A path tracking and vehicle body attitude cooperative control method

Abstract

The present invention discloses a cooperative control method for path tracking and vehicle body attitude, comprising the following steps: Step 1, based on the road information obtained by sensors, a desired path is planned in real time; Step 2, based on the vehicle inverse dynamics model, feedforward control is performed on the vehicle to preferentially meet two control objectives that have a greater impact on driving safety, namely path tracking accuracy and rollover resistance performance; Step 3, a comprehensive evaluation module is set up to evaluate the feedforward control effect of the previous control cycle to obtain the performance allocation weights of each item; Step 4, within a control cycle, the closed-loop control from the sensors in the feedforward control module to the feedback adjustment module is continuously repeated until the cycle ends and enters the next cycle from the feedforward control module. Beneficial effects: By using the inverse dynamics model, the disadvantages of the forward dynamics model, such as high order, difficult determination of system parameters, and strong nonlinearity of the chassis execution system, are overcome, the system response speed and initial output accuracy are improved, and the chattering of the output quantity is reduced.

Description

Technical Field

[0001] The present invention relates to a path tracking control method, and particularly provides a collaborative control method for path tracking and vehicle body attitude. Background Technique

[0002] In recent years, with the rapid development of artificial intelligence, chips, and sensors, intelligentization and networking have become research hotspots in the automotive industry, injecting powerful impetus into the continuous upgrading of the automotive industry. While intelligent vehicles have many advantages such as predictable driving behavior, reduced traffic accident rates, and improved commuting efficiency, they also pose challenges to the stability and development of ethics and laws and regulations. Therefore, the public's requirements for the safety of intelligent driving vehicles are significantly higher than those of traditional vehicles.

[0003] At the level of motion control, the driving safety of intelligent vehicles mainly includes the following two aspects: accurately driving along the desired path to avoid collisions and exceeding the lane lines; not rolling over or skidding when changing the driving direction, that is, maintaining the vehicle body attitude and handling stability within a certain range.

[0004] At present, the motion control of most intelligent vehicles ensures path tracking accuracy without considering vehicle body attitude control and handling stability, that is, only meeting the first point of driving safety. However, in the actual path tracking process, there is a mutual coupling and restriction between the vehicle body attitude and path tracking accuracy. While maintaining and improving the path tracking accuracy, it may be accompanied by a sharp deterioration of the vehicle body attitude and handling stability, increasing the probability of rollover and skidding.

[0005] Regarding the above problems, existing research uses modern control methods to continuously correct control quantities such as vehicle speed, steering angle, and vehicle body attitude based on information such as vehicle orientation and attitude obtained by sensors, ensuring path tracking accuracy and anti-roll safety to a certain extent. However, due to the characteristics of feedback control itself, these methods are difficult to solve problems such as large errors at the initial moment of the control cycle and output chattering. At the same time, since the path tracking accuracy control and vehicle body attitude control cannot be decoupled, it is extremely easy to cause the simultaneous deterioration of accuracy and vehicle body attitude. Summary of the Invention

[0006] Object of the Invention: Aiming at the disadvantages of large initial errors and system chattering in existing control methods, the present invention provides a collaborative control method for path tracking and vehicle body attitude based on the vehicle inverse dynamics model, constructs and uses an inverse dynamics model that can accurately reflect the lateral-vertical motion coupling characteristics of autonomous vehicles, overcomes the disadvantages of high-order forward dynamics models, difficult determination of system parameters, and strong nonlinearity of the chassis execution system, highly fits the control process of autonomous vehicles, improves the system response speed and initial output accuracy, and reduces the chattering of the output quantity.

[0007] Technical Solution:

[0008] A path tracking and vehicle body attitude cooperative control method, comprising the following steps:

[0009] Step 1, turn on the path tracking and vehicle body attitude cooperative controller, and plan an expected path in real time based on the road information obtained by the sensor;

[0010] Step 2, the on-vehicle computer performs feedforward control on the vehicle based on the driving demand and road conditions, based on the vehicle inverse dynamics model, and gives priority to meeting two control objectives that have a greater impact on driving safety, namely path tracking accuracy and rollover resistance performance;

[0011] Step 3, set a comprehensive evaluation module to evaluate the feedforward control effect of the previous control cycle. First, use the entropy weight fusion method to obtain the weights of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance in real time. The method is as follows:

[0012] Process the sensor data to obtain a quantitative evaluation of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance and perform normalization processing;

[0013] Based on the sensor data corresponding to each index, including: the path tracking accuracy corresponds to the lateral displacement deviation y e threshold, the lateral orientation deviation e threshold, and the yaw rate threshold w e , obtain the entropy value of this index, and then obtain the weight values of each index;

[0014] According to the driver's choice of driving performance preference, multiply the weights of the path tracking accuracy and vibration comfort indexes by coefficients k1 and k2 to obtain a comprehensive evaluation index H, where k1 + k2 = 2. If the driving performance preference is for sports, then k1 takes 2. If it is for comfort, then k1 = 0. In other cases, it is linearly adjusted between 0 and 2; if the comprehensive index H after entropy weight fusion is less than 0.8, then allow the feedforward control of the next cycle to enter; if H is greater than 0.8 and less than 1, then pause the feedforward control and only retain the feedback control module; if H is greater than 1, then remind the driver to intervene;

[0015] Step 4, within one control cycle, use the slowest working step size of the sensor in the feedforward control module as the negative feedback control step size, and continuously repeat the closed-loop control from the sensor in the feedforward control module to the feedback adjustment module until the end of the cycle and enter the next cycle from the feedforward control module;

[0016] If the driving state parameter exceeds the warning value, give a warning and prompt the driver to intervene, and directly jump out of this control cycle and enter the next cycle.

[0017] Further, the specific operation process of the said step 2 is as follows:

[0018] According to the driver's choice of driving style, determine the bias of performance indicators such as path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance during driving;

[0019] The on-vehicle computer then determines the performance requirements based on the road and traffic flow density information obtained by the sensors in combination with the driving style, as well as the allocation weights of the control parameters of the path tracking system and the vehicle body attitude system under the current working conditions;

[0020] Taking path tracking accuracy and rollover resistance performance as constraint conditions, based on the above weight allocation for each performance, solve for the lateral displacement deviation y e threshold, lateral azimuth deviation e threshold, and yaw angular velocity threshold w e ;

[0021] By filtering out noise, fusing multiple frames, and clustering the main features of the point cloud map obtained by the lidar and the images obtained by the on-vehicle camera, obtain the current displacement deviation and road curvature. Input the current displacement deviation, the road curvature x meters ahead, the lateral displacement deviation ye threshold, the lateral azimuth deviation e threshold, and the yaw angular velocity threshold into the lateral inverse dynamics model to obtain the vehicle speed v x and the preview distance x e ; where the value of x meters is the numerical value of the average vehicle speed in the past 2s multiplied by 5;

[0022] Measure the roll angle and roll angular velocity in real time through the on-vehicle triaxial accelerometer, obtain the roll threshold TTR value through dynamic balance analysis and input it into the vertical inverse dynamics model to obtain the vehicle speed v x and the preview distance x e ;

[0023] When the vehicle speeds v x and the preview distances x e that meet the path tracking accuracy and rollover resistance performance have an intersection, select the vehicle speed v x and the preview distance x e according to the selected driving style of the driver, and control the throttle opening, gear, and brakes through the longitudinal inverse dynamics model to control the vehicle speed;

[0024] If the vehicle speeds v x and the preview distances x e that meet the path tracking accuracy requirements and rollover resistance performance have no intersection, issue a warning to the driver and directly enter the feedback regulation system.

[0025] Furthermore, the specific control logic of step 3 is as follows:

[0026] When the path tracking accuracy, handling stability, and comfort all meet the requirements, perform path tracking in the form of feedforward + feedback;

[0027] When the path tracking accuracy is lower than the threshold, but the handling stability and vehicle body attitude both meet the requirements, reduce the preview distance for tracking accuracy compensation;

[0028] When the path tracking accuracy does not meet the requirements, there are lane-changing conditions, and there is no risk of rollover, perform attitude compensation for handling stability and comfort;

[0029] When the path tracking accuracy does not meet the requirements, there are lane-changing conditions but there is a risk of rollover during lane change, decelerate and correct the vehicle and perform attitude compensation;

[0030] When the tracking accuracy does not meet the requirements and there are no lane-changing conditions, perform tracking accuracy compensation on the vehicle, and perform attitude compensation if the handling stability deteriorates;

[0031] When neither the vehicle body attitude compensation nor the path tracking accuracy compensation can achieve the generalized control target, give an early warning, brake and decelerate, and prompt the driver to intervene.

[0032] Furthermore, with path tracking accuracy, handling stability, and roll safety as constraint conditions, based on road conditions, driving states, and performance requirement biases, calculate the roll time limit TTR value, lateral displacement deviation y e threshold, and lateral orientation deviation e threshold. Substitute the roll time limit TTR value into the vertical inverse dynamics model, and substitute the lateral displacement deviation y e threshold and lateral orientation deviation e threshold into the lateral inverse dynamics model, and solve the vehicle speed v x and preview distance x e ;

[0033] The vehicle speed v x and preview distance x e solved by the two inverse dynamics models respectively form closed planes in the two-dimensional plane. Select appropriate vehicle speed v x and preview distance x e in the overlapping area of the two closed planes according to the real-time performance requirement bias and output them to the actuator.

[0034] Furthermore, with road parameters, driving state parameters, and cooperative control constraint conditions as inputs, specifically including: the curvature of the road ahead, the current vehicle azimuth deviation y e , displacement deviation e, yaw rate, maximum displacement deviation threshold, maximum azimuth deviation threshold, the output is the vehicle speed v x and preview distance x e ;

[0035] The input of the vertical inverse dynamics model is the vehicle body roll angular velocity and the roll TTR time limit threshold, and the output is the vehicle speed v x and preview distance xe 。

[0036] Further, the method for obtaining the inverse dynamics model is as follows: According to the motion differential equations of the two-degree-of-freedom vehicle open-loop system model and the preview tracking model, the forward dynamics model is solved to obtain the input-output data set representing the path tracking accuracy under different road curvatures, different working conditions, and different preview distances, and the explicit expression of the input-output of the path tracking system is obtained;

[0037] On this basis, the generalized regression neural network is applied to the lateral inverse dynamics modeling of autonomous vehicles. With the vehicle driving normally, that is, without colliding with the road boundary and without rolling over as the constraint conditions, the input-output set is selected, and the vehicle roll angle, roll angular velocity, lateral displacement deviation y e and lateral orientation deviation e are extracted to construct the multi-input multi-output data set of the normal path tracking system of autonomous vehicles. Matching design is carried out for the weighted summation of the neuron data in the summation layer of the generalized regression neural network, and finally an inverse dynamics model that accurately reflects the lateral dynamics and preview tracking characteristics of the vehicle is obtained.

[0038] Further, the input-output set is obtained based on the vehicle two-degree-of-freedom model and the preview tracking model, and is realized through the following steps:

[0039] Based on the two-degree-of-freedom vehicle dynamics equation, a two-degree-of-freedom vehicle dynamics model is established;

[0040] According to the relationship between the side slip angle of the tire, the sideslip angle of the vehicle center of mass, the yaw angular velocity, and the distance parameters of the center of mass from the front and rear axles, combined with Newton's second law, the motion differential equation of the linear two-degree-of-freedom model is obtained as:

[0041]

[0042] where m is the vehicle mass, C f and C r are the linear side slip stiffnesses of the front and rear wheels, l f and l r are the distances of the vehicle center of mass from the front and rear axles, I Z is the moment of inertia of the whole vehicle about the Z axis, δ f is the front wheel steering angle, V x is the longitudinal vehicle speed;

[0043] Design of the preview error system; The inputs of the preview error model include the sideslip angle of the vehicle center of mass and the yaw angular velocity output by the vehicle dynamics model, as well as the external input road curvature ρ and preview distance x e , and the output of the preview error model is the lateral displacement deviation ye and the lateral orientation deviation e. The formula of the preview error model of the vehicle is constructed:

[0044]

[0045] In the formula, y is the distance deviation between the vehicle preview point and the center line, v y is the vehicle lateral displacement speed, v x is the longitudinal vehicle speed, and r is the turning radius;

[0046] Based on the data within the past one feedforward control cycle, the entropy weight fusion method is used to assign weights to the lateral displacement deviation ye and the lateral orientation deviation e, and then they are used as the comprehensive deviation E;

[0047] The obtained comprehensive deviation E is input into the PID controller, and the comprehensive deviations of at least the past two feedforward control cycles are integrated in the time domain. Based on the integration result, the PID control method is used to output the front wheel angle δ of the supplementary vehicle f as the control quantity of the vehicle dynamics model, thereby forming a closed-loop control system for path-tracking lateral motion.

[0048] Furthermore, it also includes a driving state monitoring module, which uses the optimal individual to optimize the weights to obtain the optimized weights;

[0049] The optimized initial weights and thresholds are substituted into the neural network. By performing selection, crossover, and mutation operations, the individual corresponding to the optimal fitness value is found, the network error is calculated, and the weights and thresholds are modified according to the calculation result, ultimately achieving the required accuracy;

[0050] Among them, the measured values of the triaxial acceleration and the environmental variables are used as the input nodes. The environmental variables depend on the situation and the specific values are determined after the network training. The vehicle body pitch angle and roll angle are used as the output nodes. The weights and thresholds are changed through network training until the accuracy reaches the standard.

[0051] Furthermore, the relative quantity of the ratio of the current lateral acceleration of the vehicle to the real-time lateral limit acceleration is used as the rollover determination condition, replacing the traditional method of predicting rollover by relying on the roll angle or the absolute quantity of the lateral acceleration;

[0052] The acceleration of the vehicle in the rollover critical state is:

[0053]

[0054] In the formula, a y is the lateral acceleration at the center of gravity position, a y,L is the critical lateral acceleration at the center of gravity position for rollover, m is the total vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the wheelbase, h is the distance from the center of gravity to the roll center, is the roll angle

[0055] L d = a y / ay,L

[0056]

[0057] Wherein, a y is the lateral acceleration at the center of gravity position, a y,L is the critical lateral acceleration for rollover at the center of gravity position, m is the total vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the track width, h is the distance from the center of gravity to the roll center, is the roll angle

[0058] Based on the L d value, calculate the time to rollover limit, i.e., the TTR value, and the calculation step size is 10 ms;

[0059] Determine the TTR threshold for vehicle body attitude compensation according to the performance requirement bias,

[0060] If the collaborative control performance requirement biases towards tracking accuracy, the TTR threshold takes a larger value to make the center of gravity of the whole vehicle lower when the vehicle is turning;

[0061] If the collaborative control requirement biases towards comfort, the TTR threshold takes a smaller value, and the TTR threshold is proportional to the vehicle speed.

[0062] Furthermore, when the path tracking accuracy is lower than the set threshold, but the measurement values of handling stability and comfort do not meet the set threshold, the handling stability and comfort of the vehicle are given priority, and the vehicle body attitude compensation is performed on the vehicle. The vehicle body attitude compensation control method is realized through the following steps:

[0063] Improve the vehicle body attitude by controlling the vehicle roll. The roll moment generated by the roll consists of three parts: 1. The roll moment M ΦrΙ caused by the centrifugal force of the sprung mass;

[0064] 2. The anti-roll moment M ΦrⅡ caused by the gravity of the sprung mass;

[0065] 3. The roll moment M ΦrⅢ caused by the centrifugal force of the unsprung mass;

[0066] In addition, when rolling, the vertical load is transferred between the left and right wheel loads, generating load transfer moments M ZF , M ZR ; when the vehicle body is in a roll state, additional forces Δf with opposite directions are applied by the active suspension actuators on the left and right sides of the suspension in the current state, which can form an anti-roll moment M af , which can suppress the roll of the vehicle;

[0067] The roll moment M ΦrⅠis:

[0068] M ΦrⅠ = m s ·a y ·h

[0069] M ΦrⅠ is the roll moment caused by the centrifugal force of the unsprung mass, m s is the unsprung mass, a y is the acceleration of the vehicle body centroid, and h is the centroid height;

[0070] The roll moment M caused by the gravity of the unsprung mass ΦrⅡ is:

[0071]

[0072] M ΦrⅡ is the roll moment caused by the centrifugal force of the unsprung mass, m s is the unsprung mass, e is the lateral distance from the centroid to the wheel, a y is the acceleration of the vehicle body centroid, h g is the height from the connection between the suspension on the lower side of the vehicle body and the vehicle body to the centroid, is the roll angle of the vehicle body;

[0073] The roll moment M caused by the centrifugal force of the unsprung mass ΦrⅢ is:

[0074] M ΦrⅢ = -Fuy(h0 - r)

[0075] M ΦrⅢ is the roll moment caused by the unsprung mass, F is the static friction force of the road surface on the wheel pointing to the steering center, u is the vehicle speed, y is the lateral dimension of the vehicle body, h0 is the centroid height, and r is the wheel radius;

[0076] During rolling, the vertical load causes a load transfer between the left and right wheels, generating a load transfer moment M ZF 、M ZR is:

[0077] M ZF =(F rRF - F rLF )·B / 2

[0078] M ZR =(F rRR - F rLR )·B / 2

[0079] M ZF is the front wheel transfer moment, M ZR is the rear wheel transfer moment, F rRF is the right front wheel transfer force, FrLF is the transfer force of the left rear wheel, F rRR is the transfer force of the right rear wheel, F rRR , B is the vehicle body width;

[0080] Taking moments about the longitudinal center line of the vehicle body, that is:

[0081] M ΦrⅠ -M ΦrⅡ +M ΦrⅢ +M ZF +M ZR =M af

[0082] M ΦrⅠ is the roll moment caused by the centrifugal force of the unsprung mass, M ΦrⅡ is the roll moment caused by the centrifugal force of the unsprung mass, M ΦrⅢ is the roll moment caused by the unsprung mass, M ZF is the transfer moment of the front wheels, M ZR is the transfer moment of the rear wheels, M af is the anti-roll moment provided by the actuator.

[0083] Beneficial effects: The present invention provides a method for coordinated control of path tracking and vehicle body attitude, constructs and uses an inverse dynamics model that can accurately reflect the lateral-vertical motion coupling characteristics of an autonomous vehicle, overcomes the disadvantages of high order of the forward dynamics model, difficult determination of system parameters, and strong nonlinearity of the chassis actuator system, highly conforms to the control process of an autonomous vehicle, improves the system response speed and initial output accuracy, and reduces the chattering of the output. Description of the Drawings

[0084] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0085] Figure 1 is the coordinated control logic route diagram of vehicle path tracking and vehicle body attitude of the present invention.

[0086] Figure 2 is the road driving scenario diagram of the intelligent vehicle of the present invention.

[0087] Figure 3 is the construction flow chart of the vehicle dynamics inverse model of the present invention.

[0088] Figure 4 is the two-degree-of-freedom vehicle dynamics model diagram of the present invention.

[0089] Figure 5 This is the preview error model diagram of the present invention.

[0090] Figure 6 This is the schematic diagram of the generation principle of the roll moment of the present invention.

[0091] Figure 7 This is the effect diagram of the vehicle body attitude compensation of the present invention.

[0092] Figure 8 This is the effect diagram of the path tracking compensation of the present invention.

[0093] Figure 9 This is the effect diagram of the path tracking compensation of the present invention.

[0094] Figure 10 This is the effect diagram of the path tracking compensation of the present invention. Detailed implementation manners

[0095] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, 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 those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0096] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0097] In the present invention, unless otherwise clearly defined and limited, the first feature being "on" or "under" the second feature may include direct contact between the first and second features, or may include indirect contact between the first and second features through additional features therebetween. Moreover, the first feature being "above", "over" and "on top of" the second feature includes the first feature being directly above and obliquely above the second feature, or simply indicating that the horizontal height of the first feature is higher than that of the second feature. The first feature being "under", "below" and "beneath" the second feature includes the first feature being directly below and obliquely below the second feature, or simply indicating that the horizontal height of the first feature is lower than that of the second feature.

[0098] As Figure 1 and 2 shown, a path tracking and vehicle body attitude collaborative control method includes the following steps:

[0099] Step 1: Turn on the path tracking and vehicle body attitude collaborative controller, and based on the road information obtained by the sensor, plan the desired path in real time;

[0100] Step 2: The on-vehicle computer performs feedforward control on the vehicle based on the driving requirements and road conditions and the vehicle inverse dynamics model, giving priority to meeting the two control objectives of path tracking accuracy and rollover resistance performance that have a greater impact on driving safety; the specific operation process is as follows:

[0101] According to the driver's choice of driving style, determine the bias of performance indicators such as path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance during driving;

[0102] The on-vehicle computer then determines the performance requirements based on the road and traffic flow density conditions obtained by the sensor and the driving style, makes normalized evaluation indicators for path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance, and determines the allocation weights of the control parameters of the path tracking system and the vehicle body attitude system under the current working conditions;

[0103] Taking path tracking accuracy and rollover resistance performance as constraint conditions, based on the above weight allocation for each performance, solve for the lateral displacement deviation y e threshold, lateral azimuth deviation e threshold, yaw angular velocity threshold w e ;

[0104] By filtering out noise, multi-frame fusion, and clustering of main features from the point cloud map obtained by the lidar and the images obtained by the on-vehicle camera, obtain the current displacement deviation and road curvature. The current displacement deviation, the road curvature x meters ahead, and the lateral displacement deviation ye threshold, lateral azimuth deviation e threshold, and yaw angular velocity threshold are input into the lateral inverse dynamics model to obtain the vehicle speed v that meets the tracking accuracy requirements x and preview distance x e ; where the value of x meters is the numerical value of the average vehicle speed in the past 2s multiplied by 5;

[0105] Integrate the data measured by the on-vehicle three-axis accelerometer in the time domain and frequency domain to obtain the roll angle and roll angle velocity. Through dynamic balance analysis, obtain the roll threshold TTR value and input it into the vertical inverse dynamics model to obtain the vehicle speed v that meets the rollover resistance performance x and preview distance x e ;

[0106] When the vehicle speeds v x and preview distances x e that meet path tracking accuracy and rollover resistance performance have an intersection, select the vehicle speed v x and preview distance x e, and control the throttle opening, gear, and brake through a longitudinal inverse dynamics model to control the vehicle speed;

[0107] If the vehicle speed v that meets the path tracking accuracy requirement and rollover resistance performance x and the preview distance x e have no intersection, a warning is issued to the driver and directly enter the feedback regulation system;

[0108] Step 3: Set up a comprehensive evaluation module to evaluate the feedforward control effect of the previous control cycle. First, use the entropy weight fusion method to obtain the respective weights of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance in real time. The method is as follows:

[0109] Take the processed sensor data every 0.01 s within 0.4 s of the previous feedforward control cycle, and perform normalization processing on path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance respectively;

[0110] Based on the sensor data corresponding to each index, including: the path tracking accuracy corresponding to the lateral displacement deviation y e threshold, the lateral azimuth deviation e threshold, and the yaw rate threshold w e , obtain the entropy value of this index, and then obtain the weight values of each index;

[0111] According to the driver's selection of driving performance preference, multiply the weights of the path tracking accuracy and vibration comfort indexes by coefficients k1 and k2 to obtain the comprehensive evaluation index H, where k1 + k2 = 2. If the driving performance preference is for sports, k1 takes 2; if it is for comfort, k1 = 0; in other cases, it is linearly adjusted between 0 and 2. If the comprehensive index H after entropy weight fusion is less than 0.8, the feedforward control of the next cycle is allowed; if H is greater than 0.8 and less than 1, the feedforward control is suspended and only the feedback control module is retained; if H is greater than 1, the driver is reminded to intervene. The specific control logic is as follows:

[0112] When the path tracking accuracy, handling stability, and comfort all meet the requirements, the path tracking is carried out normally with feedforward + feedback;

[0113] When the path tracking accuracy is lower than the threshold, but the handling stability and vehicle body attitude both meet the requirements, reduce the preview distance for tracking accuracy compensation;

[0114] When the path tracking accuracy does not meet the requirements, there are lane-changing conditions, and there is no rollover danger, perform attitude compensation on the handling stability and comfort;

[0115] When the path tracking accuracy does not meet the requirements, there are lane-changing conditions but there is a rollover danger during lane-changing, decelerate and correct the vehicle and perform attitude compensation;

[0116] When the tracking accuracy does not meet the requirements and there is no lane-changing condition, perform tracking accuracy compensation on the vehicle. If the handling stability deteriorates, perform attitude compensation;

[0117] When both the vehicle body attitude compensation and the path tracking accuracy compensation cannot achieve the generalized control objective, give an early warning, brake and decelerate, and prompt the driver to intervene;

[0118] Step 4: In a control cycle, use the slowest working step of the sensors in the feedforward control module as the negative feedback control step, and continuously repeat the closed-loop control from the sensors in the feedforward control module to the feedback adjustment module until the end of the cycle and enter the next cycle from the feedforward control module;

[0119] If the driving state parameters exceed the warning value, give an early warning, prompt the driver to intervene, and directly jump out of this control cycle and enter the next cycle.

[0120] At the beginning of each control cycle, the feedforward control module calculates the initial output and outputs it through the actuator;

[0121] Subsequently, enter the feedback control module. The triaxial accelerometer and lidar are used to obtain the vehicle pose information, azimuth information, and driving state parameters in real time, judge the vehicle driving state, and correspondingly select the control logic of the vehicle chassis actuator in the feedback control;

[0122] If the comprehensive evaluation index H of the feedforward control is greater than 1, that is, the control quantity that meets the requirements of coordinated control cannot be output at the initial moment of the control cycle, give an early warning to prompt the driver to pay attention to intervening at any time, and at the same time enter the feedback control.

[0123] The feedforward control only outputs the control quantity at the initial moment of each control cycle, and then enters the feedback control. The feedback control step is 10 ms, and a control cycle is 400 ms.

[0124] Taking the path tracking accuracy, handling stability, and roll safety as constraint conditions, calculate the roll time limit TTR value, lateral displacement deviation y e threshold, and lateral azimuth deviation e threshold based on the road conditions, driving state, and performance requirement bias. Substitute the roll time limit TTR value into the vertical inverse dynamics model, and substitute the lateral displacement deviation y e threshold and lateral azimuth deviation e threshold into the lateral inverse dynamics model, and solve the vehicle speed v x and preview distance x e ;

[0125] The vehicle speed v x and preview distance x e solved by the two inverse dynamics models respectively form closed planes in the two-dimensional plane. Select the appropriate vehicle speed v x and preview distance xe Output to the actuator.

[0126] Taking road parameters, driving state parameters, and cooperative control constraint conditions as inputs, specifically including: the curvature of the road ahead, the azimuth deviation y of the current vehicle e , the displacement deviation e, the yaw rate, the maximum displacement deviation threshold, the maximum azimuth deviation threshold, and the output is the vehicle speed v that meets the requirements of path tracking accuracy and handling stability x and the preview distance x e ;

[0127] The input of the vertical inverse dynamics model is the body roll rate and the roll TTR time limit threshold, and the output is the vehicle speed v that meets the roll safety x and the preview distance x e .

[0128] According to the motion differential equations of the two-degree-of-freedom vehicle open-loop system model and the preview tracking model, the forward dynamics model is solved to obtain the input-output data sets representing path tracking accuracy under different road curvatures, different working conditions, and different preview distances, and the explicit expression of the input-output of the path tracking system is obtained;

[0129] On this basis, the generalized regression neural network (GRNN) is applied to the lateral inverse dynamics modeling of autonomous vehicles. Taking the normal driving of the vehicle, that is, not colliding with the road boundary and not rolling over, as the constraint condition, the input-output set is selected, and the vehicle roll angle, roll rate, lateral displacement deviation y e , and lateral azimuth deviation e are extracted to construct a multi-input multi-output data set for the normal path tracking system of autonomous vehicles. Matching design is carried out for the weighted summation of the neurons in the summation layer of GRNN, and finally an inverse dynamics model that accurately reflects the lateral dynamics and preview tracking characteristics of the vehicle is obtained.

[0130] Based on the vehicle two-degree-of-freedom model and the preview tracking model, the input-output set is obtained and realized through the following steps:

[0131] As Figure 3 shown, based on the two-degree-of-freedom vehicle dynamics equation, a two-degree-of-freedom vehicle dynamics model is established;

[0132] According to the relationship between the sideslip angle of the tire and the sideslip angle of the vehicle center of mass, the yaw rate, and the distance parameters of the center of mass from the front and rear axles, combined with Newton's second law, the motion differential equation of the linear two-degree-of-freedom model is obtained as:

[0133]

[0134] where m is the vehicle mass, C f 、C r are the linear sideslip stiffnesses of the front and rear wheels, l f 、lr The distance from the vehicle's center of mass to the front and rear axles, I Z The moment of inertia of the whole vehicle about the Z-axis, δ f The steering angle of the front wheels, V x The longitudinal vehicle speed;

[0135] Such as Figure 4 As shown, the design of the preview error system; the inputs of the preview error model include the sideslip angle and yaw rate of the output center of mass of the vehicle dynamics model, as well as the external inputs of road curvature ρ and preview distance x e , the output of the preview error model is the lateral displacement deviation ye and the lateral orientation deviation e, and the formula for constructing the vehicle's preview error model is:

[0136]

[0137] In the formula, y is the distance deviation between the vehicle preview point and the center line, v y is the vehicle's lateral translation speed, v x is the longitudinal vehicle speed, and r is the turning radius.

[0138] Based on the data within the past 0.4 s, the entropy weight fusion method is used to assign weights to the lateral displacement deviation ye and the lateral orientation deviation e, and then used as the comprehensive deviation E;

[0139] The obtained comprehensive deviation E is input into the PID controller, and the time-domain integration of the comprehensive deviation in the past 2 s is performed. Based on the integration result, the PID control method is used to output and supplement the steering angle δ of the vehicle's front wheels f As the control quantity of the vehicle dynamics model, a closed-loop control system for path-tracking lateral motion is formed in this way.

[0140] It also includes a driving state monitoring module,

[0141] The weights are optimized using the optimal individuals to obtain the optimized weights;

[0142] The optimized initial weights and thresholds are substituted into the neural network. By performing selection, crossover, and mutation operations, the individual corresponding to the optimal fitness value is found, the network error is calculated, and the weights and thresholds are modified according to the calculation results until the required accuracy is achieved;

[0143] Among them, the measured values of the three-axis acceleration and the environmental variables are used as input nodes. The environmental variables depend on the situation and the specific values are determined after the network training. The vehicle body pitch angle and roll angle are used as output nodes. The weights and thresholds are changed through network training until the accuracy reaches the standard.

[0144] The relative quantity of the ratio of the vehicle's current lateral acceleration to the real-time lateral limit acceleration is used as the rollover determination condition, replacing the traditional method of predicting rollover by relying on the roll angle or the absolute quantity of lateral acceleration;

[0145] The acceleration of the vehicle in the critical rollover state is:

[0146]

[0147] In the formula, a y is the lateral acceleration at the center of gravity position, a y,L is the critical lateral acceleration for rollover at the center of gravity position, m is the total vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the track width, h is the distance from the center of gravity to the roll center, is the roll angle

[0148] L d = a y / a y,L

[0149]

[0150] In the formula, a y is the lateral acceleration at the center of gravity position, a y,L is the critical lateral acceleration for rollover at the center of gravity position, m is the total vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the track width, h is the distance from the center of gravity to the roll center, is the roll angle

[0151] Based on the L d value, calculate the time to rollover (TTR) value, and the calculation step size is 10 ms;

[0152] According to the performance requirement bias, determine the TTR threshold for vehicle body attitude compensation,

[0153] If the collaborative control performance requirement biases towards tracking accuracy, the TTR threshold takes a larger value to make the center of gravity of the whole vehicle lower when the vehicle is turning;

[0154] If the collaborative control requirement biases towards comfort, the TTR threshold takes a smaller value, and the TTR threshold is proportional to the vehicle speed.

[0155] Among them, when the path tracking accuracy is lower than the set threshold, but the measurement values of handling stability and comfort do not meet the set threshold, the handling stability and comfort of the vehicle are given priority, and the vehicle body attitude compensation is performed on the vehicle. The vehicle body attitude compensation control method is realized through the following steps:

[0156] Improve the vehicle body attitude by controlling the vehicle roll. The roll moment generated by the roll consists of three parts: 1. The roll moment M ΦrΙ ;

[0157] 2. Roll resistance moment M caused by the gravity of the sprung mass ΦrⅡ ;

[0158] 3. Roll moment M caused by the centrifugal force of the unsprung mass ΦrⅢ ;

[0159] In addition, as shown in Figure 5 and 6 , when rolling, the vertical load causes a load transfer between the left and right wheel loads, generating a load transfer moment M ZF , M ZR ; when the vehicle body is in a rolling state, the active suspension actuators on the left and right sides apply additional forces Δf in opposite directions in the current state, which can form a roll resistance moment M af , which can suppress the roll of the vehicle;

[0160] Roll moment M caused by the centrifugal force of the sprung mass ΦrⅠ is:

[0161] M ΦrⅠ = m s ·a y ·h

[0162] M ΦrⅠ is the roll moment caused by the centrifugal force of the sprung mass, m s is the sprung mass, a y is the centroid acceleration of the vehicle body, h is the centroid height;

[0163] Roll resistance moment M caused by the gravity of the sprung mass ΦrⅡ is:

[0164]

[0165] M ΦrⅡ is the roll moment caused by the centrifugal force of the sprung mass, m s is the sprung mass, e is the lateral distance from the centroid to the wheel, a y is the centroid acceleration of the vehicle body, h g is the height from the connection between the lower side suspension of the vehicle body and the vehicle body to the centroid, is the roll angle of the vehicle body;

[0166] Roll moment M caused by the centrifugal force of the unsprung mass ΦrⅢ is:

[0167] M ΦrⅢ = -Fuy(h0 - r)

[0168] M ΦrⅢ is the roll moment caused by the unsprung mass, F is the static friction force of the road surface on the wheel pointing to the center of the turn, u is the vehicle speed, y is the lateral dimension of the vehicle body, h0 is the centroid height, and r is the wheel radius.

[0169] When the vehicle rolls, the vertical load causes load transfer between the left and right wheels, generating a load transfer moment M ZF 、M ZR is:

[0170] M ZF =(F rRF -F rLF )·B / 2

[0171] M ZR =(F rRR -F rLR )·B / 2

[0172] M ZF is the front wheel transfer moment, M ZR is the rear wheel transfer moment, F rRF is the right front wheel transfer force, F rLF is the left rear wheel transfer force, F rRR is the right rear wheel transfer force, F rRR , and B is the vehicle body width.

[0173] Taking moments about the longitudinal center line of the vehicle body, that is:

[0174] M ΦrⅠ -M ΦrⅡ +M ΦrⅢ +M ZF +M ZR =M af

[0175] M ΦrⅠ is the roll moment caused by the centrifugal force of the sprung mass, M ΦrⅡ is the roll moment caused by the centrifugal force of the sprung mass, M ΦrⅢ is the roll moment caused by the unsprung mass, M ZF is the front wheel transfer moment, M ZR is the rear wheel transfer moment, M af is the anti-roll moment provided by the actuator.

[0176] Using the skyhook control strategy to suppress the vertical vibration of the vehicle; the skyhook control force is:

[0177]

[0178] where f si is the skyhook control force; c sky is the skyhook damping, Z si is the absolute velocity of the vehicle body in the vertical direction;

[0179] Combining the skyhook control force f si with the control force compensation f fiAfter addition, the desired semi-active control force F is obtained ci , that is:

[0180] F ci = f si + f fi

[0181] F ci is the semi-active control force, f si is the skyhook control force, f fi is the control force compensation amount.

[0182] F ci and A zi The mathematical relationship is:

[0183]

[0184] In order to keep the flow area of the adjustable damping valve within a reasonable range, the maximum flow area A max = 0.00006 m 2 of the adjustable damping valve is set according to experience, and the minimum flow area A min = 0 m 2 .

[0185] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is the difference from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and reference can be made to the description of the method part for related parts.

[0186] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A path tracking and vehicle body attitude cooperative control method, characterized in that, It includes the following steps: Step 1: Turn on the path tracking and vehicle body attitude collaborative controller, and plan the desired path in real time based on the road information obtained by the sensor; Step 2: The on-vehicle computer performs feedforward control on the vehicle based on the driving requirements and road conditions, and based on the vehicle inverse dynamics model, giving priority to meeting the two control objectives of path tracking accuracy and rollover resistance performance, which have a greater impact on driving safety; Step 3: Set up a comprehensive evaluation module to evaluate the feedforward control effect of the previous control cycle. First, use the entropy weight fusion method to obtain the weights of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance in real time. The method is as follows: Process the sensor data to obtain quantitative evaluations of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance and perform normalization processing; Based on the sensor data corresponding to each indicator, including: path tracking accuracy corresponding to lateral displacement deviation y e Threshold, lateral orientation deviation e threshold, yaw angular velocity threshold w e , get the entropy value of the indicator, and then get the weight of each indicator; According to the driver's choice of driving performance preference, multiply the weights of the path tracking accuracy and vibration comfort indicators by coefficients k1 and k2 to obtain the comprehensive evaluation index H, where k1 + k2 = 2. If the driving performance preference is for sport, then k1 is taken as 2. If it is for comfort, then k1 = 0. In other cases, it is linearly adjusted between 0 and 2. If the comprehensive index H after entropy weight fusion is less than 0.8, then allow the feedforward control of the next cycle to enter; if H is greater than 0.8 and less than 1, then pause the feedforward control and only retain the feedback control module; if H is greater than 1, then remind the driver to intervene; Step 4: In one control cycle, use the slowest working step length of the sensor in the feedforward control module as the negative feedback control step length, and continuously repeat the closed-loop control from the sensor in the feedforward control module to the feedback adjustment module until the end of the cycle and enter the next cycle from the feedforward control module; If the driving state parameters exceed the warning value, give a warning and prompt the driver to intervene, and directly jump out of this control cycle and enter the next cycle.

2. The path tracking and vehicle body attitude cooperative control method according to claim 1, characterized in that: The specific operation process of the said Step 2 is as follows: According to the driver's choice of driving style, determine the bias of the performance indicators of path tracking accuracy, vehicle body attitude maintenance performance, vibration comfort, rollover resistance performance, and anti-skid performance during driving; The on-vehicle computer then determines the performance requirements based on the road and traffic flow density conditions obtained by the sensor in combination with the driving style, as well as the distribution weights of the control parameters of the path tracking system and the vehicle body attitude system under the current working conditions; Taking the path tracking accuracy and rollover resistance performance as constraint conditions, based on the weight distribution of various performances, solve the lateral displacement deviation y that characterizes the constraint conditions e threshold, lateral azimuth deviation e threshold, yaw rate threshold w e ; The current displacement deviation and road curvature are obtained by filtering noise, fusing multiple frames, and clustering main features from the point cloud map acquired by lidar and the image acquired by in-vehicle camera. The current displacement deviation, road curvature at x meters ahead, lateral displacement deviation ye threshold, lateral azimuth deviation e threshold, and yaw rate threshold are input into the lateral inverse dynamics model to obtain the vehicle speed v that meets the tracking accuracy requirements x and preview distance x e ; where the value of x meters is the numerical value of the average vehicle speed in the past 2s multiplied by 5; The roll angle and roll angular velocity are measured in real time by an in-vehicle triaxial accelerometer. The roll threshold TTR value is obtained through dynamic balance analysis and input into the vertical inverse dynamics model to obtain the vehicle speed v that meets the anti-rollover performance x and the preview distance x e ; When the vehicle speed v x and the preview distance x e have an intersection, the vehicle speed v x and the preview distance x e are selected according to the driving style selected by the driver, and the throttle opening, gear, and brake are controlled through the longitudinal inverse dynamics model to control the vehicle speed; If the vehicle speed v x that meets the path tracking accuracy requirements and rollover resistance performance e and the preview distance x have no intersection, a warning is issued to the driver and the feedback regulation system is directly entered.

3. The path tracking and vehicle body attitude cooperative control method according to claim 1, characterized in that: The specific control logic of the said Step 3 is as follows: When the path tracking accuracy, handling stability, and comfort all meet the requirements, perform path tracking with feedforward + feedback normally; When the path tracking accuracy is lower than the threshold, but the handling stability and vehicle body attitude both meet the requirements, reduce the preview distance for tracking accuracy compensation; When the path tracking accuracy does not meet the requirements and there are lane-changing conditions and no rollover danger, perform attitude compensation for handling stability and comfort; When the path tracking accuracy does not meet the requirements, there are lane-changing conditions but there is a rollover danger during lane-changing, decelerate and correct the vehicle and perform attitude compensation; When the tracking accuracy does not meet the requirements and there are no lane-changing conditions, perform tracking accuracy compensation on the vehicle. If the handling stability deteriorates, perform attitude compensation; When neither the vehicle body attitude compensation nor the path tracking accuracy compensation can reach the generalized control objective, give a warning, brake and decelerate, and prompt the driver to intervene.

4. The path tracking and vehicle body attitude cooperative control method according to claim 1, characterized in that: With the path tracking accuracy, handling stability, and roll safety as the constraint conditions, the roll time limit TTR value, lateral displacement deviation y e threshold, and lateral orientation deviation e threshold are calculated based on the road conditions, driving state, and performance requirement bias. The roll time limit TTR value is substituted into the vertical inverse dynamics model, and the lateral displacement deviation y e threshold and lateral orientation deviation e threshold are substituted into the lateral inverse dynamics model to solve the vehicle speed v x and preview distance x e ; The vehicle speed v solved by two inverse dynamics models x and the preview distance x e respectively form closed planes in a two-dimensional plane. According to real-time performance requirements, a suitable vehicle speed v x and preview distance x e are output to the actuator.

5. The path tracking and vehicle body attitude cooperative control method according to claim 4, characterized in that: Taking road parameters, driving state parameters, and cooperative control constraint conditions as inputs, specifically including: the curvature of the road ahead, the lateral deviation y of the current vehicle e , the displacement deviation e, the yaw rate, the maximum displacement deviation threshold, the maximum lateral deviation threshold, and the output is the vehicle speed v that meets the requirements of path tracking accuracy and handling stability x and the preview distance x e ; The input of the vertical inverse dynamics model is the body roll angular velocity and the roll TTR time limit threshold, and the output is the vehicle speed v that meets the roll safety x and the preview distance x e .

6. The path tracking and vehicle body attitude cooperative control method according to claim 4 or 5, characterized in that: The method for obtaining the inverse dynamics model is as follows: According to the motion differential equations of the two-degree-of-freedom vehicle open-loop system model and the preview tracking model, the forward dynamics model is solved to obtain the input-output data set representing the path tracking accuracy under different road curvatures, different working conditions, and different preview distances, and the explicit expression of the input-output of the path tracking system is obtained. On this basis, the generalized regression neural network is applied to the lateral inverse dynamics modeling of autonomous vehicles. With the normal driving of the vehicle, i.e., not colliding with the road boundary and not rolling over, as the constraint conditions, the input and output sets are selected, and the vehicle roll angle, roll angle velocity, and lateral displacement deviation y are extracted. e , the lateral orientation deviation e, to construct a multi-input and multi-output data set for the normal path tracking system of autonomous vehicles. For the neurons in the summation layer of the generalized regression neural network, a matching design is carried out for the weighted summation of data, and finally an inverse dynamics model that accurately reflects the lateral dynamics and preview tracking characteristics of the vehicle is obtained.

7. The path tracking and vehicle body attitude cooperative control method according to claim 6, characterized in that: Based on the two-degree-of-freedom vehicle model and the preview tracking model, an input-output set is obtained, and it is realized through the following steps: Based on the two-degree-of-freedom vehicle dynamics equation, a two-degree-of-freedom vehicle dynamics model is established. According to the relationship between the tire sideslip angle and the vehicle center-of-mass sideslip angle, yaw angular velocity, and the distance parameters of the center of mass from the front and rear axles, and combined with Newton's second law, the motion differential equation of the linear two-degree-of-freedom model is: where m is the vehicle mass, C f , Cr are the linear cornering stiffnesses of the front and rear wheels, l f , l r are the distances from the vehicle's center of mass to the front and rear axles, I Z is the moment of inertia of the whole vehicle about the Z-axis, δ f is the front wheel steering angle, V x is the longitudinal vehicle speed; Design of preview error system; the inputs of the preview error model include the sideslip angle and yaw rate of the vehicle center of mass output by the vehicle dynamics model, as well as the external inputs of road curvature ρ and preview distance x e , the outputs of the preview error model are the lateral displacement deviation ye and the lateral orientation deviation e, and the formula of the vehicle preview error model is constructed: where y is the distance deviation between the vehicle preview point and the center line, v y is the vehicle lateral translation speed, v x is the longitudinal vehicle speed, and r is the turning radius; Based on the data within the past one feedforward control period, the entropy weight fusion method is used to allocate weights to the lateral displacement deviation ye and the lateral orientation deviation e, and then they are used as the comprehensive deviation E. Input the obtained comprehensive deviation E into the PID controller, perform time-domain integration on the comprehensive deviations of at least two past feedforward control cycles, and use the PID control method to output the front wheel angle δ of the supplementary vehicle based on the integration result. f As the control quantity of the vehicle dynamics model, a closed-loop control system for path-tracking lateral motion is formed thereby.

8. The path tracking and vehicle body attitude cooperative control method according to claim 1, wherein: It further includes a driving state monitoring module. The optimal individual is used to optimize the weights to obtain the optimized weights. The optimized initial weights and thresholds are substituted into the neural network. By performing selection, crossover, and mutation operations, the individual corresponding to the optimal fitness value is found, the network error is calculated, and the weights and thresholds are modified according to the calculation results, and finally the required accuracy is achieved. Among them, the measured values of the three-axis acceleration and the environmental variables are used as the input nodes. The environmental variables depend on the situation and the specific values are determined after the network training. The vehicle body pitch angle and roll angle are used as the output nodes. The weights and thresholds are changed through network training until the accuracy meets the standard.

9. The path tracking and vehicle body attitude cooperative control method according to claim 1, characterized in that: The relative quantity of the ratio of the current lateral acceleration of the vehicle to the real-time lateral limit acceleration is used as the rollover determination condition to replace the traditional method of predicting rollover by relying on the roll angle or the absolute quantity of the lateral acceleration. The acceleration of the vehicle in the rollover critical state is: Where a y is the lateral acceleration at the center of gravity, a y,L is the critical lateral acceleration for rollover at the center of gravity, m is the vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the track width, h is the distance from the center of gravity to the roll center, is the roll angle L d = a y / a y,L where a y is the lateral acceleration at the center of gravity, a y,L is the critical lateral acceleration for rollover at the center of gravity, m is the total vehicle mass, m s is the sprung mass, h cm is the height of the center of gravity, T is the track width, h is the distance from the center of gravity to the roll center, is the roll angle Based on L d The value calculation rollover time limit, i.e., the TTR value, is calculated with a step size of 10 ms; Determine the TTR threshold for vehicle body attitude compensation according to the performance requirement bias. If the cooperative control performance requirement is biased towards tracking accuracy, the TTR threshold takes a larger value to make the vehicle center of gravity lower when the vehicle turns. If the cooperative control requirement is biased towards comfort, the TTR threshold takes a smaller value, and the TTR threshold is proportional to the vehicle speed.

10. The path tracking and vehicle body attitude collaborative control method according to claim 1, characterized in that: Among them, when the path tracking accuracy is lower than the set threshold, but the measurement values of the handling stability and comfort do not meet the set threshold, the handling stability and comfort of the vehicle are preferentially considered, and the vehicle body attitude compensation is performed on the vehicle. The vehicle body attitude compensation control method is realized through the following steps: Improve the vehicle body attitude by controlling the vehicle roll. The roll moment generated by the roll consists of three parts:

1. The roll moment M caused by the centrifugal force of the sprung mass ΦrΙ ; 2. Roll resistance moment M caused by the gravity of the sprung mass ΦrⅡ ; 3. The roll moment M caused by the centrifugal force of the unsprung mass ΦrⅢ ; In addition, during roll, the vertical load is transferred between the left and right wheel loads, generating a load transfer moment M ZF , M ZR ; when the vehicle body is in a roll state, the active suspension actuators on the left and right sides of the suspension apply additional forces Δf in opposite directions in the current state, which can form an anti-roll moment M af , which can suppress the roll of the vehicle; The roll moment M caused by the centrifugal force of the unsprung mass ΦrⅠ is as follows: M ΦrⅠ = m s · a y · h M ΦrⅠ is the roll moment caused by the centrifugal force of the unsprung mass, m s is the unsprung mass, a y is the vehicle body centroid acceleration, and h is the centroid height; The roll resistance moment M caused by the gravity of the unsprung mass ΦrⅡ is as follows: M ΦrⅡ is the roll moment caused by the centrifugal force of the unsprung mass, m s is the unsprung mass, e is the lateral distance from the center of mass to the wheel, a y is the acceleration of the center of mass of the vehicle body, h g is the height from the connection between the suspension on the lower side of the vehicle body and the vehicle body to the center of mass, is the roll angle of the vehicle body; The roll moment M caused by the centrifugal force of the unsprung mass ΦrⅢ is as follows: M ΦrⅢ = -Fuy(h0 - r) M ΦrⅢ is the rolling moment caused by the unsprung mass, F is the static friction force of the road surface on the wheel pointing towards the steering center, u is the vehicle speed, y is the lateral dimension of the vehicle body, h0 is the height of the center of mass, and r is the wheel radius; When the vehicle rolls, the vertical load causes a load transfer between the left and right wheels, generating load transfer moments M ZF and M ZR which are as follows: M ZF = (F rRF - F rLF ) · B / 2 M ZR = (F rRR - F rLR ) · B / 2 M ZF is the transfer torque of the front wheel, M ZR is the transfer torque of the rear wheel, F rRF is the transfer force of the right front wheel, F rLF is the transfer force of the left rear wheel, F rRR is the transfer force of the right rear wheel, F rRR , B is the body width; Take the moment about the longitudinal center line of the vehicle body, that is: M ΦrⅠ -M ΦrⅡ +M ΦrⅢ +M ZF +M ZR = M af M ΦrⅠ is the roll moment caused by the centrifugal force of the unsprung mass, M ΦrⅡ is the roll moment caused by the centrifugal force of the unsprung mass, M ΦrⅢ is the roll moment caused by the unsprung mass, M ZF is the front wheel transfer moment, M ZR is the rear wheel transfer moment, M af is the anti-roll moment provided by the actuator.

Images