A control method and system of an active front wheel steering system based on an LQR algorithm

Abstract

The application relates to the technical field of vehicle control, and discloses a control method and system for an active front wheel steering system based on an LQR algorithm, wherein the control method estimates a vehicle mass center side slip angle in real time through a Kalman filter observer, combines a two-degree-of-freedom vehicle dynamics model to construct a state space equation, and adopts an LQR controller to calculate an optimal front wheel steering angle compensation amount; the system compares the deviation of actual and ideal yaw angular velocities and mass center side slip angles, dynamically adjusts the front wheel steering angle, and realizes high-precision trajectory tracking and stable control of the vehicle under complex dynamic working conditions. The application has the advantages of high control precision, fast response, strong anti-interference capability and low energy consumption, is suitable for improving the vehicle control stability and driving safety, and is especially suitable for intelligent driving and an advanced auxiliary driving system.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to an active front wheel steering system control method and system based on the LQR algorithm. Background Technology

[0002] With the rapid popularization of automobiles, various traffic accidents are also increasing. These accidents are caused by many factors, such as sudden crosswind disturbances, complex road conditions, driver fatigue, and vehicle electronic system malfunctions. To reduce the accident rate, vehicle handling stability needs further optimization. The quality of the steering system largely determines the vehicle's handling stability. In vehicle dynamics, yaw rate (the rate of rotation of the vehicle around its vertical axis) and sideslip angle (the angle between the vehicle's center of gravity and the longitudinal axis) are core indicators for measuring handling stability; ideally, both must be strictly constrained within safe thresholds (e.g., yaw rate deviation <0.1 rad / s, sideslip angle <3°) to avoid skidding or loss of control. However, traditional mechanical steering structures have fixed transmission ratios, making it difficult to dynamically adapt to complex driving environments. Especially under dynamic conditions such as transient steering (e.g., emergency obstacle avoidance), low-traction surfaces (e.g., icy or slippery surfaces), or sudden lateral disturbances (e.g., strong crosswinds), the system cannot adjust steering response in real time, causing yaw rate and sideslip angle to rapidly deviate from the ideal stable region, easily leading to accidents such as fishtailing and rollovers. Furthermore, steering is not light at low speeds and difficult to control at high speeds. In addition, existing mass-produced simple electronic steering systems (such as basic electric power steering EPS), while improving steering lightness, lack closed-loop control algorithms for vehicle stability. In the aforementioned complex dynamic conditions, they still cannot effectively suppress the oscillations or divergence of yaw rate and sideslip angle; their control effect is limited to static or slowly changing conditions and is insufficient to cope with transient disturbances. Therefore, ensuring that the yaw rate and sideslip angle remain within the ideal stable region in real time under complex dynamic conditions, while simultaneously improving steering ability and handling stability, is the core challenge in the current development of steering systems. In addition to steering capability and handling stability, the vehicle steering system should also save space to facilitate the vehicle chassis layout.

[0003] In recent years, some driver assistance and intelligent driving technologies have developed rapidly. Mass production of technologies aimed at improving vehicle steering performance, especially active front-wheel steering, has begun. Furthermore, with the advancement of automotive electronics, the steering system and electronic control technology are becoming increasingly intertwined. Improving vehicle steering capability through control algorithms is currently the mainstream approach in steering system development. However, existing algorithms (such as PID control or fixed-gain feedback) have significant limitations under dynamic conditions: their control laws do not fully consider the coupling characteristics of yaw rate and sideslip angle, failing to minimize control energy consumption while ensuring stability. This results in vehicles easily entering unstable regions on low-traction surfaces or under lateral disturbances.

[0004] Active Front Steering (AFS) is a crucial component of electronic steering technology in vehicles, widely used in modern automobiles, especially in high-end passenger cars and intelligent driving systems. This system actively adjusts the front wheel angle to improve vehicle handling stability and safety under various driving conditions. Specific applications include:

[0005] Emergency obstacle avoidance and transient steering: When encountering sudden obstacles at high speeds, traditional steering systems respond slowly, which can easily lead to loss of vehicle control; the AFS system can adjust the steering angle in real time according to the vehicle's status, helping the driver to complete obstacle avoidance maneuvers quickly and stably.

[0006] Driving on low-friction surfaces: On roads with low traction coefficients such as ice, snow, and wet surfaces, vehicles are prone to sideslip; the AFS system maintains vehicle stability by adjusting the front wheel steering angle to suppress excessive changes in yaw rate and center of gravity sideslip angle.

[0007] Stability control under crosswind disturbance: When encountering strong crosswinds on highways or bridges, the AFS system can compensate for vehicle attitude deviation caused by wind disturbance in real time to prevent the vehicle from deviating from the lane.

[0008] Intelligent driving and assisted steering: In autonomous driving or lane keeping assist systems, AFS, as an actuator, can adjust the steering in real time according to the planned path and vehicle status to achieve precise trajectory tracking.

[0009] Although the AFS system has been applied in some vehicle models, the following technical issues still exist in actual use:

[0010] Technical problem one: Insufficient dynamic adaptability of the control algorithm. Existing AFS systems mostly use PID or fixed gain control, which does not fully consider the dynamic coupling characteristics of yaw rate and center of gravity sideslip angle. The control effect is poor under transient or strong disturbance conditions, which can easily lead to vehicle instability.

[0011] The second technical problem is the low accuracy of state estimation. The sideslip angle of the center of gravity is a key state variable that affects vehicle stability, but it is difficult to measure directly. Existing systems mostly rely on simplified models or low-precision sensors for estimation, which results in large errors under dynamic conditions and affects the control effect.

[0012] The third technical problem is that it is difficult to balance energy consumption and stability. Traditional control methods often sacrifice energy consumption for stability, or vice versa, failing to achieve the optimal balance between the two under dynamic operating conditions.

[0013] Technical Issue 4: System response delay and overshoot. During emergency steering or sudden road changes, the existing system often exhibits response delay or overshoot, causing the yaw rate and center of gravity sideslip angle to deviate from the safe range, increasing the risk of accidents.

[0014] The fifth technical problem is the weak anti-interference capability. When faced with sudden interference from crosswinds or uneven road surfaces, the existing system lacks an effective real-time compensation mechanism, which makes vehicles prone to swaying or trajectory deviation.

[0015] In summary, those skilled in the art urgently need a control technology for an active front wheel steering system that achieves high precision, low energy consumption, and strong robustness in vehicles under complex dynamic conditions. Summary of the Invention

[0016] The core technical problem to be solved by this invention is: how to minimize control energy consumption and improve the system's response speed and stability margin under transient and disturbance conditions, while ensuring that the yaw rate and center of mass sideslip angle strictly follow the ideal values.

[0017] To address the aforementioned technical problems, this invention presents an active front wheel steering system control method and system based on the LQR algorithm, which mainly includes the following:

[0018] Construct a Kalman filter observer to estimate the centroid sideslip angle, which is difficult to measure directly, in real time with high accuracy;

[0019] A two-degree-of-freedom vehicle dynamics state-space model is established, with yaw rate and sideslip angle as the core state variables;

[0020] Design an LQR controller and obtain the optimal state feedback gain by solving the Riccati equation to achieve real-time optimal adjustment of the front wheel steering angle;

[0021] A closed-loop control system of "state observation-optimal feedback" is formed, which minimizes control energy consumption while ensuring tracking accuracy and improving the system's anti-interference capability and dynamic stability.

[0022] It should be noted that LQR (Linear Quadratic Regulator) control is a type of optimal control. This invention innovatively applies it to the steering system. By constructing a state-space model centered on yaw rate and sideslip angle, the trajectory tracking problem is transformed into a real-time optimal adjustment problem targeting the errors of these two parameters. LQR dynamically optimizes the front wheel steering input by adjusting the state feedback gain matrix, ensuring that the yaw rate and sideslip angle are always constrained within the ideal stable region (e.g., through quadratic performance index weighting, the parameter deviation convergence time is <0.5s), while significantly reducing control energy consumption. This method requires minimal resources (only the conventional computing power of the vehicle ECU) and effectively overcomes the inherent defects of traditional mechanical steering and simple electronic steering under complex dynamic conditions. It can be used to solve key problems in vehicle stability control, fundamentally improving the active safety of the vehicle.

[0023] The Kalman filter, essentially an "optima recursive data processing algorithm," is composed of a series of mathematical formulas derived from linear algebra and hidden Markov models. When obtaining an estimate of the current time, this algorithm only needs to know the current observation time and the previous time's estimate, making it a method that recursively derives the minimum variance. It can estimate the minimum mean square error, achieving the expected requirements relatively well. The Kalman filter algorithm has advantages such as fast operation, simple and efficient algorithm, and good applicability to real-time signals in control systems, making it suitable for solving many engineering problems and thus widely used by most scholars. Therefore, the Kalman filter observer can provide an accurate, reliable, and difficult-to-measure centroid sideslip angle state estimate for LQR, thus forming a complete "state observation-optimal feedback" control closed loop.

[0024] The specific technical solution of the present invention to achieve the above objectives is a control method for an active front wheel steering system based on the LQR algorithm, comprising the following steps:

[0025] S1. The vehicle's center of gravity sideslip angle is estimated in real time using a Kalman filter observer; the principle of the Kalman filter observer is as follows: Figure 2 As shown.

[0026] S2. Construct state-space equations based on a two-degree-of-freedom vehicle dynamics model; represent the vehicle's dynamics in state-space form using the state-space method. The state equations of the two-degree-of-freedom linear vehicle model are:

[0027] (2.1)

[0028] State variables in the formula Input

[0029] Specifically, it is expressed as follows:

[0030] (2.2)

[0031] in:

[0032]

[0033] The initial state of the system is ;

[0034] For an ideal two-degree-of-freedom model, the system outputs, with vehicle speed and front wheel steering angle as inputs, are the two response quantities: sideslip angle and yaw rate, which are referenced from the vehicle's actual yaw rate during operation. and centroid side slip angle To distinguish them, the output value of the derived ideal two-degree-of-freedom model is defined as the reference yaw rate. And reference centroid side slip angle This serves as a reference value for the AFS controller when controlling the vehicle's steering.

[0035] S3. Design an LQR controller and obtain the optimal state feedback gain by solving the Riccati equation;

[0036] S4. Calculate the additional front wheel steering angle compensation based on the deviation between the actual and ideal yaw rate and the center of gravity sideslip angle.

[0037] S5. The compensation amount is superimposed on the front wheel steering angle input by the driver and output to the steering actuator.

[0038] The Kalman filter observer in S1 includes two parts: time update and measurement update. The Kalman filter observer is used to estimate the minimum variance of the centroid sideslip angle.

[0039] It should be noted that the observer built in this invention is based on the Kalman filter algorithm and is used to estimate the real-time centroid sideslip angle of the vehicle. The specific idea is to use feedback control to reflect the specific process state of the algorithm. This part consists of two main parts: time and the update of the measurement equation.

[0040] (1) Update of the time equation;

[0041] When estimating state variables backward:

[0042] (1.1)

[0043] The forward calculation of the error covariance is:

[0044] (1.2)

[0045] (2) Updating the measurement equation;

[0046] The Kalman gain expression is defined as follows:

[0047] (1.3)

[0048] The estimate is updated again after observing the variables:

[0049] (1.4)

[0050] The covariance of the updated error is:

[0051] (1.5)

[0052] In the above formula: the system parameters are A and B respectively; Q represents the process excitation noise covariance matrix, which is a symmetric non-negative definite moment; R represents the observation noise covariance matrix, which is a positive definite matrix, and it is assumed that they are not affected by state changes.

[0053] The state variables in the state-space equation of S2 include: yaw rate and center of mass sideslip angle, and the input is the front wheel steering angle.

[0054] The performance index function of the LQR controller in S3 is a quadratic integral form, and optimal control is achieved by adjusting the state weighting matrix and the control weighting matrix.

[0055] The additional front wheel steering angle compensation is calculated using the following formula:

[0056]

[0057] in, This is the LQR feedback gain matrix. The actual state vector, This is the ideal state vector.

[0058] It should be noted that by setting the performance index function as an integral expression of a quadratic function, the selected control system state variables and the control law to be solved are linearly related. By adjusting the state feedback, an optimal closed-loop control is achieved. The overall concept of the LQR controller in this invention is as follows:

[0059] (2.3)

[0060] (2.4)

[0061] In the formula, For control vectors, For state variables, For terminal weighting matrix, The weighted matrix of the state vectors. This is the weighting matrix for the control vectors.

[0062] The optimal feedback coefficient matrix is ​​K, and U = -KX. We can obtain K = Q²⁻¹BTP, where the solution to the Riccati equation is expressed by P, and its expression is:

[0063] (2.5)

[0064] Through the cost function and discrete solution, the steps for solving the LQR feedback gain are as follows:

[0065] Step 1: Initialization

[0066] 1) Order .

[0067] 2) Gain error between adjacent steps and steady-state threshold These two items are used to determine whether the system is stable.

[0068] 3) Initial feedback gain

[0069] 4) Maximum number of iterations This is used to limit the number of iterations.

[0070] Step 2 Iteration .

[0071] 1) Solve for the feedback gain:

[0072] (2.6)

[0073] 2) Solve :

[0074] (2.7)

[0075] 3) Solve for the feedback gain error of adjacent steps. :

[0076] (2.8)

[0077] 4) Iteration termination check;

[0078] when The iteration ends when the maximum number of iterations is reached. An error is reported when the number of iterations exceeds the maximum number of iterations.

[0079] An active front wheel steering system based on the LQR algorithm includes:

[0080] Kalman filter observer module is used to estimate the vehicle's center of gravity sideslip angle;

[0081] The state-space modeling module is used to build vehicle dynamics models;

[0082] The LQR control module is used to calculate the optimal front wheel steering angle compensation.

[0083] The front wheel steering angle superposition module is used to output the final steering control signal.

[0084] The inputs to the Kalman filter observer module include the vehicle's yaw rate, longitudinal speed, front wheel steering angle, and lateral acceleration.

[0085] The LQR control module uses a discrete iterative algorithm to solve for the feedback gain and sets a steady-state threshold and a maximum number of iterations.

[0086] A vehicle including the system described above.

[0087] A computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, implement the steps of the method.

[0088] Compared with the prior art, the technical solution disclosed in this application has the following non-obvious technical features:

[0089] First, this application combines LQR optimal control theory with active front wheel steering system to construct a quadratic performance index with yaw rate and center of gravity sideslip angle as state variables, so as to achieve steering control while minimizing control energy consumption.

[0090] Second, this application uses the Kalman filter algorithm to estimate the centroid sideslip angle in real time, which solves the problem that this key state variable is difficult to measure directly, and provides high-precision state input for the LQR controller;

[0091] Third, this application proposes an LQR gain solution method based on discrete iteration, which sets a steady-state threshold and a maximum number of iterations to control the computational load while ensuring the solution accuracy, thus adapting to the real-time requirements of vehicle ECUs.

[0092] Fourth, a two-layer control structure of "ideal reference model + state feedback correction" is constructed. By comparing the deviation between the actual and ideal state quantities, an additional front wheel steering angle compensation signal is dynamically generated.

[0093] Fifth, the coupling relationship between yaw rate and centroid sideslip angle is explicitly introduced in the controller design, and the coordinated control of the two under dynamic conditions is achieved through state-space model and weighted matrix adjustment.

[0094] Compared with the prior art, the present invention has the following beneficial effects:

[0095] 1. Significantly improves vehicle handling stability: Under complex conditions such as transient steering, low-adhesion road surfaces, and crosswind disturbances, it can maintain yaw rate and center of gravity sideslip angle that strictly follow the ideal values ​​with minimal deviation;

[0096] 2. Enhanced system anti-interference capability: Through LQR optimal adjustment and Kalman filter observation, the system has strong robustness to external disturbances and can effectively suppress the impact of crosswinds, uneven road surfaces, and other disturbances.

[0097] 3. Achieving the optimal balance between control energy consumption and performance: The LQR controller optimizes the performance index function to minimize control input energy while ensuring tracking accuracy, thereby improving system energy efficiency;

[0098] 4. Improved trajectory tracking accuracy: Under dynamic conditions such as continuous sinusoidal motion, the vehicle can accurately track the ideal path, reduce steering overshoot and oscillation, and improve driving comfort and safety;

[0099] 5. Strong real-time performance and low computational load: The Kalman filter and LQR solution algorithm adopted has a simple structure, which is suitable for embedded system deployment and can achieve real-time control without the need for high-performance computing units;

[0100] 6. Applicable to intelligent driving systems: It can provide high-precision and high-stability underlying steering execution support for advanced driving assistance functions such as autonomous driving, lane keeping, and emergency obstacle avoidance. Attached Figure Description

[0101] Figure 1 This is a flowchart of the method described in Embodiment 1 of the present invention;

[0102] Figure 2 This is a schematic diagram of the Kalman filter observer described in Embodiment 1 of the present invention;

[0103] Figure 3 This is a diagram of the overall control framework of the active front wheel steering system described in Embodiment 1 of the present invention;

[0104] Figure 4 This is a comparison of the yaw rate response described in Embodiment 1 of the present invention;

[0105] Figure 5 This is a comparison of the centroid sideslip angle response described in Embodiment 1 of the present invention;

[0106] Figure 6 This is a structural diagram of the active front wheel steering system described in Embodiment 1 of the present invention;

[0107] Figure 7 This is a schematic diagram of the system described in Embodiment 2 of the present invention. Detailed Implementation

[0108] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings:

[0109] Example 1:

[0110] A control method for an active front wheel steering system based on the LQR algorithm, such as Figures 1 to 5 As shown, its implementation is as follows:

[0111] First, accurate real-time values ​​of vehicle response parameters are obtained by a centroid sideslip angle observer installed on the vehicle, providing compensation for the active front wheel steering controller;

[0112] Secondly, the ideal yaw rate and the ideal center of gravity sideslip angle are input into the active front wheel steering system controller based on the LQR algorithm;

[0113] Then, the active front wheel steering system controller based on the LQR algorithm compares the real-time parameters of the vehicle with the ideal vehicle center of gravity sideslip angle and yaw rate.

[0114] Finally, when the vehicle is in a steady state, the actual yaw rate and sideslip angle are not significantly different from the ideal values, and the vehicle can then drive normally.

[0115] It is important to note that when the real-time vehicle parameters deviate significantly from the ideal vehicle parameters (i.e., sideslip angle and yaw rate), the active front wheel steering system controller will intervene, and the control algorithm will adjust the calculated compensating front wheel steering angle. This output is provided to the vehicle and, when superimposed on the actual front wheel steering angle, becomes the vehicle's current front wheel steering angle. This improves vehicle steering and enhances vehicle handling stability.

[0116] A continuous sinusoidal typical operating condition test was used to evaluate the vehicle's dynamic performance under external crosswind interference. The simulated vehicle speed was 80 km / h, the road adhesion coefficient was 0.8, and the external crosswind speed was 12 m / s. The front wheel steering angle input under the continuous sinusoidal test is shown in the figure below. Specific simulation results can be found in [the figure below]. Figure 4 and Figure 5 .

[0117] Depend on Figure 4 It can be seen that under continuous sinusoidal conditions, the constructed active front wheel steering system controller can track the ideal value of yaw rate very well. Compared with the uncontrolled case, the vehicle with LQR control has better trajectory tracking performance, and the vehicle's yaw rate can always follow the ideal reference value with almost no error. The designed LQR controller can improve the vehicle's steering performance and has good resistance to crosswind interference.

[0118] Depend on Figure 5 It can be seen that the constructed active front wheel steering system controller can track the ideal value of the center of gravity sideslip angle very well, and the vehicle body posture is stable when driving. Compared with the uncontrolled vehicle, the trajectory tracking effect of the vehicle with LQR control is better, and it can drive along the expected path. Moreover, the actual dynamic response parameter of the vehicle, i.e., the center of gravity sideslip angle, can always follow the ideal reference value with almost no error, thus achieving the goal of stable vehicle operation under the set conditions.

[0119] Example 2:

[0120] An active front-wheel steering system based on the LQR algorithm, such as Figure 7 As shown, this system is integrated into the vehicle's electronic control unit (ECU) and works in conjunction with the vehicle's sensor network and actuators. The system is a complete hardware and software integration, with its core control flow following the method described in Example 1, while the system's structure corresponds to... Figure 3 The structure shown; it includes

[0121] Main controller: A 32-bit microcontroller that meets automotive-grade standards (such as the Infineon TC3xx series) is used as the core of the system's computing.

[0122] Sensor Input Interface: The system receives signals from the following sensors in real time via the vehicle's CAN bus and a dedicated analog / digital interface:

[0123] Yaw rate sensor: provides the angular velocity signal of the vehicle's rotation about its vertical axis. ;

[0124] Longitudinal speed sensor: Typically takes the wheel speed signal from the anti-lock braking system (ABS) to calculate the vehicle speed. ;

[0125] Steering wheel angle sensor: measures the steering wheel angle input by the driver and converts it into the front wheel angle after passing through the steering gear ratio. ;

[0126] Lateral acceleration sensor: provides lateral acceleration at the vehicle's center of gravity. ;

[0127] Actuator output interface: The system controls an electronic power steering motor or a superimposed steering motor via PWM (Pulse Width Modulation) signals or CAN commands to apply a calculated additional front wheel steering angle. .

[0128] The system software runs on the ECU and adopts a modular design, mainly including the following four core modules:

[0129] (1) Kalman filter observer module: This module uses the vehicle's yaw rate as the basis for its observation. Longitudinal speed Driver's front wheel steering angle and lateral acceleration As input, the module embeds the aforementioned Kalman filtering algorithm, which maintains a state vector containing the centroid sideslip angle to be estimated. For other relevant states, the algorithm performs recursive calculations according to equations (1.1) to (1.5), where the system parameter matrices A and B are determined online or obtained by looking up tables from vehicle parameters (mass, wheelbase, tire lateral stiffness, etc.); the process noise covariance matrix... and observation noise covariance Through experimental calibration, the module ultimately outputs a high-precision estimate of the centroid sideslip angle. ;

[0130] (2) State-space modeling module: This module is based on the current vehicle speed Based on the vehicle's inherent parameters, the module constructs or calls a pre-stored two-degree-of-freedom vehicle model online; the module calculates the coefficient matrix of the state-space equations in real time according to equations (2.1) and (2.2). It outputs the Kalman filter observer module's... With sensor input Combined into actual state vector Meanwhile, this module uses the same vehicle model and current vehicle speed. and driver's front wheel steering angle Calculate the ideal reference state vector , as the tracking target for control;

[0131] (3) LQR control module: This module is the "decision center" of the system, responsible for calculating the optimal additional steering angle. Its working process is as follows:

[0132] Gain Calculation: A discrete iterative algorithm is adopted. When the system is initialized or the vehicle speed changes significantly, the module will call the LQR solver based on the coefficient matrix (A, B) of the current state-space model. The solver is based on the Riccati equation (2.5) and iterates according to S1 and S2 as described in the manual to quickly calculate the optimal state feedback gain matrix under the current operating condition. ;

[0133] Error Calculation and Control Output: The module calculates the state error in real time. Then, according to the formula Calculate the additional front wheel steering angle compensation amount that minimizes the performance index function (Equation 2.3). Weighted matrix and Offline optimization was performed to ensure the best balance between yaw rate tracking accuracy and centroid sideslip angle stability, while limiting control energy.

[0134] (4) Front wheel steering angle superposition module: This module is the synthesis and output terminal of control commands. The module receives commands from the LQR control module. The driver requests the steering angle from the steering wheel angle sensor. It superimposes the two: Final total front wheel steering angle command The signal is sent to the steering actuator (such as an overlay steering motor) to drive the wheels to rotate, thereby achieving active steering control.

[0135] The specific workflow of the system in this embodiment is as follows:

[0136] After the vehicle is powered on, all system modules initialize. During driving, sensor data is sent to the system at fixed intervals (e.g., 10ms), and the Kalman filter observer module continuously estimates... The state-space modeling module synchronously updates the actual and ideal states; the LQR control module judges the state error, and if the error exceeds the preset quiet threshold (e.g., yaw rate error > 0.05 rad / s), it immediately calculates and outputs the result. The front wheel steering angle superposition module completes the steering angle synthesis and drives the actuator. The entire process forms a high-speed, closed-loop "perception-decision-execution" circuit, ensuring the stability of the vehicle under various dynamic conditions.

[0137] Example 3:

[0138] A vehicle equipped with the active front-wheel steering system based on the LQR algorithm described in Embodiment 2, the vehicle being a passenger car, including a body, chassis, powertrain, conventional steering mechanism, and electronic and electrical architecture; its core feature lies in the integration of the active front-wheel steering system described in Embodiment 2 into the chassis control system, the integration and arrangement of which are as follows:

[0139] Controller deployment: The software algorithm of the active front wheel steering system is written into the vehicle stability control domain controller or a dedicated steering control ECU, which is located in the electronic control box at the front of the vehicle's cockpit;

[0140] Sensor integration: The vehicle comes standard with all the sensors required for the system, including:

[0141] The yaw rate and lateral acceleration sensors are integrated into the inertial measurement unit (IMU) and mounted on the vehicle floor near the vehicle's center of gravity.

[0142] Longitudinal vehicle speed information is provided by wheel speed sensors through the ABS system;

[0143] The steering wheel angle sensor is integrated inside the steering column assembly;

[0144] Actuator Integration: This vehicle employs a variant of the dual pinion electric power steering (DP-EPS) system, in which the steering assist motor not only provides assistance but also outputs an additional steering angle upon command from the controller. This motor is connected to the second pinion of the steering rack via a reduction mechanism, thereby achieving active superposition of the front wheel steering angle.

[0145] The vehicle operates in the following modes:

[0146] When the vehicle is in normal operation, the system is in a state of readiness.

[0147] Normal driving: When cruising on a smooth highway, the state error is small, and the system output is... With a near-zero sensitivity, the driver gains a direct feel indistinguishable from traditional steering;

[0148] Emergency Obstacle Avoidance: When the driver sharply turns the steering wheel to perform an emergency obstacle avoidance maneuver, the system detects that the yaw rate response may be too rapid, and there is a risk of an increased sideslip angle. The LQR controller will quickly calculate an angle that is the same as the driver's steering direction but slightly smaller. The superimposed layer helps to suppress oversteer and allows the vehicle to return to its original lane more smoothly.

[0149] Cornering on low-adhesion surfaces: When cornering on icy or snowy surfaces, the system uses Kalman filtering to accurately detect when the vehicle's center of gravity sideslip angle is close to its limit. The LQR controller outputs a small compensation angle to adjust the tire lateral force, ensuring that the actual center of gravity sideslip angle and yaw rate closely follow the ideal value with reduced amplitude, thereby preventing the vehicle from fishtailing.

[0150] Crosswind compensation: When encountering strong crosswinds, the IMU detects yaw movements not intended for driving. The system treats this as a disturbance, and the LQR controller calculates a reverse yaw rate. This drives the vehicle to generate a yaw moment that resists crosswinds, enabling the vehicle to maintain a straight line without requiring the driver to frequently correct the steering wheel.

[0151] Through the above integration, compared with the same model without this system, this vehicle exhibits less trajectory deviation, faster transient response convergence speed, and less steering correction burden on the driver in handling stability tests (such as the moose test and the sine stop test), significantly improving active safety.

[0152] Example 4:

[0153] A computer-readable storage medium stores program instructions executable by a processor of a vehicle ECU. When executed, these program instructions fully implement the active front-wheel steering system control method based on the LQR algorithm described in Embodiment 1. The physical form of the computer-readable storage medium is a flash memory chip conforming to automotive-grade standards, such as NOR Flash. It is soldered onto the motherboard of the vehicle steering control ECU. Alternatively, the program instructions can also be stored in the internal mask ROM or OTP memory of a microcontroller.

[0154] After the vehicle is started, the ECU begins execution, and the processor reads instructions from a fixed address in the storage medium to start execution. The system first performs initialization, including clearing variables and loading calibration parameters (such as vehicle parameters, ...). matrix, (Matrix); then, the timer interrupt is started and repeatedly triggered with a period of 10ms; the processor continuously completes the closed-loop control process of "sensor data acquisition -> state estimation -> reference model calculation -> optimal control decision -> actuator drive" by executing these stored instructions, thereby realizing the control method described in this invention at the physical level, enabling the vehicle to have excellent active steering stability control function.

[0155] The above technical solutions only embody the preferred technical solutions of the present invention. Any modifications that may be made by those skilled in the art to certain parts thereof embody the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. A control method for an active front wheel steering system based on the LQR algorithm, characterized in that, Includes the following steps: S1. Estimate the vehicle's center of gravity sideslip angle in real time using a Kalman filter observer; S2. Construct state-space equations based on a two-degree-of-freedom vehicle dynamics model; S3. Design an LQR controller and obtain the optimal state feedback gain by solving the Riccati equation; S4. Calculate the additional front wheel steering angle compensation based on the deviation between the actual and ideal yaw rate and the center of gravity sideslip angle. S5. The compensation amount is superimposed on the front wheel steering angle input by the driver and output to the steering actuator.

2. The method according to claim 1, characterized in that, The Kalman filter observer in S1 includes two parts: time update and measurement update. The Kalman filter observer is used to estimate the minimum variance of the centroid sideslip angle.

3. The method according to claim 1, characterized in that, The state variables in the state-space equation of S2 include: yaw rate and center of mass sideslip angle, and the input is the front wheel steering angle.

4. The method according to claim 1, characterized in that, The performance index function of the LQR controller in S3 is a quadratic integral form, and optimal control is achieved by adjusting the state weighting matrix and the control weighting matrix.

5. The method according to claim 1, characterized in that, The additional front wheel steering angle compensation in S4 is calculated using the following formula: in, This is the LQR feedback gain matrix. The actual state vector, This is the ideal state vector.

6. An active front wheel steering system based on the LQR algorithm, characterized in that, include: Kalman filter observer module is used to estimate the vehicle's center of gravity sideslip angle; The state-space modeling module is used to build vehicle dynamics models; The LQR control module is used to calculate the optimal front wheel steering angle compensation. The front wheel steering angle superposition module is used to output the final steering control signal.

7. The system according to claim 6, characterized in that, The inputs to the Kalman filter observer module include the vehicle's yaw rate, longitudinal speed, front wheel steering angle, and lateral acceleration.

8. The system according to claim 6, characterized in that, The LQR control module uses a discrete iterative algorithm to solve for the feedback gain and sets a steady-state threshold and a maximum number of iterations.

9. A vehicle, characterized in that, The system includes any one of claims 6 to 8.

10. A computer-readable storage medium having program instructions stored thereon, characterized in that, When the program instructions are executed by the processor, they implement the steps of the method as described in any one of claims 1 to 5.

Images