New energy vehicle driving stability active control method based on adaptive optimization MPC algorithm

Abstract

The present application relates to the technical field of automobile active safety control, and particularly to a new energy automobile driving stability active control method based on adaptive optimization MPC algorithm, which comprises the following steps: S1: high-precision assembly and initial calibration are performed on a vehicle chassis air suspension system to establish an accurate hardware basis for model predictive control; S2: calibrated vehicle state information is acquired in real time; S3: a model predictive controller is constructed based on an eight-degree-of-freedom vehicle dynamics model, and a parallel chaos optimization algorithm is used to perform rolling optimization solution on the objective function of the controller to generate optimal control instructions for stability control; and S4: the optimal control instructions are executed, and vehicle states are fed back to form a closed-loop control. Through the innovative multi-wheel position high-precision assembly and calibration method, the consistency of the chassis hardware system is fundamentally improved, and an accurate physical basis is provided for high-order control algorithms.

Description

Technical Field

[0001] This invention relates to the field of active safety control technology for automobiles, specifically to an active control method for driving stability of new energy vehicles based on an adaptive optimization MPC algorithm. Background Technology

[0002] Current electric vehicle stability control largely relies on Electronic Stability Program (ESP). However, traditional ESP has significant tracking errors in reference trajectories under extreme conditions (even with high tracking accuracy, fluctuations in the vehicle's stability performance during assembly can affect the actual mass production results of the developed vehicles). It struggles to balance vehicle yaw stability and driving comfort, considering the actual conditions of different models within the same production batch. While Model Predictive Control (MPC) algorithms can achieve high-precision control through a "prediction-optimization-feedback" mechanism, conventional MPC has low effectiveness in solving its objective function and insufficient adaptability to the multi-degree-of-freedom dynamics of vehicles (due to issues such as uneven load distribution caused by chassis assembly precision, especially air suspension assembly, deviations between sensor information and actual conditions, and vehicle posture and driving status). This limits its application in distributed drive electric heavy-duty trucks (most heavy-duty trucks use electric drive axles for the middle and rear axles), resulting in low effectiveness of the calculated theoretical solutions. Summary of the Invention

[0003] To address the problems in existing technologies, this invention provides an active control method for the driving stability of new energy vehicles based on an adaptive optimization MPC algorithm.

[0004] The technical solution adopted by this invention to solve its technical problem is: an active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm, characterized by including the following steps:

[0005] S1: Perform high-precision assembly and initial calibration of the vehicle chassis air suspension system to establish a precise hardware foundation for model predictive control;

[0006] S2: Real-time acquisition of calibrated vehicle status information;

[0007] S3: A model predictive controller is constructed based on an eight-degree-of-freedom vehicle dynamics model, and a parallel chaotic optimization algorithm is used to perform rolling optimization of the controller's objective function to generate the optimal control command for stability control.

[0008] S4: Execute the optimal control command and provide feedback on the vehicle status to form a closed-loop control.

[0009] Preferably, the high-precision assembly and initial calibration in step S1 includes using a multi-wheel positioning assembly method to correct the wheel positions of all axles so that the geometric shape formed by connecting the center points of each wheel approaches a preset rectangular frame.

[0010] Preferably, the multi-wheel positioning assembly method specifically involves: establishing a rectangular theoretical frame based on the center point of the vehicle's front wheel; measuring the center point positions of each wheel on the middle and rear axles, and calculating their parallelism deviation and propulsion line deviation relative to the rectangular theoretical frame;

[0011] Depending on the direction and magnitude of the deviation, it is corrected by adding an adjusting shim of a specific thickness at the connection between the corresponding thrust rod and the balance shaft bracket.

[0012] Preferably, the specific rules for correcting the parallelism deviation and the propulsion line deviation include: if the propulsion line deviation of the middle axle is positive, a shim is added at the thrust rod connection on its left side; if it is negative, a shim is added at the thrust rod connection on its right side; the correction of the rear axle is based on the sum of the propulsion line deviation of the middle axle and the parallelism deviation of the rear axle itself, with a shim added on the right side for positive values ​​and on the left side for negative values.

[0013] Preferably, the high-precision assembly and initial calibration in step S1 includes applying controllable downward pressure to the vehicle's load-bearing crossbeam using a suspension system calibration compression tool to overcome the inherent damping of the chassis system, so that the air suspension system accurately reaches and stabilizes at the preset initial calibration height.

[0014] Preferably, the high-precision assembly and initial calibration in step S1 includes using a split jack device to independently and synchronously support adjacent axles when repairing and replacing air springs, maintaining the relative position and posture of adjacent axles when lifting the frame, thereby ensuring that the parallelism of the guide spring installation and the coaxiality of the lugs on the axle being repaired meet the assembly requirements.

[0015] Preferably, in step S3, the objective function of the model prediction controller is configured to simultaneously optimize vehicle trajectory tracking accuracy, control input quantity, and control input change rate; the parallel chaotic optimization algorithm expands the search range and improves the solution efficiency through chaotic variable mapping iteration.

[0016] Preferably, the rolling optimization solution process in step S3 needs to meet the constraints determined by the vehicle dynamics characteristics. The constraints include limits on the center of gravity sideslip angle, yaw rate, wheel slip ratio, and drive motor output torque.

[0017] Preferably, the feedback correction model at the next time step Collect actual output , and the predicted output Compare and correct the model parameters, then repeat the S1 to S4 process.

[0018] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements an active control method for driving stability of new energy vehicles based on an adaptive optimization MPC algorithm.

[0019] The beneficial effects of this invention are:

[0020] (1) The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm described in this invention fundamentally improves the consistency of the chassis hardware system through innovative multi-wheel high-precision assembly and calibration methods, providing a precise physical basis for high-order control algorithms. A laser measurement and shim correction method based on the "rectangular theoretical framework" is proposed, which improves the parallelism control accuracy of the middle / rear axle and the front axle of multi-axle heavy trucks from the industry common ±10mm to within ±3mm; at the same time, the controllable downward pressure tool is used to overcome chassis damping for calibration, ensuring the consistency of the initial state of the air suspension system. This effectively solves the industry problems such as uneven suspension load distribution and sensor input distortion caused by assembly errors and chassis deformation, significantly reduces vehicle deviation and tire wear, and makes the vehicle state information on which the model predictive control (MPC) algorithm depends more realistic and reliable.

[0021] (2) The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm described in this invention optimizes vehicle network resources and ensures the accuracy of control timing while ensuring control accuracy through an adaptive collaborative processing mechanism. Specifically, it reduces invalid data transmission by dynamically adjusting the sensor compensation cycle; it reduces the amount of data by about 60% by extracting and compressing features from the original data, and ensures the real-time performance of key parameters by combining priority coding; and it strictly aligns the compensation results with the MPC rolling optimization cycle through hardware clock synchronization. This collaborative scheme reduces the vehicle CAN bus occupancy rate from 30% in the traditional method to below 27%, avoids data packet loss, and ensures that the MPC controller can obtain high-quality information under the correct timing. As a result, under dynamic conditions such as double lane change, it reduces suspension travel fluctuation by 45% and controls the roll angle acceleration deviation within ±0.5 rad / s², achieving a balance between driving stability and the real-time performance of the control system. Attached Figure Description

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] Figure 1 A flowchart of the steps provided for this invention;

[0024] Figure 2 The overall control closed-loop flowchart provided by this invention;

[0025] Figure 3 The multi-round positioning flowchart provided by this invention;

[0026] Figure 4 The MPC controller construction and optimization process provided by this invention;

[0027] Figure 5 The feedback correction iteration flowchart provided for this invention. Detailed Implementation

[0028] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0029] It should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0030] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0031] like Figures 1-5 As shown, the active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm of the present invention includes the following steps:

[0032] S1: Perform high-precision assembly and initial calibration of the vehicle chassis air suspension system to establish a precise hardware foundation for model predictive control;

[0033] High-precision assembly and initial calibration of the vehicle chassis air suspension system are performed to establish a precise hardware foundation for model predictive control. The high-precision assembly and initial calibration include multi-wheel alignment assembly, air suspension calibration, and maintenance assembly support, which are compatible with 6×4, 8×4, 10×6 and other multi-wheel distributed drive new energy heavy trucks and special vehicles.

[0034] S2: Real-time acquisition of calibrated vehicle status information;

[0035] The vehicle status information after calibration is acquired in real time. The status information includes vehicle speed, center of gravity sideslip angle, yaw rate, suspension dynamic travel and vehicle attitude parameters. During the acquisition process, adaptive collaborative processing is performed on the sensor data, including dynamic compensation cycle, data compression and priority encoding, and compensation control timing alignment.

[0036] The adaptive compensation cycle adjustment involves setting a dynamic threshold for the compensation cycle based on the dynamic characteristics of the air suspension system (such as the rate of change of vehicle height and the degree of road bumps). When the suspension is in a steady state (e.g., on a flat road with a height change rate ≤ 0.5 mm / s), the compensation cycle is extended to 50 ms; when in dynamic conditions (e.g., on a bumpy road with a height change rate > 0.5 mm / s), it is shortened to 10 ms to reduce invalid data transmission.

[0037] Data compression and priority coding: Feature extraction is performed on the raw sensor data (such as multi-digit acceleration values) to retain key features that are strongly related to MPC control (such as peak value and mean value). The amount of data after compression is reduced by 60%. At the same time, priority coding is used to mark core parameters such as suspension travel and yaw rate as the highest priority and give them priority to occupy bus bandwidth.

[0038] Compensation-control timing alignment: By synchronizing the hardware clock (accuracy ±1ms), the dynamic compensation results of the sensor are strictly aligned with the rolling optimization cycle of the MPC algorithm (e.g., 20ms), avoiding control deviations caused by compensation delays.

[0039] Technical Effects - CAN Bus Resource Optimization: Bus occupancy rate has been reduced from 30% in the traditional algorithm to below 27% (referring to the CAN bus occupancy rate of the entire vehicle), avoiding data packet loss caused by bus congestion.

[0040] Improved vehicle performance and stability: Under conditions such as double lane change and rapid deceleration, the fluctuation range of suspension dynamic deflection is reduced by 45%, and the maximum deviation of roll angle acceleration is controlled within ±0.5rad / s², ensuring the MPC algorithm's precise control of the air suspension system.

[0041] The difference from existing technologies - traditional technologies only focus on the compensation accuracy of sensor data, without considering bus resource constraints and control timing matching; this invention, through the collaborative design of "dynamic cycle + data compression + timing alignment", solves the contradiction between excessive bus resource occupation and vehicle performance fluctuation while ensuring compensation accuracy. It is suitable for high-precision control scenarios of distributed drive electric new energy heavy truck air suspension systems (with two or more drive axles in the middle and rear axles).

[0042] S3: A model predictive controller is constructed based on an eight-degree-of-freedom vehicle dynamics model, and a parallel chaotic optimization algorithm is used to perform rolling optimization of the controller's objective function to generate the optimal control command for stability control.

[0043] A model predictive controller is constructed based on an eight-degree-of-freedom vehicle dynamics model. The model includes longitudinal, lateral, yaw motion, and four-wheel rotation degrees of freedom. The influence of front wheel longitudinal force, lateral force, steering angle, and direct yaw torque on vehicle motion is characterized by dynamic equations. The objective function of the controller is solved by a parallel chaotic optimization algorithm (PCOA) to generate the optimal control command for stability control. The optimal control command includes the front wheel steering angle adjustment and the four-wheel drive torque distribution value.

[0044] Objective function (existing technology): comprehensively optimizes trajectory tracking accuracy, control input smoothness, and energy consumption; its expression is:

[0045] in, For the actual state (β, γ), For reference trajectory, For control inputs (front wheel steering angle, four-wheel torque). To control the rate of change of the input, Q, R, and A are weight matrices;

[0046] S4: Execute the optimal control command and provide feedback on the vehicle status to form a closed-loop control.

[0047] The vehicle's current state, obtained by S2, is compared with the vehicle's real-time state, including speed, obtained by the sensors. yaw rate Inclination angle etc., denoted as ( (At the current moment).

[0048] Model predicts future output - Substitute into the vehicle dynamics model (e.g., considering tire forces, suspension characteristics), combine with current control variables (e.g., braking torque), and predict the next N steps (e.g., ...). +1, k+2, ..., Output at time +N) ( =1,2,...,N);

[0049] Rolling optimization control sequence - Define optimization objective (e.g., minimize) , Using the reference trajectory and constraints (such as braking force not exceeding actuator limits), solve for the optimal control sequence. Take the first control quantity It applies to vehicles.

[0050] The optimal control command is executed, and the vehicle status is fed back to form a closed-loop control. During the execution process, the control command is sent to the actuator via the CAN bus. During the feedback process, the actual state is compared with the predicted state in real time, and the model parameters are dynamically corrected.

[0051] An eight-DOF vehicle dynamics model is established, including longitudinal, lateral, yaw, and four-wheel rotational degrees of freedom. Its core equations are as follows:

[0052] Longitudinal dynamic equation:

[0053] Lateral dynamic equations: sin

[0054] Equation of yaw motion:

[0055] Where m is the total vehicle mass. These are longitudinal and lateral accelerations, respectively. , For the longitudinal and lateral forces of the front wheels, This refers to the front wheel steering angle. Let γ be the moment of inertia about the center of mass, and γ be the yaw rate. This is the direct yaw torque.

[0056] As a preferred technical solution, the high-precision assembly and initial calibration in step S1 includes using a multi-wheel positioning assembly method to correct the wheel positions of all axles so that the geometric shape formed by connecting the center points of each wheel approaches the preset rectangular frame.

[0057] Based on the "rectangular parameter theory", the wheel positions of all axles are corrected. A rectangular theoretical framework is established with the center point of the front wheel as the reference, so that the geometric shape formed by connecting the center points (non-contact points) of each wheel is close to the preset rectangular framework, which is suitable for the chassis structure of multi-wheel models.

[0058] It abandons the limitation of traditional four-wheel alignment being only applicable to passenger cars and solves the problem of multi-wheel alignment for new energy heavy-duty trucks; by constructing a rectangular frame by the center point of the wheel, it avoids the impact of frame deformation on positioning accuracy when half-loaded / fully loaded, so that the parallelism between the middle and rear wheels and the front wheels is controlled within ±3mm, effectively reducing phenomena such as driving deviation and tire wear, and improving the consistency of the air suspension system.

[0059] As a preferred technical solution, the multi-wheel positioning assembly method specifically involves: establishing a rectangular theoretical frame based on the center point of the vehicle's front wheel; measuring the center point positions of each wheel on the middle and rear axles, and calculating their parallelism deviation and propulsion line deviation relative to the rectangular theoretical frame;

[0060] Depending on the direction and magnitude of the deviation, it is corrected by adding an adjusting shim of a specific thickness at the connection between the corresponding thrust rod and the balance shaft bracket.

[0061] A rectangular theoretical framework is established with the center point of the front wheel of the vehicle as the reference. The center point positions of each wheel of the middle and rear axles are measured by a laser camera (measurement accuracy ±0.1mm), and the parallelism deviation and thrust line deviation relative to the rectangular theoretical framework are calculated. According to the direction and magnitude of the deviation, the correction is made by adding 1mm, 3mm and 5mm adjusting shims at the connection between the corresponding thrust rod and the balance shaft bracket. No more than two adjusting shims are allowed to be added at the same connection.

[0062] High-precision deviation measurement is achieved using laser cameras, and accurate correction is achieved through standardized adjustment shims. The operation is simple and highly repeatable. It effectively corrects deviations amplified by the assembly process, significantly reducing the consistency fluctuation of the suspension of vehicles in the same batch, providing a precise hardware foundation for the MPC algorithm, and reducing model prediction bias.

[0063] As a preferred technical solution, the specific rules for correcting the parallelism deviation and the propulsion line deviation include: if the propulsion line deviation of the middle axle is positive, a shim is added at the thrust rod connection on its left side; if it is negative, a shim is added at the thrust rod connection on its right side; the correction of the rear axle is based on the sum of the propulsion line deviation of the middle axle and the parallelism deviation of the rear axle itself, with a shim added on the right side for positive values ​​and on the left side for negative values.

[0064] Define differentiated correction rules to address the parallelism and propulsion line deviation issues of the middle and rear axles, ensuring that the parallelism between the rear axle and the front and middle axles is controlled within ±3mm; avoid the phenomenon of the drive axle pushing the vehicle, reduce tire wear and power loss, and improve the vehicle's straight-line driving stability and tire life.

[0065] The parallel alignment method for the middle and rear wheels solves the industry problem of large parallelism between the middle and rear wheels (here, the middle and rear wheels refer to the wheels mounted on the middle and rear axles, which are called the middle and rear axles in the industry) and the front wheels of new energy heavy trucks (in the industry, the parallelism between the middle axle and the front axle is within ±10mm).

[0066] By correcting parallelism deviation and propulsion line deviation, parallelism can be reduced to within ±3mm. This significantly solves the industry-wide problem of vehicle deviation, uneven wear, and tire wear caused by uneven load distribution in the suspension system after assembly, which leads to errors in data input from height and angle sensors. There's an industry curse that fully loaded heavy trucks inevitably veer off course and 8x4 trucks inevitably suffer tire wear. Truck drivers in the industry generally accept this phenomenon (a fully loaded heavy truck weighs 49 tons; this load distribution on the chassis load-bearing system causes changes in load-bearing capacity due to large assembly precision errors in the suspension system. The air spring load distribution is inevitably uneven, resulting in significant fluctuations in driving feedback data, affecting ECU calculation response time and accurate perception). This invention fills a technological gap in the industry.

[0067] In the traditional application of existing MPC technology in autonomous driving scenarios of intelligent vehicles, MPC algorithms can be used for vehicle trajectory tracking and control, helping vehicles to accurately travel along a planned path. The following are specific application examples:

[0068] In trajectory tracking control for autonomous driving in intelligent vehicles, the MPC algorithm, with its closed-loop mechanism of "prediction-optimization-feedback," can effectively address the impact of vehicle dynamics, road disturbances, and dynamic traffic environments. The following are more specific application examples:

[0069] Highway navigation assistance scenario

[0070] When a smart car is in highway navigation mode (such as following other vehicles or lane keeping), MPC needs to achieve centimeter-level trajectory accuracy to avoid deviating from the lane or rear-ending the vehicle in front.

[0071] Model construction: A five-degree-of-freedom vehicle model is adopted, which includes longitudinal (vehicle speed, acceleration) and lateral (lateral displacement, yaw angle). At the same time, a tire lateral slip characteristic model (such as Pacejka magic formula) is introduced to accurately reflect the tire force changes on road surfaces with different adhesion coefficients (such as wet and slippery road surfaces in rainy weather).

[0072] Reference trajectory generation: Based on lane line data from high-precision maps, a smooth reference path (including curvature and heading angle information) is generated, and the desired vehicle speed is dynamically adjusted in combination with the speed of the vehicle in front (such as maintaining a cruising speed of 100km / h or maintaining a safe distance of 2 seconds from the vehicle in front when following).

[0073] Optimization objectives: Prioritize minimizing lateral offset (60% weighting) and yaw rate deviation (20%), while limiting steering wheel angle change rate (10%) and longitudinal acceleration fluctuation (10%) to avoid passenger discomfort caused by sudden steering or rapid acceleration / deceleration.

[0074] Constraints: Strictly limit steering wheel angle (±450°), braking deceleration (≤-3m / s²), and accelerator pedal opening (≤80%) to ensure that control inputs are within the vehicle's physical limits.

[0075] Actual results: When driving on a highway curve (curvature radius 500m), the lateral deviation controlled by MPC can be kept within ±30cm, which is 40% more accurate than PID control; in the face of sudden situations (such as the vehicle in front decelerating suddenly), it can complete the deceleration response within 0.5 seconds and maintain the trajectory stability to avoid deviating from the lane.

[0076] Traffic scenarios at complex intersections in urban areas

[0077] In urban roads, sudden disturbances such as pedestrians crossing and non-motorized vehicles weaving through are frequent, so MPC needs to have the ability to make rapid adjustments while ensuring trajectory accuracy.

[0078] Model adaptation: Considering the low-speed (≤50km / h) characteristics of urban areas, the longitudinal dynamics model is simplified (air resistance is ignored), but the lateral response speed is enhanced (sampling period is shortened to 0.01 seconds) to cope with frequent turning requirements.

[0079] Reference trajectory dynamic update: Combining real-time perception data (such as LiDAR detecting a pedestrian intrusion), the path planning module will instantly generate detour actions (such as shifting 0.5m to the left to avoid the pedestrian), and MPC needs to complete the trajectory switch within 2 seconds.

[0080] Optimization logic adjustment: Temporarily increase the "obstacle avoidance priority" to allow lateral acceleration to exceed the comfort threshold in the short term (e.g., from ±1.5m / s² to ±2.0m / s²), but limit high-frequency steering wheel vibration (angle change rate ≤100° / s) through weight allocation.

[0081] Feedback correction enhancement: Using lane line recognition results from the vehicle camera (30Hz refresh rate) and real-time attitude data from the IMU, the model prediction bias (such as lateral slippage caused by road bumps) is corrected every 0.02 seconds.

[0082] Actual effect: At unprotected left-turn intersections, vehicles can accurately track the reference trajectory of curvature changes (lateral deviation ≤20cm during turning). When encountering pedestrians suddenly crossing, they can smoothly avoid them and return to the original trajectory within 1.5 seconds, avoiding dangers caused by sudden braking or sharp steering.

[0083] In these driving scenarios, the core value of MPC lies in: balancing trajectory accuracy and driving stability through multi-objective optimization, responding to environmental changes in real time via rolling time domain, and compensating for system errors using model feedback, ultimately achieving high reliability of autonomous driving under complex conditions. Note: The above information is from relevant academic literature, and the specific theoretical applications and conclusions are also from the literature descriptions. It can be concluded that in the autonomous driving scenarios of intelligent vehicles, the MPC algorithm has a positive effect on vehicle trajectory tracking control, helping the vehicle to accurately drive along the planned path.

[0084] As a preferred technical solution, the high-precision assembly and initial calibration in step S1 includes applying controllable downward pressure to the vehicle's load-bearing crossbeam using a suspension system calibration compression tool to overcome the inherent damping of the chassis system, so that the air suspension system can accurately reach and stabilize at the preset initial calibration height.

[0085] Using a suspension system calibration compression tool (pressure adjustment range 0-50kN), controllable downward pressure is applied to the vehicle's load-bearing crossbeam to overcome the inherent damping generated by components such as shock absorbers and thrust rods in the chassis system, so that the air suspension system can accurately reach and stabilize at a preset calibration initial height; the calibration initial height is based on the relative initial position of the air spring and the frame crossbeam, rather than the traditional ground clearance.

[0086] Abandoning the traditional ground clearance calibration method, it solves the calibration inaccuracy problem caused by internal damping of the vehicle frame; by controlling downforce, the air suspension is accurately and stably stabilized at the calibration position, ensuring that the dispersion of calibration parameters of the same batch of vehicles is greatly reduced, avoiding uneven load distribution under half load / full load, and improving the authenticity and reliability of the input data of the MPC algorithm.

[0087] As a preferred technical solution, the high-precision assembly and initial calibration in step S1 includes using a split jack device to independently and synchronously support adjacent axles when repairing and replacing air springs, maintaining the relative position and posture of adjacent axles when lifting the frame, thereby ensuring that the parallelism of the guide spring installation and the coaxiality of the lugs on the axle being repaired meet the assembly requirements.

[0088] The multi-objective optimization objective function balances stability and comfort, and the parallel chaotic algorithm improves the efficiency of solving the optimal solution of the objective function. Compared with the traditional extreme value search control (ESC) method, the side slip angle tracking error is reduced by more than 30%. The algorithm's lightweight design is adapted to the computing power of the vehicle ECU, avoids control delay, and ensures that the calculation is completed within a 10-50ms control cycle.

[0089] This addresses the industry challenge of large errors in the parallelism of the guide spring installation and the coaxiality of the coil lugs after replacing the air springs in a vehicle.

[0090] As a preferred technical solution, in step S3, the objective function of the model prediction controller is configured to simultaneously optimize the vehicle trajectory tracking accuracy, control input quantity, and control input change rate; the parallel chaotic optimization algorithm expands the search range and improves the solution efficiency through chaotic variable mapping iteration.

[0091] Clearly define physical constraint boundaries to avoid tire slippage, power system overload, or vehicle loss of control caused by exceeding hardware limits; ensure the safety and effectiveness of the control process so that the vehicle can maintain a stable state under extreme conditions and improve driving safety.

[0092] As a preferred technical solution, the rolling optimization solution process in step S3 needs to meet the constraints determined by the vehicle dynamics characteristics. The constraints include limits on the sideslip angle of the center of gravity, yaw rate, wheel slip ratio, and output torque of the drive motor.

[0093] Dynamically correcting control errors caused by model bias and external disturbances improves algorithm robustness; avoids the decrease in control accuracy caused by the mismatch between the model and actual hardware characteristics, and ensures that the vehicle can maintain stable control performance under complex working conditions such as load changes and road bumps.

[0094] Optimization Solution: The PCOA algorithm is used to perform rolling optimization on the objective function. Chaotic variable mapping is used to expand the search range and improve the efficiency of finding the optimal solution. The iterative formula is as follows: ;

[0095] Constraints (existing technology): Limit the centroid sideslip angle β ∈ yaw rate γ∈ Slip ratio K∈ The torque T ≤ 185 N·m [T ≤ 185 N·m] is set as a constraint condition in vehicle stability control. It is mainly based on the power system parameters and safe operation requirements of distributed drive electric vehicles. It comprehensively considers the wheel hub motor performance parameters (motor rated torque, maximum output capacity), tire adhesion limit and vehicle dynamic stability requirements of the vehicle model tested. It aims to avoid tire slippage, power system overload or vehicle loss of control due to excessive torque, thereby ensuring the safety and effectiveness of the control process. Because the safety parameter settings are reserved during the verification process, the corresponding values ​​can be set according to the actual situation of new energy heavy trucks to ensure the physical safety of the vehicle.

[0096] As a preferred technical solution, the feedback correction model – next moment Collect actual output , and the predicted output Compare and correct the model parameters, then repeat the S1 to S4 process.

[0097] In the autonomous driving scenario of intelligent vehicles, the following are specific application examples of existing MPC technology in traditional new energy vehicle suspension calibration systems:

[0098] The application of MPC technology in traditional new energy vehicle suspension calibration systems primarily focuses on improving ride comfort and handling stability. By establishing a vehicle dynamics model, setting reference trajectories and optimization targets, and performing rolling optimization and feedback correction under constraints, precise control of the suspension is achieved. The following are three application examples:

[0099] The paper "Model predictive control of a semi-active suspension with a shift delay compensation using preview road information" proposes a semi-active suspension control method based on model predictive control (MPC) to compensate for shift delay using preview road information. By establishing a vehicle dynamics model, the suspension state is estimated, and the feasible region of the control input is determined. Considering the high computational load of MPC and the "shift delay" problem that occurs when the vehicle's electronic control unit processes the algorithm, the suspension state is predicted using the vehicle dynamic model and road preview information for compensation. Computer simulation and real-vehicle testing show that this shift compensation algorithm significantly improves ride comfort.

[0100] Vertical Vibration Control of Hub Motor-Driven Electric Vehicles: In the study "Research on Vertical Vibration of Hub Motor-Driven Electric Vehicles Based on MPC," a seven-degree-of-freedom (DOF) half-vehicle model consisting of a hub motor and an air suspension system was constructed to address the problem of deteriorated vertical vibration caused by unbalanced electromagnetic forces in hub motor-driven electric vehicles. The accuracy of the model was verified through real-vehicle testing. A model predictive controller was established using the suspension dynamic travel and the damping force generated by the adjustable damper as constraints. Simulation results show that the MPC-based semi-active suspension effectively improves the problem of deteriorated vertical vibration. The power spectral density of various vehicle evaluation indicators significantly decreases within the human-sensitive frequency range, improving ride comfort.

[0101] Active control of electronically controlled air suspension systems: The paper "Research on Active Suspension Control of Electric Vehicles Based on Fractional-Order Air Spring Modeling" takes the electronically controlled air suspension system (ECAS) of a passenger vehicle as the research object. It optimizes the air spring model using fractional-order theory, establishes a 14-DOF vehicle ECAS dynamic model, and proposes the MPC active suspension control method. In simulations of rapid deceleration and double lane change conditions, as well as in vehicle bench tests, MPC control effectively reduces suspension dynamic deflection, pitch acceleration, and roll acceleration, showing better optimization results than PID control, thus improving vehicle ride comfort and stability.

[0102] Feedback correction: At the next sampling time, the actual state of the vehicle is obtained through sensors on the vehicle (such as GPS, inertial measurement unit, IMU, etc.), compared with the predicted state, and the model is corrected by feedback to correct the model parameters or prediction error. Then the above process is repeated to realize the real-time trajectory tracking control of the vehicle.

[0103] Application Results: Through trajectory tracking control using the MPC algorithm, vehicles can travel relatively accurately along a planned trajectory under various complex road conditions and interference. For example, in urban roads where other vehicles frequently change lanes, the MPC algorithm can quickly adjust the vehicle's speed and steering wheel angle to keep the vehicle stable within the designated lane, while avoiding collisions with surrounding vehicles, greatly improving the safety and comfort of autonomous driving.

[0104] In these ECAS (Electronic Stability and Compression) control system scenarios for air suspension, the core value of MPC (Multi-Purpose Control) lies in its application in traditional new energy vehicle suspension calibration systems, primarily focusing on improving ride comfort and handling stability. By establishing a vehicle dynamics model, setting a reference trajectory and optimization objectives, and performing rolling optimization and feedback correction under constrained conditions, precise control of the suspension is achieved. Note: The above information is from relevant academic literature, and the specific theoretical applications and conclusions are also derived from literature descriptions. In summary, the application in traditional new energy vehicle suspension calibration systems primarily focuses on improving ride comfort and handling stability. Achieving precise suspension control through establishing a vehicle dynamics model, setting a reference trajectory and optimization objectives, and performing rolling optimization and feedback correction under constrained conditions is a positive outcome.

[0105] In summary, the traditional application of existing MPC technology in the autonomous driving scenario of intelligent vehicles is as follows: MPC algorithms can be used for vehicle trajectory tracking and control, helping vehicles to accurately travel along planned paths. Below are specific application examples, including an application case study of existing MPC technology in the traditional suspension calibration system of new energy vehicles within the autonomous driving scenario of intelligent vehicles. The industry issues are analyzed below:

[0106] In air suspension systems, assembly precision (such as beam coaxiality and frame symmetry) and calibration accuracy (such as sensor parameters and actuator response thresholds) directly affect the effectiveness of the MPC algorithm's "prediction-optimization-feedback" closed loop.

[0107] Insufficient assembly precision can lead to a greater deviation between the vehicle dynamics model and reality, reducing the accuracy of MPC prediction (such as inaccurate calculation of suspension dynamic travel constraints).

[0108] Calibration deviations can weaken the feedback correction effect (e.g., errors in air spring stiffness parameters can cause an imbalance in the optimization target weights), ultimately affecting the MPC's ability to improve comfort (vertical vibration suppression) and stability (tilt / pitch control).

[0109] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the active control method for driving stability of new energy vehicles based on the adaptive optimization MPC algorithm as described in any one of claims 1 to 9.

[0110] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0111] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm, characterized in that: Includes the following steps: S1: Perform high-precision assembly and initial calibration of the vehicle chassis air suspension system to establish a precise hardware foundation for model predictive control; S2: Real-time acquisition of calibrated vehicle status information; S3: A model predictive controller is constructed based on an eight-degree-of-freedom vehicle dynamics model, and a parallel chaotic optimization algorithm is used to perform rolling optimization of the controller's objective function to generate the optimal control command for stability control. S4: Execute the optimal control command and provide feedback on the vehicle status. The feedback correction model corrects the data to form a closed-loop control.

2. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 1, characterized in that: The high-precision assembly and initial calibration in step S1 includes using a multi-wheel positioning assembly method to correct the wheel positions of all axles so that the geometric shape formed by connecting the center points of each wheel approaches the preset rectangular frame.

3. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 2, characterized in that: The multi-wheel positioning assembly method specifically involves: establishing a rectangular theoretical frame based on the center point of the vehicle's front wheel; measuring the center point positions of each wheel on the middle and rear axles, and calculating their parallelism deviation and propulsion line deviation relative to the rectangular theoretical frame; Depending on the direction and magnitude of the deviation, it is corrected by adding an adjusting shim of a specific thickness at the connection between the corresponding thrust rod and the balance shaft bracket.

4. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 3, characterized in that: The specific rules for correcting the parallelism deviation and the propulsion line deviation include: if the propulsion line deviation of the middle axle is positive, a shim is added at the thrust rod connection on its left side; if it is negative, a shim is added at the thrust rod connection on its right side; the correction of the rear axle is based on the sum of the propulsion line deviation of the middle axle and the parallelism deviation of the rear axle itself, with a shim added on the right side for positive values ​​and on the left side for negative values.

5. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 1, characterized in that: The high-precision assembly and initial calibration in step S1 includes applying controllable downward pressure to the vehicle's load-bearing crossbeam using a suspension system calibration compression tool to overcome the inherent damping of the chassis system, so that the air suspension system can accurately reach and stabilize at the preset initial calibration height.

6. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 5, characterized in that: The high-precision assembly and initial calibration in step S1 include using a split jack device to independently and synchronously support adjacent axles when repairing and replacing air springs, maintaining the relative position and attitude of adjacent axles when lifting the frame, thereby ensuring that the parallelism of the guide spring installation and the coaxiality of the lugs on the axle being repaired meet the assembly requirements.

7. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 6, characterized in that: In step S3, the objective function of the model prediction controller is configured to simultaneously optimize vehicle trajectory tracking accuracy, control input quantity, and control input change rate; the parallel chaotic optimization algorithm expands the search range and improves the solution efficiency through chaotic variable mapping iteration.

8. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 1 or 7, characterized in that: The rolling optimization solution process in step S3 must meet the constraints determined by the vehicle dynamics characteristics. These constraints include limits on the sideslip angle, yaw rate, wheel slip ratio, and drive motor output torque.

9. The active control method for driving stability of new energy vehicles based on adaptive optimization MPC algorithm according to claim 1, characterized in that: In step S4, the feedback correction model is used in the next time step. Collect actual output , and the predicted output Compare and correct the model parameters, then repeat the S1 to S4 process.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the active control method for driving stability of new energy vehicles based on the adaptive optimization MPC algorithm as described in any one of claims 1 to 9.

Images