Vehicle stability adaptive control method under low road surface multi-source interference

Abstract

The application discloses a vehicle stability adaptive control method under low adhesion road multi-source interference and belongs to the field of vehicle control, and comprises the following steps: S1, multi-source interference sensing and adhesion coefficient prediction; S2, dynamic correction of tire dynamics parameters: online correction of tire cornering stiffness and slip stiffness; S3, multi-source interference cooperative observation and estimation: dynamic estimation of interference sources by using an enhanced extended state observer; S4, phase plane stability envelope analysis and weight adjustment: dynamic allocation of control priorities; S5, dynamic torque distribution and actuator cooperative control under the premise of considering interference sources: optimization of four-wheel torque by using a quadratic programming algorithm to obtain a torque distribution result. By using the above-mentioned vehicle stability adaptive control method under low adhesion road multi-source interference, the problems of insufficient modeling accuracy, weak interference suppression, target conflict and redundancy loss in low adhesion road vehicle stability control are solved, and the trajectory tracking accuracy and yaw stability are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle control, and in particular to a vehicle stability adaptive control method under low adhesion road surface multi-source interference. BACKGROUND

[0002] When driving on low adhesion road surfaces (road adhesion coefficient ) such as ice and snow, wet, etc., the friction between the vehicle tires and the road surface decreases significantly, which easily leads to stability problems such as side slip and spin. At the same time, the vehicle often faces the superimposed effects of crosswind (> 15 m / s), road adhesion sudden change (such as alternating ice and dry road), brake hydraulic delay (0.8 ms / ℃), etc. under low adhesion conditions, which makes it difficult for traditional control systems to balance path tracking accuracy and yaw stability. For example, when crosswind and adhesion sudden change act together, the yaw rate overshoot of the traditional MPC system is 12° / s, and the side slip risk increases by 68%.

[0003] To solve the above problems, the existing technology mainly starts from the following aspects:

[0004] 1. Tire dynamics modeling: the existing technology generally uses linear tire models such as magic formula to simplify the non-linear relationship between side stiffness and slip rate through steady-state assumption. For example, the magic formula fits the relationship between side force and side angle through a polynomial, but ignores the coupling effect of slip rate and side angle. In addition, the traditional model does not consider the change of tire rubber hardness with temperature (-2.3% hardness / ℃) and the dynamic characteristics of road adhesion, resulting in a side force prediction error of more than 40% under low adhesion conditions.

[0005] 2. Single sensor sensing and fixed parameter control strategy: the existing system relies on ABS (anti-lock braking system), ESP (electronic stability program) and other single sensors (such as wheel speed, steering wheel angle) for stability control. When disturbances such as crosswind, adhesion sudden change, execution delay, etc. are superimposed, the traditional control strategy is difficult to dynamically adjust the compensation priority. For example, when crosswind (> 15 m / s) and adhesion sudden change act together, the yaw rate overshoot is 12° / s, and the response time of the traditional PID controller is more than 100 ms, which cannot effectively suppress the dynamic deviation. In addition, single sensors (such as wheel speed) cannot accurately distinguish the difference between ice and snow areas (laser radar reflection intensity <30) and wet road surfaces, resulting in insufficient sensing accuracy. Some studies attempt to use fuzzy control to compensate for crosswind interference, but the joint suppression rate of multiple source disturbances is less than 45%, and the delay characteristics of brake fluid viscosity with temperature are not considered, and the hydraulic response error is enlarged by 23%.

[0006] ​3. Limitations of actuator redundancy design: Existing redundancy mechanisms mainly focus on the controller level, such as the use of double-winding motors in steer-by-wire systems to achieve hardware redundancy, but there is a lack of actuator-level redundancy. For example, traditional ESP systems do not design compensation strategies for single-motor failure or hydraulic failure, and the loss of control distance in extreme scenarios can reach 55m. Although some studies attempt to achieve torque distribution through four-wheel independent drive, there is a lack of dynamic weight adjustment, and when the double electric drive bridge slips, equal torque reduction leads to a 19% increase in lateral force deviation.

[0007] 4. Insufficient decoupling of control objectives: Existing methods usually use fixed weight allocation to distribute path tracking and stability control objectives, resulting in decreased overall performance. For example, in the snow double lane shift test, the fixed weight strategy causes the center of mass side slip angle to exceed 5°, triggering a side slip warning. Although existing patents such as CN202411844853A propose a work-condition cooperative control strategy, they do not address the strong coupling contradiction between low-attachment road path tracking (lateral deviation <0.2m) and yaw stability (yaw rate <0.2rad / s).

[0008] In summary, existing technologies have significant defects in multi-source disturbance perception, tire model adaptability, redundancy fault tolerance capability, and control objective decoupling. SUMMARY

[0009] The purpose of the present application is to provide a vehicle stability adaptive control method under low-attachment road multi-source disturbance, which solves the above technical problems.

[0010] To achieve the above purpose, the present application provides a vehicle stability adaptive control method under low-attachment road multi-source disturbance, comprising the following steps:

[0011] S1, multi-source disturbance perception and adhesion coefficient prediction: using collected multi-modal environmental data and vehicle dynamics state data to estimate the road adhesion coefficient, and then fusing it with the road adhesion coefficient predicted by the spatio-temporal convolution network;

[0012] S2, dynamic correction of tire dynamics parameters: based on the change trend of the fused road adhesion coefficient prediction value and the tire temperature data, online correction of tire cornering stiffness and slip stiffness;

[0013] S3, multi-source disturbance cooperative observation and estimation: using an enhanced extended state observer to dynamically estimate the disturbance source;

[0014] S4, phase plane stability envelope analysis and weight adjustment: based on the dynamic prediction value of the adhesion coefficient, using Lyapunov stability theory to define the phase plane boundary and dynamically allocate control priority;

[0015] ​S5, dynamic torque distribution and actuator cooperative control: under the premise of considering the disturbance source, based on the weight dynamic distribution result and the corrected tire cornering stiffness, the four-wheel torque is optimized by using the quadratic programming algorithm to obtain the torque distribution result.

[0016] Preferably, step S1 specifically comprises the following steps:

[0017] S11, collecting multi-modal environment data and vehicle dynamics state data, and performing time-space alignment after filtering noise, wherein the multi-modal environment data includes point cloud elevation map and road image data in front of the vehicle, and environmental temperature and humidity, cross-wind disturbance moment, the vehicle dynamics state data includes vehicle roll angle , pitch angle , vehicle longitudinal speed and vehicle lateral speed ;

[0018] S12, comparing the reflection intensity value of each point in the point cloud elevation map with the set threshold value, if it is less than the set threshold value, it is determined as ice and snow covered area, and the ice and snow covered area is regarded as low adhesion road, if it is greater than the set threshold value, it is determined as non-low adhesion road;

[0019] and calculating the road adhesion coefficient of the low adhesion road based on the point cloud elevation map :

[0020] ;

[0021] In the formula, , the reflection intensity value is represented;

[0022] At the same time, the road image data is input into the YOLOv5 network for road texture feature segmentation, and the road adhesion coefficient of the low adhesion road is output by combining the ResNet-50 classification network :

[0023] ;

[0024] In the formula, , the probability of low adhesion road as visual semantics is represented;

[0025] Collecting historical road adhesion coefficient, environmental temperature and humidity, and vehicle dynamics state data, and inputting the collected data into the space-time convolution network, and predicting the road adhesion coefficient in the next 5s by using the space-time convolution network ;

[0026] And compensating the road slope interference based on the vehicle dynamics state data:

[0027] ;

[0028] In the formula, represents a slope compensation amount;

[0029] S13, the road adhesion coefficient of the low adhesion road calculated based on the point cloud elevation map , the road adhesion coefficient of the low adhesion road output by the ResNet-50 classification network , and the road adhesion coefficient of the future 5s predicted by the spatio-temporal convolution network are fused to obtain a fused road adhesion coefficient :

[0030] ;

[0031] wherein, , , and all represent weight coefficients, and .

[0032] Preferably, in step S11, a Velodyne VLS-128 laser radar is used to emit laser pulses to the road 50m in front of the vehicle, and a point cloud elevation map is generated by recording the three-dimensional coordinates of the reflected signals. The point cloud elevation map includes the three-dimensional coordinates of each point and its reflection intensity value.

[0033] Preferably, step S2 specifically includes the following steps:

[0034] S21, based on real-time dynamic correction of the cornering stiffness, compensating for the slip rate-cornering angle coupling effect:

[0035] ;

[0036] wherein, represents the corrected cornering stiffness; represents the initial cornering stiffness;

[0037] S22, using a double Kalman filter to update the slip stiffness decay factor and the temperature compensation coefficient in real time through tire slip rate feedback and tire temperature data, respectively;

[0038] S23, based on the updated slip stiffness decay factor and the temperature compensation coefficient and the corrected cornering stiffness, correcting the tire cornering force.

[0039] Preferably, step S3 specifically includes the following steps:

[0040] S31, setting the disturbance source as the crosswind disturbance moment, the road adhesion gradient and the brake hydraulic delay, regarding the disturbance source as an extended state, constructing an augmented state space containing vehicle dynamics state data and extended state, and obtaining the disturbance source estimation value :

[0041] ;

[0042] In the formula, represents a weight coefficient; represents a low-pass filter, represents a time constant, represents a Laplace variable; represents a measured value; represents an output prediction value based on a vehicle dynamics model; represents a type of disturbance source, , respectively corresponding to the crosswind disturbance moment, the road adhesion gradient and the brake hydraulic pressure delay;

[0043] S32, considering the space-time correlation of crosswind, road sudden change, actuator error, dynamically updating the weight coefficient of the enhanced extended state observer by the unscented Kalman filter ;

[0044] S33, outputting the estimated crosswind disturbance moment, road adhesion gradient and brake hydraulic pressure delay.

[0045] Preferably, step S4 specifically comprises the following steps:

[0046] S41, defining the yaw rate boundary and the center of mass side slip angle boundary using Lyapunov stability theory.

[0047] ;

[0048] ;

[0049] In the formula, represents the acceleration of gravity;

[0050] S42, dividing the warning levels according to the distance between the vehicle state and the stability boundary , and switching the control mode based on the divided warning levels.

[0051] Preferably, in step S42, a dynamic weight adjustment mechanism is set:

[0052] ;

[0053] In the formula, represents a dynamic weight; represents an adjustment factor, and ; represents a critical threshold, and ;

[0054] When When the , and activate longitudinal acceleration limitation;

[0055] when When the warning is level 2, , activate torque distribution and limit the front wheel steering angle change rate to ≤15° / s;

[0056] when It is designated as Level 3 warning. , switch to stability priority mode and turn off ESP intervention delay.

[0057] Preferably, step S5 specifically includes the following steps:

[0058] S51. Construct a 19-DOF vehicle model that includes suspension geometric nonlinearity, tire relaxation effects, and motor torque saturation characteristics:

[0059] ;

[0060] Where, and Represent the vehicle longitudinal acceleration and vehicle lateral acceleration respectively; Indicates the vehicle mass; 、 、 and Represent the longitudinal forces of the front left, front right, rear left, and rear right wheels respectively; 、 、 and Represents the lateral forces of the front left, front right, rear left, and rear right wheels respectively; represents the yaw angular velocity; represents the derivative of the yaw rate; Indicates that the vehicle is around The moment of inertia of the shaft; 、 、 and Represent the torque of the front left, front right, rear left, and rear right wheels respectively;

[0061] S52: Minimizing the torque deviation and yaw moment deviation is the goal, and optimizing the torque distribution by considering motor saturation, battery power constraint, and slip ratio constraint. The objective function expression is as follows:

[0062] ;

[0063] Where, Indicates wheels The actual torque; and Represents wheels target torque and maximum torque; denotes an interference source, denote a crosswind disturbance moment, a road adhesion gradient and brake hydraulic delay; denotes a weight coefficient; denotes a yaw moment compensation amount;

[0064] S53, solving the objective function to obtain four-wheel actual torques and a yaw moment compensation amount .

[0065] Preferably, in step S5, a redundant fault-tolerant control and actuator compensation are also provided: whether ESP failure or single motor failure occurs is detected, and if so, stability is maintained through differential torque and steering angle compensation based on the torque distribution result.

[0066] Preferably, when single-wheel slip ratio > 30% or ESP failure signal is detected, inverse yaw moment is generated through double-motor differential, and the maximum adjustment range is set to ±1200 Nm;

[0067] and based on actively adjusting the front wheel steering angle, and limiting the steering angle speed to 30° / s.

[0068] Therefore, the vehicle stability adaptive control method under the above low adhesion road multi-source disturbance has the beneficial effects of:

[0069] 1. High-precision adhesion coefficient estimation: through laser radar (reflection intensity threshold), visual semantic segmentation (YOLOv5+ResNet-50), IMU (inertial unit) slope compensation, and spatio-temporal convolution network (ST-CNN) prediction, multi-source data fusion is realized, the low adhesion road adhesion coefficient estimation error is controlled to be less than or equal to 0.03, and the value mutation (such as ice-asphalt boundary) is predicted 5 seconds in advance, solving the problem of misjudgment of traditional single sensor (such as night ice and snow recognition accuracy improvement of 50%);

[0070] 2. Full-scene environment recognition: the laser radar point cloud reflection intensity accurately identifies the ice and snow area, the visual system supplements the texture feature recognition under thin ice and strong light, and the IMU compensates for the slope interference, realizing real-time detection of low adhesion road in a temperature range of -40°C to 85°C and a 15m pre-look distance, covering complex working conditions such as ice and snow, wet and slippery, and sandstone;

[0071] 3. Introducing a slip rate-side angle coupling term, the nonlinear tire force is calculated in real time through the finite difference method, breaking through the steady-state assumption of the traditional linear model (such as the magic formula), and improving the tire force prediction accuracy in extreme conditions (such as ice surface sharp turning); ​

[0072] 4、Enhanced Extended State Observer (EESO) combined with UKF (Unscented Kalman Filter) dynamic weight update, realizes the cooperative observation of cross-wind disturbance torque (range ±200Nm, resolution ±5Nm), road adhesion gradient (range 0-1 / m, resolution 0.01 / m), brake hydraulic delay (compensation accuracy ±2ms), and the disturbance estimation accuracy is ≥92% (RMS error <0.02), which is 47% higher than that of the traditional observer;

[0073] 5、Under the condition of cross-wind (20m / s) + sudden change (0.2→0.1), the yaw rate overshoot is reduced from 12° / s to 5.1° / s, and the side slip risk is reduced by 67%; when the hydraulic response delay is >50ms, the control command tracking error is ≤±2%, and the response time is shortened to 4ms, solving the problem of loss of control caused by disturbance superposition in traditional systems;

[0074] 6、Through multi-modal perception (laser radar + vision + IMU) and spatio-temporal convolution network (ST-CNN), the adhesion coefficient prediction with a 15m preview distance is realized, and the lateral control demand of L4 level automatic driving is supported; for example, in the snow deer test, the system response speed is 30 times faster than human neural reflex, and the trajectory deviation is reduced by 72%.

[0075] The technical solutions of the present application will be further described in detail below with the aid of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0076] Figure 1 is a flow chart of the vehicle stability adaptive control method of the present application under low adhesion road multi-source disturbance. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be further described in detail below with the aid of the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present application and do not limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application. The examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions throughout.

[0078] ​It is to be understood that the terms "including", "comprising", "having" and their conjugates mean "including but not limited to", e.g. a process, method, object, or apparatus that includes a list of steps or elements is not necessarily limited to those specifically listed and can include other steps or elements not expressly listed or inherent to such process, method, object, or apparatus.

[0079] The embodiments of the present application will be described in detail below with reference to the drawings.

[0080] As shown in the drawings, a vehicle stability adaptive control method under low adhesion road multi-source interference includes the following steps: Figure 1

[0081] S1, multi-source interference perception and adhesion coefficient prediction: using the collected multi-modal environment data and vehicle dynamics state data to estimate the road adhesion coefficient, and then fusing it with the road adhesion coefficient predicted by the spatio-temporal convolution network;

[0082] Step S1 specifically includes the following steps:

[0083] S11, collect multi-modal environment data and vehicle dynamics state data, and perform spatio-temporal alignment after filtering noise, wherein the multi-modal environment data includes point cloud elevation map and road image data in front of the vehicle, as well as environmental temperature and humidity, cross-wind disturbance torque, and the vehicle dynamics state data includes vehicle roll angle , pitch angle , vehicle longitudinal speed and vehicle lateral speed ;

[0084] In step S11, a Velodyne VLS-128 laser radar is used to emit laser pulses to the road in front of the vehicle 50m away, and record the three-dimensional coordinates of the reflected signals to generate a point cloud elevation map, which includes the three-dimensional coordinates of each point and its reflection intensity value.

[0085] S12, compare the reflection intensity value of each point in the point cloud elevation map with the set threshold value, if it is less than the set threshold value, it is determined as an ice and snow covered area, and the ice and snow covered area is regarded as a low adhesion road, if it is greater than the set threshold value, it is determined as a non-low adhesion road;

[0086] and the road adhesion coefficient of the low adhesion road is calculated based on the point cloud elevation map :

[0087] ;

[0088] In the formula, represents the reflection intensity value;

[0089] ​Meanwhile, the pavement image data is input into the YOLOv5 network to perform pavement texture feature segmentation, and the pavement adhesion coefficient of the low adhesion pavement is output by combining the ResNet-50 classification network :

[0090] ;

[0091] In the formula, represents the probability of the low adhesion pavement being a visual semantic;

[0092] The historical pavement adhesion coefficient (10 frames, 0.5s interval), environmental temperature and humidity, and vehicle dynamics state data are collected, and the collected data is input into the spatio-temporal convolution network to predict the pavement adhesion coefficient in the next 5s ;

[0093] And the pavement slope interference is compensated based on the vehicle dynamics state data:

[0094] ;

[0095] In the formula, represents the slope compensation amount;

[0096] S13, the pavement adhesion coefficient of the low adhesion pavement calculated based on the point cloud elevation map , the pavement adhesion coefficient of the low adhesion pavement output by the ResNet-50 classification network , and the pavement adhesion coefficient of the low adhesion pavement predicted by the spatio-temporal convolution network in the next 5s are fused to obtain the fused pavement adhesion coefficient :

[0097] ;

[0098] In the formula, , , and all represent weight coefficients, and .

[0099] S2, dynamic correction of tire dynamics parameters: based on the change trend of the fused pavement adhesion coefficient prediction value and the tire temperature data, the tire cornering stiffness and slip stiffness are corrected online;

[0100] Step S2 specifically includes the following steps:

[0101] S21, based on real-time dynamic correction of the cornering stiffness, compensation of the slip rate-cornering angle coupling effect:

[0102] ;

[0103] Where, represents the corrected cornering stiffness; represents the initial cornering stiffness;

[0104] After testing and correction, the lateral force prediction error is The range is reduced to within 12%.

[0105] S22: Using a dual Kalman filter to update the slip stiffness attenuation factor (0.6-1.2) and temperature compensation coefficient (-0.0085 / °C) in real time using tire slip rate feedback and tire temperature data.

[0106] S23. Correct the tire lateral force based on the updated slip stiffness attenuation factor and temperature compensation coefficient and the corrected cornering stiffness.

[0107] S3. Collaborative observation and estimation of multi-source interference: Dynamically estimate interference sources using the enhanced extended state observer;

[0108] Step S3 specifically includes the following steps:

[0109] S31. Set the interference sources as crosswind disturbance torque, road adhesion gradient and brake hydraulic delay, and regard the interference sources as expanded states. Construct an augmented state space containing vehicle dynamics state data and expanded states to obtain the interference source estimation value. :

[0110] ;

[0111] Where, represents the weight coefficient; represents a low-pass filter, represents the time constant, represents the Laplace variable; Indicates the measured value; represents the output prediction value based on the vehicle dynamics model; Indicates the type of interference source, , corresponding to the crosswind disturbance torque, road adhesion gradient and brake hydraulic delay respectively;

[0112] S32, considering the spatiotemporal correlation of crosswind, road surface mutation, and actuator error, dynamically updating the weight coefficients of the enhanced extended state observer through unscented Kalman filtering ;

[0113] S33, output estimated crosswind disturbance torque, road adhesion gradient and brake hydraulic delay.

[0114] S4, phase plane stability envelope analysis and weight adjustment: based on the dynamic prediction of the adhesion coefficient, the phase plane boundary is defined by using Lyapunov stability theory, and the control priority is dynamically allocated;

[0115] Step S4 specifically comprises the following steps:

[0116] S41, define the yaw rate boundary by using Lyapunov stability theory and the center of mass side slip angle boundary

[0117]

[0118]

[0119] In the formula, g represents the acceleration of gravity;

[0120] When , , the yaw rate boundary is , and the center of mass side slip angle boundary is .

[0121] S42, according to the distance between the vehicle state and the stability boundary , the warning level is divided, and the control mode is switched based on the divided warning level.

[0122] Preferably, in step S42, a dynamic weight adjustment mechanism is set:

[0123]

[0124] In the formula, g represents the dynamic weight; g represents the adjustment factor, and g represents the critical threshold, and

[0125] When , it is classified as a first-level warning, , and the longitudinal acceleration limit is started;

[0126] When , it is classified as a second-level warning, , the torque distribution is activated, and the front wheel steering angle change rate is limited to ≤15° / s;

[0127] When , it is classified as a third-level warning, , the stability priority mode is switched on, and the ESP intervention delay is closed.

[0128] ​​​​​​S5. Dynamic torque distribution and actuator coordinated control: Taking interference sources into consideration, the quadratic programming algorithm is used to optimize the four-wheel torque based on the weighted dynamic distribution results and the corrected tire cornering stiffness to obtain the torque distribution result.

[0129] Step S5 specifically includes the following steps:

[0130] S51, build a system that includes suspension geometric nonlinearity, tire relaxation effect (first-order inertia link, time constant ) and motor torque saturation (peak torque ±1500Nm) and characteristics of the 19-DOF vehicle model:

[0131] ;

[0132] Where, and Represent the vehicle longitudinal acceleration and vehicle lateral acceleration respectively; Indicates the vehicle mass; 、 、 and Respectively represent the longitudinal forces of the front left, front right, rear left, and rear right wheels, which are calculated in step S23; 、 、 and Respectively represent the lateral forces of the front left, front right, rear left, and rear right wheels, which are calculated in step S23; represents the yaw angular velocity; represents the derivative of the yaw rate; Indicates that the vehicle is around The moment of inertia of the shaft; 、 、 and Represent the torque of the front left, front right, rear left, and rear right wheels respectively;

[0133] S52: Minimizing the torque deviation and yaw moment deviation is the goal, and optimizing the torque distribution by considering motor saturation, battery power constraint, and slip ratio constraint. The objective function expression is as follows:

[0134] ;

[0135] Where, Indicates wheels The actual torque; and Represents wheels Target torque and maximum torque; Indicates the source of interference, Represent the crosswind disturbance torque and road adhesion gradient respectively and brake hydraulic pressure delay; denotes a weight coefficient; denotes a yaw moment compensation amount;

[0136] For example, in the split road surface (left wheel , right wheel ) working condition, the optimizer reduces the right front wheel torque by 40% (300→180 Nm) and increases the left rear wheel torque by 25% (320→400 Nm), forming an inverse yaw moment of 12 Nm·m.

[0137] S53, solving the objective function to obtain the actual torque of the four wheels and the yaw moment compensation amount .

[0138] In step S5, a redundant fault-tolerant control and actuator compensation are also provided: whether ESP failure or single motor failure occurs is detected, and if so, based on the torque distribution result, differential torque and steering angle compensation are used to maintain stability. Specifically, when a single wheel slip rate > 30% or an ESP failure signal is detected, an inverse yaw moment is generated by differential torque of the double motors, and the maximum adjustment range is set to ±1200 Nm;

[0139] and based on active adjustment of the front wheel steering angle, and the steering angle speed is limited to 30° / s.

[0140] To prove the present application, the following working conditions are further verified:

[0141] Ice and snow double lane change test: lateral error ≤0.25 m (53% lower than traditional MPC), peak yaw rate ≤6° / s (58% lower);

[0142] Ice surface braking: braking distance 28 m (38% shorter), vehicle body offset angle <3°.

[0143] Crosswind Abrupt working condition: when the crosswind 20 m / s is superimposed from 0.2→0.1, the center of mass sideslip angle overshoot is <1.2°, and the sideslip risk is reduced by 67%;

[0144] Actuator delay compensation: when the hydraulic delay is >50 ms, the control command tracking error is ≤±2%, and the response time is shortened to 4 ms.

[0145] ESP failure scenario: the ASC / DST module maintains a 0.3g deceleration, and the out-of-control distance is shortened by 42% (from 55 m→32 m);

[0146] Single motor failure: after the remaining three motors are redistributed, the yaw angular velocity deviation is <2° / s, and the lateral deviation is <0.5m.

[0147] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, but not to limit them. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can still be modified or equivalently replaced, and these modifications or equivalent replacements should not make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for adaptively controlling vehicle stability under multi-source disturbances on low-adhesion roads, characterized by: The following steps are involved: S1. Multi-source interference perception and adhesion coefficient prediction: The road adhesion coefficient is estimated using collected multimodal environmental data and vehicle dynamics data, and then fused with the road adhesion coefficient predicted using a spatiotemporal convolutional network. S2. Dynamic correction of tire dynamic parameters: Based on the changing trend of the fused road adhesion coefficient prediction value and tire temperature data, the tire cornering stiffness and slip stiffness are corrected online; Step S2 specifically includes the following steps: S21, road adhesion coefficient based on real-time fusion Dynamically correct cornering stiffness to compensate for the slip rate-slip angle coupling effect: ； Where, represents the corrected cornering stiffness; represents the initial cornering stiffness; S22, using a dual Kalman filter to update the slip stiffness attenuation factor and the temperature compensation coefficient in real time based on tire slip rate feedback and tire temperature data; S23. Correcting the tire lateral force based on the updated slip stiffness attenuation factor and temperature compensation coefficient and the corrected cornering stiffness; S3. Collaborative observation and estimation of multi-source interference: Dynamically estimate interference sources using the enhanced extended state observer; Step S3 specifically includes the following steps: S31. Set the interference sources as crosswind disturbance torque, road adhesion gradient and brake hydraulic delay, and regard the interference sources as expanded states. Construct an augmented state space containing vehicle dynamics state data and expanded states to obtain the interference source estimation value. : ； Where, represents the weight coefficient; represents a low-pass filter, represents the time constant, represents the Laplace variable; Indicates the measured value; represents the output prediction value based on the vehicle dynamics model; Indicates the type of interference source, , corresponding to the crosswind disturbance torque, road adhesion gradient and brake hydraulic delay respectively; S32, considering the spatiotemporal correlation of crosswind, road surface mutation, and actuator error, dynamically updating the weight coefficients of the enhanced extended state observer through unscented Kalman filtering ; S33, output estimated crosswind disturbance torque, road adhesion gradient and brake hydraulic delay; S4, Phase plane stability envelope analysis and weight adjustment: Based on the dynamic prediction value of the adhesion coefficient, the Lyapunov stability theory is used to define the phase plane boundary and dynamically allocate control priorities; S5. Dynamic torque distribution and actuator coordinated control: Taking interference sources into consideration, the torque distribution result is obtained by optimizing the four-wheel torque using a quadratic programming algorithm based on the weighted dynamic distribution results and the corrected tire cornering stiffness.

2. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11. Collect multimodal environmental data and vehicle dynamics data, and perform spatiotemporal alignment after filtering noise. The multimodal environmental data includes the point cloud elevation map and road image data in front of the vehicle, as well as the ambient temperature and humidity, and crosswind disturbance torque. The vehicle dynamics data includes the vehicle's roll angle. , pitch angle , vehicle longitudinal speed and vehicle lateral speed ; S12. Compare the reflection intensity value of each point in the point cloud elevation map with a set threshold. If the reflection intensity value is less than the set threshold, the area is determined to be covered by ice and snow and regarded as a low-adhesion road surface. If the reflection intensity value is greater than the set threshold, the area is determined to be a non-low-adhesion road surface. And calculate the road adhesion coefficient of low adhesion road based on point cloud elevation map : ； Where, Indicates the reflection intensity value; At the same time, the road surface image data is input into the YOLOv5 network for road surface texture feature segmentation, and combined with the ResNet-50 classification network to output the road adhesion coefficient of low-adhesion road surface : ； Where, Indicates the probability that low-adhesion road surface is a visual semantic; Collect historical road adhesion coefficient, ambient temperature and humidity, and vehicle dynamics status data, input the collected data into the spatiotemporal convolutional network, and use the spatiotemporal convolutional network to predict the road adhesion coefficient in the next 5 seconds. ; And compensate for road slope disturbance based on vehicle dynamics data: ； Where, Indicates the slope compensation amount; S13, the road adhesion coefficient of the low-adhesion road surface calculated based on the point cloud elevation map , Road adhesion coefficient of low-adhesion road based on the output of ResNet-50 classification network And the road adhesion coefficient predicted by the spatiotemporal convolutional network in the next 5 seconds Fusion is performed to obtain the fused road adhesion coefficient : ； Where, 、 、 and are weight coefficients, and .

3. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 2, characterized in that: In step S11, a Velodyne VLS-128 laser radar is used to emit laser pulses to the road surface 50 meters in front of the vehicle, and the three-dimensional coordinates of the reflected signals are recorded to generate a point cloud elevation map. The point cloud elevation map includes the three-dimensional coordinates of each point and its reflection intensity value.

4. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 3, characterized in that: Step S4 The specific steps include: S41. Define the yaw rate boundary using Lyapunov stability theory and the center of mass sideslip angle boundary : ； ； Where, represents the acceleration due to gravity; S42, based on the distance between the vehicle state and the stability boundary The warning levels are divided and the control mode is switched based on the divided warning levels.

5. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 4, characterized in that: In step S42, a dynamic weight adjustment mechanism is set: ； Where, Represents dynamic weight; represents the adjustment factor, and ; represents the critical threshold, and ; when When the , and activate longitudinal acceleration limitation; when It is designated as Level 2 warning. , activate torque distribution and limit the front wheel steering angle change rate to ≤15° / s; when It is designated as Level 3 warning. , switch to stability priority mode and turn off ESP intervention delay.

6. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 5, characterized in that: Step S5 specifically includes the following steps: S51. Construct a 19-DOF vehicle model that includes suspension geometric nonlinearity, tire relaxation effects, and motor torque saturation characteristics: ； Where, and Represent the vehicle longitudinal acceleration and vehicle lateral acceleration respectively; Indicates the vehicle mass; 、 、 and Represent the longitudinal forces of the front left, front right, rear left, and rear right wheels respectively; 、 、 and Represents the lateral forces of the front left, front right, rear left, and rear right wheels respectively; represents the yaw angular velocity; represents the derivative of the yaw rate; Indicates that the vehicle is around The moment of inertia of the shaft; 、 、 and Represent the torque of the front left, front right, rear left, and rear right wheels respectively; S52: Minimizing the torque deviation and yaw moment deviation is the goal, and optimizing the torque distribution by considering motor saturation, battery power constraint, and slip ratio constraint. The objective function expression is as follows: ； Where, Indicates wheels The actual torque; and Represents wheels Target torque and maximum torque; Indicates the source of interference, Represent the crosswind disturbance torque and road adhesion gradient respectively and brake hydraulic delay; represents the weight coefficient; Indicates the yaw moment compensation amount; S53. Solve the objective function to obtain the actual torque of the four wheels and yaw moment compensation .

7. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 6, characterized in that: In step S5, redundant fault-tolerant control and actuator compensation are also provided: whether ESP failure or single motor failure occurs is detected. If so, stability is maintained through differential torque and steering angle compensation based on the torque distribution result.

8. The method for adaptively controlling vehicle stability under multi-source interference on low-adhesion roads according to claim 7, characterized in that: When a single wheel slip rate > 30% or an ESP failure signal is detected, a counter-yaw torque is generated through the dual motor differential, and the maximum adjustment range is set to ±1200Nm; Based on Actively adjust the front wheel steering angle and limit the steering angle speed to 30° / s.

Images