig. 9 is a schematic diagram of the vehicle speed tracking response of the uncontrolled vehicle in test scenario 2;
[0085] Fig.10 It is a schematic diagram of the comparison of the vehicle center of mass sideslip angle in test scenario 2;
[0086] Fig.11 It is a schematic diagram of the comparison of vehicle yaw angular velocity in test scenario 2. DETAILED DESCRIPTION
[0087] In order to be able to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other. In the following description, many specific details are set forth in order to fully understand the present invention, and the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the art without making creative work are within the scope of protection of the present invention. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.
[0088] See attached for details Figure 1 This embodiment provides a torque distribution control method for a distributed intelligent electric vehicle, which is characterized by: adopting a hierarchical structure: the upper layer is a yaw moment decision layer, and the lower layer is a torque distribution layer; the sliding mode expected yaw moment is calculated in the upper layer of the control system and adaptively compensated; the expected torque of the four wheels is solved in the lower layer of the control system and the wheel torque is corrected, so as to eliminate the influence of load uncertainty on system control and improve the safety and stability of the distributed electric drive vehicle. The specific steps are as follows:
[0089] Step 1): Establish vehicle state space equation;
[0090] Step 2): Calculate the desired center of mass sideslip angle and the desired yaw rate according to the vehicle longitudinal speed and the front wheel turning angle;
[0091] Step 3): Based on the backstepping method, a new sliding mode function is constructed in which the sideslip angle and yaw rate of the center of mass converge to the expected value at the same time, and the sliding mode expected yaw moment is calculated;
[0092] Step 4): adaptively compensating the sliding mode desired yaw moment based on the RBF neural network to calculate the desired yaw moment;
[0093] Step 5): for the yaw moment, the desired torque of the four wheels is solved by using an optimal allocation method considering tire utilization;
[0094] Step 6): In order to solve the problem of wheel slippage when the vehicle starts, a PID control method is used to correct the desired wheel torque.
[0095] The preferred solution of this embodiment is that the vehicle state space equation described in step 1) is specifically:
[0096] S21: The vehicle state space equation is as follows:
[0097]
[0098] Among them, x1 = β; x2 = ω rc ; κ is the interference term caused by parameter uncertainty, and its upper bound is Δm is the unknown additional load mass; C f Indicates the cornering stiffness of the front tire; C r represents the cornering stiffness of the rear tire; β represents the sideslip angle of the center of mass; ω rc represents the yaw rate; M represents the yaw moment around the center of mass; m represents the vehicle mass; δ f Indicates the front wheel turning angle; I z represents the yaw moment of inertia; a and b represent the distance from the center of mass of the vehicle to the front and rear axles respectively; v x Indicates the longitudinal velocity of the vehicle.
[0099] The preferred solution of this embodiment is that in step 2), the desired center of mass sideslip angle and the desired yaw rate are specifically:
[0100] S31: The expected center of mass sideslip angle is calculated as follows:
[0101]
[0102] The meanings of the parameters in the formula are consistent with those in formulas S21 and S22. d represents the expected sideslip angle of the center of mass; K represents the stability factor; μ represents the wheelbase; μ represents the road adhesion coefficient; g represents the acceleration of gravity;
[0103] S32: The expected yaw rate is calculated as follows:
[0104]
[0105] The meanings of the parameters in the formula are consistent with those in the S21 and S31 formulas. rcd represents the desired yaw rate.
[0106] The preferred solution of this embodiment is that in step 3), the sliding mode expected yaw moment is specifically:
[0107] S41: The center of mass sideslip angle tracking error is calculated as follows:
[0108] e1=x1-x 1d
[0109] Among them, e1 represents the center of mass sideslip angle tracking error, x1 represents the vehicle center of mass sideslip angle, and x1d Represents the expected value of the system state x1, that is, the expected center of mass sideslip angle β d ;
[0110] S42: The second tracking error is calculated as follows:
[0111]
[0112] The parameters in the formula have the same meaning as those in formulas S22 and S41. x2 represents the virtual input value; x 2d represents the expectation of the virtual input quantity; ζ1 represents the sliding mode parameter, which is a positive constant;
[0113] S43: The sliding mode function is as follows:
[0114] s=ζ2e1+e2
[0115] The meanings of the parameters in the formula are consistent with those in formulas S41 and S42. s represents the sliding mode function; ζ2 represents the sliding mode parameter, which is a positive constant;
[0116] S44: The reaching law is as follows:
[0117]
[0118] The parameters in the formula have the same meaning as those in the S43 formula. k1 and k2 represent controller parameters and satisfy
[0119] S45: The expected yaw moment of sliding mode is calculated as follows:
[0120]
[0121] The meanings of the parameters in the formula are consistent with those in formulas S22, S44, S43, S42, and S41;
[0122] The preferred solution of this embodiment is that in step 4), the desired yaw moment of the adaptive dynamic sliding membrane is specifically:
[0123] S51: The neural network input is as follows:
[0124]
[0125] The meanings of the parameters in the formula are consistent with those in formulas S43 and S44;
[0126] S52: The output matrix of the hidden layer neurons is as follows:
[0127]
[0128] The parameters in the formula have the same meaning as those in the S51 formula. i represents the number of network inputs, j represents the number of nodes in the hidden layer, and b represents the number of nodes in the hidden layer. ij represents the neuron width value of the jth hidden node corresponding to the i-th input, c ij Indicates the central parameter value of the jth neuron corresponding to the i-th input;
[0129] S53: The estimated value of the interference term is as follows:
[0130]
[0131] The meanings of the parameters in the formula are consistent with those in formulas S51 and S52. j is the network weight matrix;
[0132] S54: The adaptive dynamic sliding membrane desired yaw moment M is calculated as follows:
[0133]
[0134] The parameters in the formula have the same meaning as those in formulas S22, S44, S43, S42, S41, and S53. k1 and k2 represent controller parameters and satisfy
[0135] S55: The update rate of the network weight matrix is as follows:
[0136]
[0137] The parameters in the formula have the same meanings as those in formulas S21, S52, and S43. η represents the update rate coefficient, which is a normal number.
[0138] S56: b ij and c ij The update rule for is calculated as follows:
[0139]
[0140] The parameters in the formula have the same meanings as those in formulas S21, S51, S52, S53, S43, and S44. γ∈(0,1) represents the learning rate. represents the momentum factor.
[0141] The preferred solution of this embodiment is that in step 5), the expected torques of the four wheels are specifically:
[0142] S61: Use quadratic programming to optimize torque distribution, take minimum tire utilization as the objective function, and obtain the output torque of each wheel as:
[0143]
[0144] The parameters in the formula have the same meanings as those in formulas S21 and S54. r represents the rolling radius of the tire, d represents the wheel track, and T i Indicates the motor output torque, F zi Represents the vertical force of the tire, i=1,2,3,4.
[0145] The preferred solution of this embodiment is that the PID control method described in step 6) is used to correct the desired wheel torque, specifically:
[0146] S71: The target slip ratio is calculated as follows:
[0147] S goal =(S min +S max ) / 2
[0148] Among them, S goal Indicates the target slip rate; S min With S max Respectively represent the maximum and minimum values of the target slip rate;
[0149] S72: The slip ratio difference is calculated as follows:
[0150] λ i =S goal -S i-real ,i=1,2,3,4
[0151] The meanings of the parameters in the formula are consistent with those in the S71 formula. i_real represents the actual slip rate of the i-th wheel;
[0152] S73: The torque correction amount based on PID is calculated as follows:
[0153]
[0154] The meanings of the parameters in the formula are consistent with those in the S72 formula. P represents the proportional gain, k I represents the integral gain, k D represents the differential gain;
[0155] S74: The expected torques of the four wheels are calculated as follows:
[0156] T i d =T i +ΔT i ,i=1,2,3,4
[0157] The meanings of the parameters in the formula are consistent with those in the S73 formula.
[0158] Example
[0159] The present invention designs a simulation experiment under a double lane-changing condition based on a Carsim / Simulink software joint simulation platform, and conducts comparative experimental verification based on the present invention, based on traditional first-order sliding film control, and without control.
[0160] The present invention adopts a hierarchical structure: the upper layer is the yaw moment decision layer, and the lower layer is the torque distribution layer. In the upper layer of the control system, a new sliding mode function is constructed based on the backstepping method, in which the center of mass sideslip angle and yaw angular velocity converge to the expected value at the same time, and the RBF neural network is used to adaptively compensate the unknown interference in the dynamic system to obtain the yaw moment output. In the lower layer of the control system, the optimal distribution method considering the tire utilization rate is used to solve the torque of the four wheels; in view of the problem that the wheels are prone to slip when the vehicle starts, the PID control method is used to correct the wheel torque.
[0161] 1. Vehicle model parameters and simulation scenarios
[0162] (1) The vehicle model parameters are as follows:
[0163]
[0164] (2) There are two experimental scenarios:
[0165] 1) The speed is 65 km / h, the road adhesion coefficient is 0.4, and the additional load mass is 0;
[0166] 2) Speed 65km / h, adhesion 0.4, increase vehicle mass by 200kg;
[0167] 2. Test results
[0168] (1) The test results of scenario 1 are as follows Figure 4-Figure 7 As shown. Figure 4 It can be seen that the uncontrolled vehicle has a large side slip, and the driving trajectory deviates greatly from the expected trajectory; the vehicle based on the traditional first-order sliding film control can basically ensure safe driving, but there is also a large overshoot at the turn; the vehicle based on the present invention has significantly improved the accuracy of tracking the expected trajectory, has a strong cornering posture correction ability, and basically does not have a large side slip. Figure 5 It can be seen that the speed tracking response of the uncontrolled vehicle is slow and there is a certain fluctuation; the speed tracking response of the vehicle based on the traditional first-order sliding film control and the vehicle based on the present invention is fast, but the vehicle based on the traditional first-order sliding film control has a certain overshoot. Figure 6 and Figure 7The figure shows the comparison of the vehicle's center of mass sideslip angle and yaw rate under the three control modes. Compared with no control, the traditional first-order sliding film control and the actual center of mass sideslip angle and yaw rate of the vehicle based on the present invention can accurately and quickly track the expected values.
[0169] (2) The test results of scenario 2 are as follows Figure 8-11 As shown. Figure 8 It can be seen that the uncontrolled vehicle and the vehicle based on the traditional first-order sliding film control have a large side slip, and the vehicle's driving trajectory seriously deviates from the expected trajectory at the longitudinal coordinate of 280m; although the vehicle based on the present invention has a certain lateral displacement, it can still track the expected trajectory. Fig. 9 It can be seen that the vehicle speed tracking response of the uncontrolled and the traditional first-order sliding film control is lower than that of the scenario 1, and there is a certain amount of jitter; the vehicle speed tracking response of the present invention is significantly improved, and the jitter is also significantly improved. Fig.10 and Fig.11 It can be seen that the yaw rate and sideslip angle deviation of the uncontrolled vehicle are the largest, and the vehicle stability is poor. After adopting the traditional first-order sliding film control, the yaw rate and sideslip angle deviation of the center of mass are significantly reduced, but there is still a certain fluctuation; the yaw rate and sideslip angle deviation of the vehicle adopting the present invention are the smallest, and the change trend is relatively gentle. Therefore, the vehicle based on the present invention has better adaptability to the uncertainty of the vehicle mass, and has the vehicle posture maintenance ability and motion tracking ability.
[0170] The above-mentioned detailed description of a torque distribution control method for a distributed intelligent electric vehicle with reference to the embodiment is illustrative rather than restrictive, and several embodiments can be listed according to the limited scope. Therefore, changes and modifications without departing from the overall concept of the present invention should fall within the scope of protection of the present invention.
Claims
1. A torque distribution control method for a distributed intelligent electric vehicle, characterized in that: A hierarchical structure is adopted, with the upper layer being the yaw moment decision layer and the lower layer being the torque distribution layer; the sliding mode desired yaw moment is calculated in the upper layer of the control system and adaptively compensated; the desired torque of the four wheels is solved in the lower layer of the control system, which can eliminate the influence of load uncertainty on system control and improve the safety and stability of distributed electric drive vehicles. The specific steps are as follows: Step 1) Establishing vehicle state space equation; Step 2) calculating the desired center of mass sideslip angle and the desired yaw rate according to the vehicle longitudinal velocity and the front wheel turning angle; Step 3) constructing a sliding mode function in which the sideslip angle and yaw rate of the center of mass converge to the expected value at the same time, and calculating the sliding mode expected yaw moment; Step 4) adaptively compensating the sliding mode desired yaw moment to obtain the desired yaw moment by calculation; Step 5) for the yaw moment, using an optimization allocation method to solve the expected torque of the four wheels; Step 6) correcting the wheel expected torque in view of the problem that the wheel may slip when the vehicle starts; The vehicle state space equation in step 1) is specifically: S21: The vehicle state space equation is as follows: Among them, x1 = β; x2 = ω rc ; κ is the interference term caused by parameter uncertainty, and its upper bound is Δm is the unknown additional load mass; C f Indicates the cornering stiffness of the front tire; C r represents the cornering stiffness of the rear tire; β represents the sideslip angle of the center of mass; ω rc represents the yaw rate; m represents the vehicle mass; δ f Indicates the front wheel turning angle; I z represents the yaw moment of inertia; a and b represent the distance from the center of mass of the vehicle to the front and rear axles respectively; v x Indicates the longitudinal speed of the vehicle; The desired center of mass sideslip angle and the desired yaw rate in step 2) are specifically: S31: The expected center of mass sideslip angle is calculated as follows: The meanings of the parameters in the formula are consistent with those in the S21 formula, β d represents the expected sideslip angle of the center of mass; K represents the stability factor; L represents the wheelbase; μ represents the road adhesion coefficient; g represents the acceleration of gravity; S32: The expected yaw rate is calculated as follows: The meanings of the parameters in the formula are consistent with those in the S21 and S31 formulas. rcd represents the desired yaw rate; The expected yaw moment of the sliding mode in step 3) is: S41: The center of mass sideslip angle tracking error is calculated as follows: e1=x1-x 1d Among them, e1 represents the center of mass sideslip angle tracking error, x1 represents the vehicle center of mass sideslip angle, and x 1d Represents the expected value of the system state x1, that is, the expected center of mass sideslip angle β d ; S42: The second tracking error is calculated as follows: The meanings of the parameters in the formula are consistent with those in formulas S22 and S41. x2 represents the virtual input value; x 2d represents the expectation of the virtual input quantity; ζ1 represents the sliding mode parameter, which is a positive constant; S43: The sliding mode function is as follows: s=ζ2e1+e2 The meanings of the parameters in the formula are consistent with those in formulas S41 and S42. s represents the sliding mode function; ζ2 represents the sliding mode parameter, which is a positive constant; S44: The reaching law is as follows: The parameters in the formula have the same meaning as those in the S43 formula. k1 and k2 represent controller parameters and satisfy S45: The expected yaw moment of sliding mode is calculated as follows: The meanings of the parameters in the formula are consistent with those in formulas S22, S44, S43, S42, and S41; The desired yaw moment of the adaptive dynamic sliding membrane in step 4) is specifically: S51: The neural network input is as follows: The meanings of the parameters in the formula are consistent with those in formulas S43 and S44; S52: The output matrix of the hidden layer neurons is as follows: The parameters in the formula have the same meaning as those in the S51 formula. i represents the number of network inputs, j represents the number of nodes in the hidden layer, and b represents the number of nodes in the hidden layer. ij represents the neuron width value of the jth hidden node corresponding to the i-th input, c ij Indicates the central parameter value of the jth neuron corresponding to the i-th input; S53: The estimated value of the interference term is as follows: The meanings of the parameters in the formula are consistent with those in formulas S51 and S52. j is the network weight matrix; S54: The adaptive dynamic sliding membrane desired yaw moment M is calculated as follows: The parameters in the formula have the same meaning as those in formulas S22, S44, S43, S42, S41, and S53. k1 and k2 represent controller parameters and satisfy S55: The update rate of the network weight matrix is as follows: The parameters in the formula have the same meanings as those in formulas S21, S52, and S43. η represents the update rate coefficient, which is a normal number. S56: b ij and c ij The update rule for is calculated as follows: The parameters in the formula have the same meanings as those in formulas S21, S51, S52, S53, S43, and S44. γ∈(0,1) represents the learning rate. represents the momentum factor; The expected torque of the four wheels in step 5) is specifically: S61: Use quadratic programming to optimize torque distribution, take minimum tire utilization as the objective function, and obtain the output torque of each wheel as: The parameters in the formula have the same meanings as those in formulas S21 and S54. r represents the rolling radius of the tire, d represents the wheel track, and T i Indicates the motor output torque, F zi Indicates the vertical force of the tire, i=1,2,3,4; Step 6) Use the PID control method to correct the desired wheel torque, specifically: S71: The target slip ratio is calculated as follows: S goal =(S min +S max ) / 2 Among them, S goal Indicates the target slip rate; S min With S max Respectively represent the maximum and minimum values of the target slip rate; S72: The slip ratio difference is calculated as follows: λ i =S goal -S i_real ,i=1,2,3,4 The meanings of the parameters in the formula are consistent with those in the S71 formula. i_real represents the actual slip rate of the i-th wheel; S73: The torque correction amount based on PID is calculated as follows: The meanings of the parameters in the formula are consistent with those in the S72 formula. P represents the proportional gain, k I represents the integral gain, k D represents the differential gain; S74: The expected torques of the four wheels are calculated as follows: T i d =T i +ΔT i ,i=1,2,3,4 The meanings of the parameters in the formula are consistent with those in the S73 formula.
