Method for judging stability of electric vehicle based on tire lateral deviation characteristic and phase plane

Abstract

The invention discloses an electric vehicle stability determination method based on tire side deviation characteristics and a phase plane, and belongs to the technical field of vehicle stability determination, and the method comprises the steps: S1, building a basic simulation model, including simulation model building and configuration of an information acquisition module; s2, phase plane construction and preliminary partitioning are carried out; according to the electric vehicle stability judgment method based on the tire side deviation characteristics and the phase plane, vehicle state parameters and a road adhesion coefficient are obtained through multi-model fusion estimation, then a side slip angle-side slip angle speed phase plane is constructed, preliminary partitioning is conducted in combination with a five-characteristic rhombus method, and the stability of the electric vehicle is judged. Critical parameters are solved based on tire lateral deviation characteristics to complete secondary partitioning, and finally, the stability margin is quantified through the distance between a state point and the instability boundary, so that accurate judgment is realized, the limitation of single-phase plane analysis is made up, the instability boundary is accurately quantified, stable region division is refined, and a clear basis is provided for intervention of a control system. And the stability judgment accuracy and robustness under the limiting working condition are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of automobile stability determination, in particular to an electric vehicle stability determination method based on tire cornering characteristics and phase planes. BACKGROUND

[0002] Electric vehicles can reduce fuel consumption and exhaust emission, have excellent environmental protection benefits, can relieve oil price fluctuation pressure and air pollution problems, are important directions of automobile industry transformation, and help green and sustainable economic development. For example, a kind of electric vehicle lateral stability prediction method is disclosed in Chinese patent application No. 202110244636.0 and application date is March 5, 2021. The method takes the best prototype model as a platform, comprehensively analyzes the dynamic characteristics of the electric vehicle under the steering and braking conditions, applies the response surface method to establish the electric vehicle lateral stability prediction model under various operating conditions, analyzes the correlation of the uncertainty variable variation of the electric vehicle suspension system assembly position, simplifies the electric vehicle lateral stability prediction model, and applies the preference integration function to solve the prediction result of the electric vehicle lateral stability prediction model. The motor subsystem and hydraulic subsystem in the electric vehicle braking system are combined to construct the electric vehicle lateral stability control strategy, so that the lateral stability of the electric vehicle under various operating conditions can be improved. Chinese patent application No. 202010677081.4 and application date is July 14, 2020 discloses a kind of electric vehicle on-board power supply stability detection system and method, which has the advantage of fully researching the stability of electric vehicle power battery.

[0003] However, the single-phase analysis only relies on the vehicle state parameters and does not take into account the tire cornering characteristics, so it cannot fully represent the stability. If the two-core analysis dimensions cannot be integrated, the stability determination will be limited to a single dimension information, it is difficult to fully represent the overall dynamic response of the vehicle and the nonlinear force characteristics of the tire, resulting in fuzzy quantization of the instability boundary, rough region division, and further affecting the accuracy of the control system intervention time, reducing the vehicle stability control effect under extreme conditions, and affecting the accuracy of the determination. Therefore, an electric vehicle stability determination method based on tire cornering characteristics and phase planes is proposed to solve the above problems. SUMMARY

[0004] The purpose of the present application is to provide an electric vehicle stability determination method based on tire cornering characteristics and phase planes to solve the problem of not being able to integrate the two-core analysis dimensions, which is difficult to fully represent the overall dynamic response of the vehicle and the nonlinear force characteristics of the tire, and ultimately affects the accuracy of the determination.

[0005] To achieve the above objectives, this invention provides the following technical solution: a method for determining the stability of electric vehicles based on tire side slip characteristics and phase planes, comprising the following steps: S1, basic simulation model construction, including simulation model construction and configuration information acquisition module, constructing a simulation platform containing core dynamics models and sensor data interfaces, providing a data acquisition and experimental basis for stability analysis; S2, phase plane construction and preliminary partitioning, including selection of phase plane characteristic quantities, multi-condition simulation tests and analysis of the influence of condition parameters, and five-characteristic rhombus method partitioning, establishing a two-dimensional phase plane and dividing the preliminary stable domain and instability domain through multi-condition tests, clarifying the influence law of core parameters on stability; S3, tire side slip characteristic analysis and critical parameter solution, including side slip characteristic curve partitioning, parameter relationship simulation study and critical front wheel steering angle solution, and center of gravity side slip characteristic analysis and critical parameter solution. The process includes: S4, S5, S6, S7, S8, S9, S1, S2, S3, S4, S6, S7, S8, S9, S1, S2, S3, S4, S1, S2, S3, S4, S1, S2, S3, S4, S1, S3, S4, S1, S4, S1, S3, S4, S1, S4, S1, S1, S3, S1, S4, S1 ...

[0006] Preferably, when building the simulation model in S1, a seven-degree-of-freedom vehicle dynamics model and a magic formula tire model are constructed based on the MATLAB / Simulink-CarSim co-simulation platform to form the core simulation framework. In S1, when configuring the information acquisition module, a steering wheel angle torque sensor and a gyroscope are connected to acquire key parameters such as steering wheel angle, longitudinal and lateral acceleration, and yaw rate.

[0007] Preferably, when selecting phase plane characteristic quantities in S2, the centroid sideslip angle and its rate of change are determined as core characteristic quantities, and a two-dimensional phase plane of centroid sideslip angle-centroid sideslip angular velocity is constructed. In the multi-condition simulation test in S2, simulation tests of straight-line driving condition, double lane change condition, and sinusoidal hysteresis condition are carried out by changing the initial values ​​of the center of gravity sideslip angle and the center of gravity sideslip angular velocity.

[0008] Preferably, in step S2, when performing the influence analysis of working condition parameters, phase plane data under different front wheel steering angles, road surface adhesion coefficients, and longitudinal vehicle speeds are collected to generate phase trajectory manifolds and analyze the effect of each parameter on vehicle stability. When performing the five-feature rhombus method partitioning in S2, five feature values ​​are extracted: the phase plane equilibrium point, the upper and lower endpoints of the stable region, and the left and right saddle points. The stable region and the unstable region are divided, and a feature value database under specific working conditions is established.

[0009] Preferably, in step S3, when dividing the side slip characteristic curve, the tire side slip angle-tire lateral force characteristic curve is plotted and divided into three characteristic regions: linear region, nonlinear region, and saturation region. In S3, when conducting parameter relationship simulation, multi-condition simulation is carried out to analyze the corresponding relationship between front wheel steering angle, yaw rate, and tire lateral force under different road surface adhesion coefficients and vehicle speeds.

[0010] Preferably, in step S3, when solving for the critical front wheel steering angle, the critical front wheel steering angle under different working conditions is calculated based on the turning characteristics of each region of the characteristic curve, in the linear region-nonlinear region and the nonlinear region-saturation region. In S3, when deriving the critical value of the center of gravity sideslip angle, the critical value of the center of gravity sideslip angle is further derived based on the critical front wheel steering angle, and the mapping relationship between the critical parameter and the working condition parameter is established.

[0011] Preferably, in step S4, when formulating the secondary partitioning rules, the critical value of the centroid side deviation angle is used as the basis, and the secondary partitioning rules of the stability region are formulated in combination with the preliminary partitioning results of the phase plane. In S4, when constructing the three-level regional system, the original stable region of the phase plane is subdivided into a stable region and a cooperative region. The cooperative region requires the coordinated intervention of AFS and DYC, while the original unstable region is retained, forming a three-level division system of stable region-cooperative region-unstable region. The boundary parameter database construction in S4 records the boundary coordinates of each area under different combinations of front wheel steering angle, road surface adhesion coefficient, and longitudinal vehicle speed, and establishes a complete boundary feature parameter database.

[0012] Preferably, when defining core indicators in S5, the distance from the vehicle state point to the instability boundary is taken as the core to define vehicle stability evaluation indicators, and the indicator values ​​are positively correlated with stability margin. When optimizing and correcting the indicators in S5, dynamic weighting coefficients are introduced to optimize the static distance calculation logic and improve the accuracy of the indicators, taking into account the vehicle state point movement trend, i.e. the convergence and divergence directions of the phase trajectory.

[0013] Preferably, in step S6, when conducting the full-condition verification test, it is necessary to carry out a full-condition coverage verification test based on the joint simulation platform and compare the judgment results with the vehicle stability status measured by the sensors. In step S6, when performing error source analysis, the error sources in the phase plane feature value extraction accuracy and tire side slip characteristic partitioning threshold are investigated.

[0014] Preferably, in step S6, when performing parameter iterative optimization, it is necessary to adjust the mapping relationship between the region boundary parameters and critical parameters, and to optimize the iterative optimization judgment method to ensure its accuracy and robustness.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: It employs a novel structural design, first obtaining vehicle state parameters and road adhesion coefficients through multi-model fusion estimation, then constructing a phase plane of center-of-gravity sideslip angle and center-of-gravity sideslip angular velocity, and initially partitioning using the five-feature rhombus method. Secondary partitioning is completed by solving critical parameters based on tire sideslip characteristics. Finally, the stability margin is quantified by the distance from the state point to the instability boundary, achieving accurate determination. This overcomes the limitations of single-phase-plane analysis, accurately quantifies the instability boundary, refines the stable region division, provides a clear basis for control system intervention, improves the accuracy and robustness of stability determination under extreme conditions, and contributes to improving the handling stability and driving safety of distributed drive electric vehicles. The specific details are as follows: This method for determining the stability of electric vehicles based on tire sideslip characteristics and phase planes overcomes the limitations of single-analysis and improves the comprehensiveness of the determination. It breaks through the limitation of single phase plane analysis relying solely on vehicle state parameters, and deeply integrates tire sideslip characteristics with phase plane analysis of center of gravity sideslip angle-center of gravity sideslip angular velocity. It considers both the overall dynamic response of the vehicle and accurately captures the nonlinear stress state of the tire. At the same time, by simultaneously incorporating key influencing factors such as front wheel steering angle, road adhesion coefficient, and longitudinal vehicle speed, it comprehensively covers multi-dimensional variables during vehicle operation, effectively avoiding the stability misjudgment caused by the one-sided perspective of traditional methods. This makes the determination results more reflective of the actual driving state of the vehicle and provides a more comprehensive decision-making basis for subsequent control strategy formulation.

[0016] This method for determining the stability of electric vehicles based on tire sideslip characteristics and the phase plane refines the region division system to adapt to actual control needs. Compared with the traditional binary coarse division of the stable and unstable regions, this method further subdivides the original stable region of the phase plane into a stable region and a cooperative region by solving the critical values ​​of the critical front wheel steering angle and the center of gravity sideslip angle, thus constructing a three-level division system of stable region, cooperative region, and unstable region. At the same time, it uses the five-feature rhombus method to extract key feature parameters and establish a boundary feature database under multiple operating conditions to accurately define the boundaries of each region. This refined division clarifies the control intervention requirements corresponding to different regions, solves the problem of ambiguous control intervention timing in traditional methods, and provides a clear and operable basis for the independent control and cooperative control switching of AFS and DYC systems.

[0017] This electric vehicle stability assessment method based on tire side-slip characteristics and phase plane can quantify instability boundaries and evaluation indicators, improving the accuracy of assessment. By constructing a dataset through multi-condition simulation tests and optimizing the expression of instability boundaries in combination with tire side-slip characteristics, the method achieves accurate quantification of instability boundaries, overcoming the shortcomings of traditional phase plane methods that rely on model accuracy and have ambiguous boundaries under extreme conditions. At the same time, based on the distance from the vehicle's state point to the instability boundary and the motion trend, a quantitative evaluation indicator is designed to intuitively reflect the vehicle's stability margin. The indicator has clear physical meaning and is easy to quantify. This indicator can dynamically adapt to changes in road surface adhesion coefficients, vehicle speeds, and other conditions, effectively reducing the impact of parameter uncertainty and model errors, and significantly improving the accuracy of stability assessment under extreme driving conditions.

[0018] This electric vehicle stability determination method based on tire side slip characteristics and phase plane can enhance the adaptability and robustness of operating conditions, facilitating practical application. A dataset is constructed through multi-condition simulation tests including straight-line driving, double lane change, serpentine driving, and fishhook driving, fully considering the parameter variation patterns under complex driving scenarios. It is specifically optimized for the characteristics of strong coupling, nonlinearity, and time-varying behavior of vehicles under extreme conditions, significantly improving adaptability to complex operating conditions. Through iterative optimization via co-simulation, hardware-in-the-loop simulation, and real-vehicle testing, this method demonstrates strong robustness to uncertainties such as tire side slip stiffness changes and lateral wind interference. Its determination results can provide precise support for the chassis collaborative control of distributed drive electric vehicles, effectively improving vehicle handling stability and driving safety, and possessing promising engineering application prospects. Attached Figure Description

[0019] Fig. 1 This is a schematic diagram of the S1-S3 workflow of the present invention; Fig. 2 This is a schematic diagram of the S4-S6 workflow of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figs. 1-2 The present invention provides the following technical solution: a method for determining the stability of electric vehicles based on tire lateral slip characteristics and phase plane.

[0022] This includes S1, basic simulation model construction, including simulation model construction and configuration information acquisition module, and building a simulation platform that includes core dynamic model and sensor data interface, providing a data acquisition and experimental basis for stability analysis; When building the simulation model in S1, the MATLAB / Simulink-CarSim co-simulation platform is used to construct a seven-degree-of-freedom vehicle dynamics model and a magic formula tire model, forming the core simulation framework. The simulation model in S1 above aims to support accurate estimation of vehicle state parameters and road adhesion coefficient, and scientific determination of stability status. It integrates multi-dimensional models and data interfaces to construct a complete simulation system. During operation, a seven-degree-of-freedom vehicle dynamics model is first built, comprehensively covering the vehicle's longitudinal, lateral, and yaw motion characteristics, accurately depicting the vehicle's dynamic response. Simultaneously, a magic formula tire model is introduced to accurately represent the nonlinear relationship between tire slip angle and lateral force, closely reflecting the tire's force characteristics during actual driving. Secondly, in terms of data acquisition interface configuration, it connects to a steering wheel angle and torque sensor to acquire steering wheel angle and torque data in real time; and it connects to a gyroscope to collect longitudinal acceleration data. Key motion parameters such as angle of deviation, lateral acceleration, and yaw rate provide measured data support for model input. Furthermore, parameter estimation model integration is a crucial step. Based on the square root commensurate Kalman filter algorithm, combined with the vehicle's longitudinal, lateral, and lateral dynamic equations, a filtering estimation model for the center of gravity sideslip angle and tire lateral force is established, clearly defining the filtering observation vector and input vector. Addressing the issue of unknown statistical characteristics of system process noise and measurement noise under extreme conditions, an adaptive antlion optimization algorithm is introduced. With the goal of minimizing the sum of the mean square errors of the actual and estimated measurement outputs, the algorithm identifies noise characteristics and improves its global search capability by adaptively adjusting the antlion's random walk range, avoiding local optima. Meanwhile, based on the kinematic relationship between the sideslip angular velocity of the center of mass and the longitudinal and lateral accelerations, an integral estimation model for the sideslip angular velocity of the center of mass is constructed. A two-dimensional extension set is established using extension theory, dividing the domain into classical, fusion, and non-domains. The fusion weight of each region is determined by the extension correlation function, realizing the fusion of filtering estimation and integral estimation, thus overcoming the shortcomings of single-model estimation. Finally, relying on the MATLAB / Simulink-CarSim co-simulation platform, a multi-model data interaction channel is established, laying the foundation for subsequent multi-condition simulation experiments, parameter optimization, and verification of stability determination methods.

[0023] In S1, when configuring the information acquisition module, the steering wheel angle torque sensor and gyroscope are connected to realize the acquisition of key parameters such as steering wheel angle, longitudinal and lateral acceleration, and yaw rate.

[0024] When configuring the information acquisition module in S1 above, core sensing devices are specifically selected for sensor selection and interface adaptation: the steering wheel angle and torque sensor is responsible for collecting steering wheel angle and torque data to accurately capture the driver's operating intentions; the gyroscope is used to obtain the vehicle's longitudinal acceleration, lateral acceleration, and yaw rate, comprehensively reflecting the vehicle's dynamic motion state. The module needs to complete the interface adaptation between the sensors and the MATLAB / Simulink-CarSim co-simulation platform to ensure that various sensor data can be smoothly integrated into the simulation model, providing raw input for subsequent dynamic analysis and parameter estimation. Then, a real-time data acquisition and transmission mechanism is constructed. The information acquisition module is configured to synchronously acquire data from multiple sensors according to a set sampling frequency to avoid data timing misalignment affecting analysis accuracy. Dedicated data transmission channels are established for the output characteristics of different sensors to ensure that key parameters such as angle, acceleration, and angular velocity are transmitted without distortion, guaranteeing the real-time performance and integrity of the data, and meeting the timeliness requirements of vehicle dynamic response analysis. Furthermore, the raw data preprocessing stage is indispensable. Considering that onboard sensors are easily affected by environmental interference in actual operation, and the acquired raw data contains noise, the module is configured with appropriate preprocessing mechanisms. The data acquisition module integrates a limiting filter algorithm to perform preliminary data purification, eliminating outliers and reducing the impact of environmental interference on data quality. Simultaneously, it reserves a data calibration interface to address potential sensor calibration and drift errors, laying the foundation for subsequent error correction and improved data accuracy through algorithms. Finally, multi-source data is collaboratively integrated. The configuration data acquisition module collaboratively integrates pre-processed multi-dimensional data such as steering wheel angle, torque, longitudinal and lateral acceleration, and yaw rate to form a standardized dataset. This dataset directly provides input for the seven-DOF vehicle dynamics model and the magic formula tire model, and also provides observation and input vectors for the square root commensurate Kalman filter estimation model and the integral estimation model. This achieves seamless integration of multi-source data with subsequent estimation and judgment models, ensuring the smooth operation of the entire simulation and analysis system.

[0025] S2. Phase plane construction and preliminary partitioning, including phase plane characteristic quantity selection, multi-condition simulation test and condition parameter influence analysis, and five-character rhombus method partitioning, to establish a two-dimensional phase plane and divide the preliminary stable domain and unstable domain through multi-condition test, and to clarify the influence law of core parameters on stability. When selecting phase plane characteristic quantities in S2, the centroid sideslip angle and its rate of change are determined as core characteristic quantities, and a two-dimensional phase plane of centroid sideslip angle-centroid sideslip angular velocity is constructed. The aforementioned S2 phase plane construction and preliminary partitioning first selects the sideslip angle and its rate of change as core feature quantities to construct a two-dimensional phase plane of sideslip angle-slip angular velocity, providing a basic framework for stability analysis. Subsequently, relying on the MATLAB / Simulink-CarSim co-simulation platform, the initial values ​​of the sideslip angle and yaw rate are changed to carry out multi-condition simulations such as straight driving, double lane change, and serpentine driving. Phase plane data under different front wheel steering angles, road adhesion coefficients, and longitudinal vehicle speeds are collected to generate phase trajectory manifolds to analyze the influence of each parameter on stability. Finally, the five-feature rhombus method is used to extract five feature values: the phase plane equilibrium point, the upper and lower endpoints of the stable region, and the left and right saddle points, establishing a feature value database under specific conditions and completing the preliminary division of the stable and unstable regions, laying the foundation for subsequent accurate determination.

[0026] In the multi-condition simulation test in S2, simulation tests of straight driving condition, double lane change condition, and sinusoidal hysteresis condition are carried out by changing the initial values ​​of the center of gravity sideslip angle and the center of gravity sideslip angular velocity.

[0027] In the above S2, the core objectives of the multi-condition simulation test are to obtain comprehensive vehicle stability data and support phase plane partitioning. The test is conducted using the MATLAB / Simulink-CarSim co-simulation platform. First, based on the established seven-DOF vehicle dynamics model and magic formula tire model, diverse driving conditions such as straight-line driving, double lane change, sinusoidal hysteresis, serpentine, figure-eight steering, and fishhook-like maneuvers are set, covering both normal and extreme driving scenarios. Second, by adjusting the initial values ​​of the center of gravity sideslip angle and angular velocity, as well as key parameters such as front wheel angle, road adhesion coefficient, and longitudinal vehicle speed, the dynamic response of the vehicle under different driving conditions is simulated. During the test, phase plane characteristic data (center of gravity sideslip angle and angular velocity) and related parameters are collected simultaneously to generate multi-condition phase trajectory manifolds. The influence of each parameter on the vehicle stability domain boundary is analyzed, providing comprehensive and realistic dataset support for subsequent division of the stability and instability domains using the five-feature rhombus method and the establishment of an eigenvalue database.

[0028] In S2, when performing the influence analysis of working condition parameters, phase plane data are collected under different front wheel steering angles, road surface adhesion coefficients, and longitudinal vehicle speeds to generate phase trajectory manifolds and analyze the effect of each parameter on vehicle stability. The above-mentioned analysis of the influence of working condition parameters aims to clarify the effects of front wheel steering angle, road surface adhesion coefficient, and longitudinal vehicle speed on vehicle stability. The working principle revolves around data correlation analysis and pattern extraction. First, based on the dataset collected from multi-working-condition simulation tests, phase plane characteristic data (center of gravity sideslip angle, center of gravity sideslip angular velocity) and phase trajectory manifold information under different parameter combinations are extracted. Second, the controlled variable method is used to fix two parameters and change another parameter, and the differences in the changes of the phase plane stability domain boundary position and phase trajectory shape are compared and analyzed. Finally, the influence of each parameter on the stability domain range and instability risk is quantitatively extracted. For example, a decrease in road surface adhesion coefficient will shrink the stability domain, and a large front wheel steering angle will easily cause the phase trajectory to shift towards the instability domain. Thus, a mapping relationship between parameters and stability is established, providing a theoretical basis for subsequent five-feature rhombus method partitioning and critical parameter solution.

[0029] When performing the five-feature rhombus method partitioning in S2, five feature values ​​are extracted: the equilibrium point of the phase plane, the upper and lower endpoints of the stable region, and the left and right saddle points. This divides the stable region into the unstable region and establishes a feature value database for specific working conditions.

[0030] The aforementioned five-feature rhombus method for zoning aims to accurately divide the stable and unstable regions of the phase plane. Its working principle revolves around feature value extraction, boundary definition, and region division. First, based on the phase trajectory data of the center-of-mass sideslip angle and angular velocity obtained from multi-condition simulations, key feature points of the phase plane are selected, and five core feature values ​​are extracted: the equilibrium point (the reference point for stable vehicle operation), the upper and lower endpoints of the stable region (longitudinal limit feature points), and the left and right saddle points (lateral critical feature points). Second, a rhombus boundary is constructed using these five feature values ​​as vertices, clearly defining the interior of the rhombus as the stable region (the area where the vehicle can maintain stability autonomously) and the exterior of the rhombus as the unstable region (the area where the vehicle is prone to sideslip and fishtailing). Finally, a database of the five feature values ​​is established under different combinations of front wheel steering angle, road surface adhesion coefficient, and longitudinal vehicle speed, forming a dynamic zoning model adapted to different operating conditions. This provides a basic boundary basis for subsequent secondary division of the stable region and stability determination.

[0031] S3. Tire side slip characteristic analysis and critical parameter solution, including side slip characteristic curve partitioning, parameter relationship simulation study, critical front wheel angle solution and centroid side slip critical value derivation, dividing the tire side slip characteristic region and solving the critical front wheel angle, providing key parameter support for accurate division of stable region. In S3, when dividing the side slip characteristic curve, the tire side slip angle-tire lateral force characteristic curve is plotted and divided into three characteristic regions: linear region, nonlinear region, and saturation region. In S3 above, the core objective of partitioning the side slip characteristic curve is to accurately define different stress states of the tire and provide a basis for subsequent critical parameter solutions. The working principle revolves around curve plotting, feature recognition, and region division. First, based on multi-condition simulation test data, the tire side slip angle and lateral force data under different front wheel steering angles and road adhesion coefficients are extracted, and the tire side slip angle-lateral force characteristic curve is plotted to intuitively present the nonlinear relationship between the two. Second, the curve morphology is analyzed to identify key turning points: the approximately linear growth segment of the curve is the linear region (lateral force increases uniformly with the side slip angle), the transition segment where the growth rate slows down is the nonlinear region, and the saturation segment where the growth tends to level off (lateral force stabilizes or decreases after reaching its peak). Finally, the side slip angle corresponding to the turning point is used as the threshold to complete the division of the linear region, nonlinear region, and saturation region. The correspondence between the threshold of each region and the operating condition parameters is recorded simultaneously to establish a side slip characteristic partitioning database, laying the foundation for subsequent critical front wheel steering angle solutions.

[0032] In S3, when conducting simulations to explore parameter relationships, multi-condition simulations are carried out to analyze the corresponding relationships between the front wheel steering angle, yaw rate, and tire lateral force under different road surface adhesion coefficients and vehicle speeds.

[0033] The parameter relationship simulation exploration in S3 above aims to clarify the correlation between front wheel steering angle, yaw rate, and tire lateral force, providing a basis for solving critical parameters. The working principle revolves around setting operating conditions, data acquisition, and correlation analysis. First, relying on the established co-simulation platform, the basic vehicle parameters are fixed, and multiple combinations of operating conditions with different road adhesion coefficients and longitudinal vehicle speeds are set, covering both normal and extreme driving scenarios. Second, the front wheel steering angle input is gradually adjusted under each operating condition, and real-time data of yaw rate and tire lateral force under the corresponding operating conditions are collected simultaneously to establish a multi-dimensional parameter dataset. Finally, through data correlation analysis and curve fitting, the quantitative correspondence between front wheel steering angle, yaw rate, and tire lateral force under different combinations of operating condition parameters is extracted, clarifying the coupling influence of parameter changes and providing data support and theoretical basis for subsequent critical front wheel steering angle solutions.

[0034] In S3, when solving for the critical front wheel steering angle, the critical front wheel steering angle under different working conditions is calculated based on the turning characteristics of each region of the characteristic curve in the linear region-nonlinear region and the nonlinear region-saturation region. The above-mentioned solution for the critical front wheel steering angle in S3 first determines the critical steering angle (i.e., the steering angle corresponding to the turning point where the lateral force growth tends to level off) based on the partitioning results of the side slip characteristic curve, where the tire side slip characteristic transitions from the nonlinear region to the saturation region. Then, a quantitative mapping model between the front wheel steering angle and the tire side slip angle, obtained from parameter relationship simulation, is established to establish a functional relationship between the two. Finally, the critical steering angle is substituted into the mapping model to derive the front wheel steering angle value under the corresponding working condition. At the same time, the accuracy of the solution is verified by combining the abrupt change characteristics of vehicle dynamic parameters such as yaw rate and lateral acceleration. The above process is repeated for different road adhesion coefficients and longitudinal vehicle speeds to form a database of critical front wheel steering angles under multiple working conditions, providing key parameter support for the subsequent secondary partitioning of the stable region.

[0035] In S3, when deriving the critical value of the center of gravity sideslip angle, the critical value of the center of gravity sideslip angle is further derived based on the critical front wheel rotation angle, and the mapping relationship between the critical parameters and the working condition parameters is established.

[0036] The derivation of the critical value of the center of gravity sideslip angle in S3 above aims to clarify the sideslip limit for stable vehicle operation. Its working principle revolves around the correlation of critical sideslip angles, support from dynamic models, and verification under multiple operating conditions. First, based on the partitioning results of the sideslip characteristic curve, the critical sideslip angle for the tire entering the saturation zone is extracted. This angle is the key node where the tire's lateral force reaches its limit and is also a core correlation parameter for the vehicle's stability transition. Second, combined with a seven-DOF vehicle dynamics model, a dynamic mapping relationship between the center of gravity sideslip angle and the tire sideslip angle is established, clarifying the coupling law between the two with the vehicle's dynamic response. Finally, the critical sideslip angle is substituted into the mapping relationship to derive the critical value of the center of gravity sideslip angle under the corresponding operating condition. Simultaneously, the derivation error is corrected by combining measured data of the center of gravity sideslip angle during vehicle instability in multi-condition simulations. A multi-condition critical value database is constructed by repeating the process for different road surface adhesion coefficients and vehicle speeds, providing core threshold support for subsequent secondary partitioning of stable regions.

[0037] S4. Secondary division of stable regions, including the formulation of secondary partitioning rules, the construction of a three-level regional system and the construction of a boundary parameter database. The original stable region of the phase plane is subdivided into stable regions and cooperative regions by combining critical parameters. A three-level regional division system is constructed to adapt to different control requirements. In S4, when formulating the secondary partitioning rules, the critical value of the centroid side slip angle is used as the basis, combined with the preliminary partitioning results of the phase plane, to formulate the secondary partitioning rules of the stability domain; The secondary partitioning rules in S4 above aim to refine the stability region and clarify the control intervention boundary. The working principle revolves around the initial parameter fusion, threshold definition, and rule construction. First, it integrates the stability domain boundary data of the initial partitioning using the five-feature diamond method in S2, as well as the critical front wheel steering angle and the critical value of the center of gravity sideslip angle solved in S3, to establish a multi-dimensional parameter correlation model. Second, it uses the critical parameters as the core thresholds to define the hierarchical standards within the stability domain: areas below the critical threshold where the vehicle can stabilize autonomously are classified as stability domains that do not require control; areas between the critical threshold and the boundary of the initial stability domain that require system coordinated intervention are classified as coordinated control domains. Finally, it formulates multi-condition adaptation rules, combining the differences in parameter thresholds under different road surface adhesion coefficients and vehicle speeds to establish a dynamic partitioning mapping table, clarifying the judgment conditions and boundary ranges of each region, forming a three-level partitioning rule system of stability domain, coordinated domain, and instability domain, providing a basis for accurate judgment and control strategy switching.

[0038] In S4, when constructing a three-level regional system, the original stable region of the phase plane is subdivided into a stable region and a cooperative region. The cooperative region requires the coordinated intervention of AFS and DYC, while the original unstable region is retained, forming a three-level division system of stable region-cooperative region-unstable region. The three-level regional system construction in S4 above first reuses the basic boundaries of the stable and unstable regions determined by the five-feature rhombus method in S2, retaining the core judgment logic of stability within the rhombus and instability outside the rhombus. Then, it embeds core parameters such as the critical value of the centroid sideslip angle and the critical front wheel steering angle solved in S3 as key thresholds for the internal stratification of the stable region, clarifying the dividing standard between autonomous vehicle stability and the need for control intervention. Finally, through the superposition of parameter thresholds and basic boundaries, a three-level system is formed: the autonomous stable region below the critical threshold, the cooperative control region between the critical threshold and the rhombus boundary, and the unstable region outside the rhombus boundary. At the same time, it is associated with a multi-condition parameter database to achieve dynamic adaptation under different road surfaces and vehicle speeds, providing a clear regional basis for subsequent stability judgment and control intervention switching. The construction of the three-level regional system aims to accurately define different stability states of the vehicle and support differentiated control strategies.

[0039] The boundary parameter database in S4 is constructed to record the boundary coordinates of each area under different combinations of front wheel steering angle, road surface adhesion coefficient, and longitudinal vehicle speed, thus establishing a complete boundary feature parameter database.

[0040] The boundary parameter database construction in S4 above aims to achieve accurate adaptation of the three-level region boundaries under multiple working conditions, providing data support for dynamic stability determination. Its working principle revolves around parameter acquisition, classification and integration, and correlation mapping. First, it summarizes core data from each stage, including the stability domain boundary parameters of the five-feature rhombus method under multiple working conditions in S2, the critical front wheel steering angle and centroid sideslip angle critical values ​​corresponding to different working conditions solved in S3, and basic parameters such as road adhesion coefficient and longitudinal vehicle speed for each working condition. Second, it classifies and integrates data according to the correspondence between working condition parameters and boundary parameters, establishing a standardized data structure and clarifying the boundary thresholds between the autonomous stability domain and the cooperative control domain, and the boundary parameters between the cooperative control domain and the instability domain under different road surface and vehicle speed conditions. Finally, through data correlation modeling and verification optimization, a dynamically adaptable boundary parameter database is formed, enabling the rapid matching of corresponding region boundary parameters upon input of working condition parameters, providing reliable data assurance for the dynamic application of the three-level region system.

[0041] S5. Design of stability quantitative evaluation index, including the definition of core index and index optimization and correction, defining a quantitative index based on the distance from the state point to the instability boundary to achieve accurate characterization of vehicle stability margin. When defining core indicators in S5, the distance from the vehicle's state point to the instability boundary is used as the core to define vehicle stability evaluation indicators, and the indicator values ​​are positively correlated with stability margin. In the aforementioned S5, the core objective in defining the core indicators is to accurately quantify vehicle stability margin and support stability level determination. The working principle revolves around indicator selection, quantitative modeling, and operating condition adaptation. First, based on the boundary characteristics of the three-level regional system, the distance from the vehicle's real-time state point to the boundary of each region is selected as the core determination indicator. This indicator can intuitively reflect the degree to which the vehicle deviates from the stable state and has a clear physical meaning. Second, combined with the dynamic boundary parameters in the boundary parameter database, an indicator quantification model is established: by calculating the real-time centroid sideslip angle, centroid sideslip angular velocity, and the Euclidean distance between the boundary of the autonomous stable domain and the boundary of the instability domain under the corresponding operating conditions, a standardized stability margin value is converted. Finally, an indicator threshold grading standard is set, corresponding to the three-level regional states (high stability margin corresponds to the autonomous stable domain, medium to the cooperative control domain, and low to the instability domain), and multi-operating condition parameters are associated to achieve dynamic adaptation, ensuring the consistency and accuracy of indicator determination under different road surfaces and vehicle speeds, and providing a quantitative basis for the final stability determination.

[0042] When optimizing and correcting indicators in S5, dynamic weighting coefficients are introduced to optimize the static distance calculation logic and improve the accuracy of indicators by combining the vehicle state point movement trend, i.e. the convergence and divergence direction of the phase trajectory.

[0043] When optimizing and correcting the indicators in S5 above, the core objective is to improve the accuracy of the stability margin indicator and adapt to complex operating conditions. The working principle revolves around error identification, correction modeling, and verification optimization. First, by comparing multi-condition simulation and real vehicle test data, deviations in the initial indicators are identified, such as inaccurate quantification of boundary distances under extreme conditions and poor consistency in judgment under different road surface adhesion coefficients. Second, operating condition weight coefficients and error compensation models are introduced. Based on historical data in the boundary parameter database, an indicator correction function is established: the weights are dynamically adjusted using road surface adhesion coefficient and longitudinal vehicle speed as variables to compensate for errors in the initial stability margin value and correct quantification deviations under extreme conditions. Finally, iterative verification is performed using a large amount of multi-condition sample data to continuously optimize the correction function parameters, ensuring that the corrected indicators can accurately match the actual stability state of the vehicle under both normal and extreme conditions, thereby improving the reliability and robustness of subsequent stability judgments.

[0044] S6. Method validation and parameter optimization, including full-condition validation tests, error source analysis and parameter iterative optimization. The accuracy of the method is determined by full-condition tests, and key parameters are iteratively optimized to improve the robustness and practicality of the method.

[0045] In S6, when conducting full-condition verification tests, it is necessary to carry out full-condition coverage verification tests based on the joint simulation platform and compare the judgment results with the vehicle stability status measured by the sensors. The aforementioned S6 full-condition verification test aims to comprehensively test the accuracy, robustness, and multi-condition adaptability of the stability assessment method. Its working principle revolves around condition coverage, data acquisition, and result verification. First, based on actual driving scenario requirements, a full-condition matrix is ​​designed, including conventional (straight-line, constant-speed steering), extreme (emergency avoidance, low-adhesion road steering), and special (lateral wind interference, load changes) conditions to ensure the test covers the application boundaries of the assessment method. Second, relying on a joint simulation platform and real-vehicle testing linkage, real-time vehicle state parameters, assessment method output results, and actual stability state data are simultaneously collected under each condition. Finally, the accuracy of the assessment is quantitatively evaluated by comparing and analyzing the consistency between the assessment results and the actual conditions. For deviation conditions, the boundary parameter database and index correction model are optimized, and iterative verification is performed until the accuracy of the assessment under all conditions reaches the target, ensuring that the assessment method can be applied to actual driving scenarios.

[0046] In S6, when performing error source analysis, the error sources in the phase plane feature value extraction accuracy and tire side slip characteristic partitioning threshold are investigated.

[0047] In S6 above, the core objective of error source analysis is to accurately locate the root cause of stability judgment deviations and provide targeted basis for subsequent method optimization. The working principle revolves around deviation identification, source analysis, and classification. First, based on full-condition verification test data, deviation samples that do not match the actual stability state of the vehicle are extracted, and the corresponding operating condition parameters and judgment links for each deviation sample are clarified. Second, the source is traced according to the dimensions of modeling-data-operating condition-algorithm: at the model level, the inherent errors caused by the simplification of the dynamic model and the assumptions of the tire model are analyzed; at the data level, sensor noise, data transmission distortion, and database parameter deviations are investigated; at the operating condition level, the adaptation errors caused by parameter coupling under extreme and special operating conditions are focused on; at the algorithm level, the logical defects of filtering estimation and index calculation are examined. Finally, the contribution of each error source to the judgment deviation is quantified to form a list of classified errors, clarifying the core error root causes and the scope of influence, and providing precise targets for subsequent model correction, parameter calibration, and algorithm optimization.

[0048] In S6, when performing parameter iterative optimization, it is necessary to adjust the mapping relationship between region boundary parameters and critical parameters, and to determine the iterative optimization method to ensure its accuracy and robustness.

[0049] The aforementioned S6 parameter iterative optimization aims to accurately address the judgment biases discovered during full-condition verification and improve the accuracy and robustness of the stability judgment method. Its working principle revolves around three core aspects: error-targeted matching, multi-dimensional parameter correction, and closed-loop iterative verification. First, based on the preliminary error source analysis results, a mapping relationship between error type and key parameters is established to precisely identify the parameters to be optimized: for inherent model errors, the focus is on core parameters such as the inertial parameters of the seven-DOF vehicle dynamics model and the lateral stiffness coefficient of the tire model; for data-related errors, the critical front wheel steering angle threshold and the critical value of the center of gravity lateral slip angle under different conditions are identified in the boundary parameter database; for algorithm calculation errors, the focus is on optimizing the condition weight coefficient of the index correction function and the noise covariance matrix of the filtering estimation algorithm, and clarifying the reasonable adjustment range of each parameter to avoid parameter overflow leading to judgment logic failure. A multi-objective optimization model is constructed, with the core objective of maximizing the consistency between the judgment results and the actual stability state of the vehicle under all working conditions, while also considering the balance of judgment accuracy under different extreme working conditions (low-adhesion road surface, emergency avoidance). An adaptive particle swarm optimization algorithm is adopted, and the key parameters are adjusted in batches and in multiple rounds through steps such as population initialization, fitness function calculation, and particle position and velocity update: first, the core parameters that contribute the most to the judgment deviation (such as boundary thresholds) are corrected, and then the secondary parameters (such as algorithm weights) are optimized. After each round of adjustment, the parameter combination and the corresponding judgment accuracy are recorded. A closed-loop verification mechanism is then established: After each round of parameter optimization, the new parameters are substituted into the stability judgment model, and the previous deviation scenarios are reproduced through full-condition verification tests to verify the effect of the improved judgment accuracy. If the preset judgment accuracy standard (such as 95% full-condition accuracy) is not reached, the error analysis stage is returned to, the root cause of the deviation is repositioned, the optimization direction and parameter adjustment step size are adjusted, and the parameter adjustment-condition verification-deviation analysis process is repeated until the judgment result is stable and accurate under full-condition conditions, and has good adaptability to uncertain factors such as changes in road surface adhesion coefficient and crosswind interference, forming the final optimized parameter system, which provides a reliable guarantee for the engineering application of the judgment method.

[0050] The above is the entire working process of the device, and all contents not described in detail in this specification are existing technologies known to those skilled in the art.

[0051] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for determining the stability of electric vehicles based on tire lateral slip characteristics and phase plane, characterized in that, Includes the following steps: S1. Basic simulation model construction, including simulation model construction and configuration information acquisition module, to build a simulation platform containing core dynamic model and sensor data interface, providing data acquisition and experimental basis for stability analysis; S2. Phase plane construction and preliminary partitioning, including phase plane characteristic quantity selection, multi-condition simulation test and condition parameter influence analysis, and five-character rhombus method partitioning, to establish a two-dimensional phase plane and divide the preliminary stable domain and unstable domain through multi-condition test, and to clarify the influence law of core parameters on stability. S3. Tire side slip characteristic analysis and critical parameter solution, including side slip characteristic curve partitioning, parameter relationship simulation study, critical front wheel angle solution and centroid side slip critical value derivation, dividing the tire side slip characteristic region and solving the critical front wheel angle, providing key parameter support for accurate division of stable region. S4. Secondary division of stable regions, including the formulation of secondary partitioning rules, the construction of a three-level regional system and the construction of a boundary parameter database. The original stable region of the phase plane is subdivided into stable regions and cooperative regions by combining critical parameters. A three-level regional division system is constructed to adapt to different control requirements. S5. Design of stability quantitative evaluation index, including the definition of core index and index optimization and correction, defining a quantitative index based on the distance from the state point to the instability boundary to achieve accurate characterization of vehicle stability margin. S6. Method validation and parameter optimization, including full-condition validation tests, error source analysis and parameter iterative optimization. The accuracy of the method is determined by full-condition tests, and key parameters are iteratively optimized to improve the robustness and practicality of the method.

2. The method for determining the stability of an electric vehicle based on tire lateral slip characteristics and phase plane as described in claim 1, characterized in that: When building the simulation model in S1, a seven-degree-of-freedom vehicle dynamics model and a magic formula tire model are constructed based on the MATLAB / Simulink-CarSim joint simulation platform to form the core simulation framework. In S1, when configuring the information acquisition module, a steering wheel angle torque sensor and a gyroscope are connected to acquire key parameters such as steering wheel angle, longitudinal and lateral acceleration, and yaw rate.

3. The method for determining the stability of an electric vehicle based on tire lateral slip characteristics and phase plane as described in claim 1, characterized in that: When selecting phase plane characteristic quantities in S2, the centroid side deflection angle and its rate of change are determined as core characteristic quantities, and a two-dimensional phase plane of centroid side deflection angle-centroid side deflection angular velocity is constructed. In the multi-condition simulation test in S2, simulation tests of straight-line driving condition, double lane change condition, and sinusoidal hysteresis condition are carried out by changing the initial values ​​of the center of gravity sideslip angle and the center of gravity sideslip angular velocity.

4. The method for determining the stability of an electric vehicle based on tire lateral slip characteristics and phase plane as described in claim 1, characterized in that: In S2, when performing the influence analysis of working condition parameters, phase plane data under different front wheel steering angles, road surface adhesion coefficients, and longitudinal vehicle speeds are collected to generate phase trajectory manifolds and analyze the effect of each parameter on vehicle stability. When performing the five-feature rhombus method partitioning in S2, five feature values ​​are extracted: the phase plane equilibrium point, the upper and lower endpoints of the stable region, and the left and right saddle points. The stable region and the unstable region are divided, and a feature value database under specific working conditions is established.

5. The method for determining the stability of an electric vehicle based on tire side slip characteristics and phase plane as described in claim 1, characterized in that: In S3, when partitioning the side slip characteristic curve, the tire side slip angle-tire lateral force characteristic curve is plotted and divided into three characteristic regions: linear region, nonlinear region, and saturation region. In S3, when conducting parameter relationship simulation, multi-condition simulation is carried out to analyze the corresponding relationship between front wheel steering angle, yaw rate, and tire lateral force under different road surface adhesion coefficients and vehicle speeds.

6. The method for determining the stability of an electric vehicle based on tire lateral slip characteristics and phase plane as described in claim 1, characterized in that: In S3, when solving for the critical front wheel steering angle, the critical front wheel steering angle under different working conditions is calculated based on the turning characteristics of each region of the characteristic curve in the linear region-nonlinear region and the nonlinear region-saturation region. In S3, when deriving the critical value of the center of gravity sideslip angle, the critical value of the center of gravity sideslip angle is further derived based on the critical front wheel steering angle, and the mapping relationship between the critical parameter and the working condition parameter is established.

7. The method for determining the stability of an electric vehicle based on tire side slip characteristics and phase plane as described in claim 1, characterized in that: In S4, when formulating the secondary partitioning rules, the critical value of the centroid side deviation angle is used as the basis, and the secondary partitioning rules of the stability domain are formulated in combination with the preliminary partitioning results of the phase plane. In S4, when constructing the three-level regional system, the original stable region of the phase plane is subdivided into a stable region and a cooperative region. The cooperative region requires the coordinated intervention of AFS and DYC, while the original unstable region is retained, forming a three-level division system of stable region-cooperative region-unstable region. The boundary parameter database construction in S4 records the boundary coordinates of each area under different combinations of front wheel steering angle, road surface adhesion coefficient, and longitudinal vehicle speed, and establishes a complete boundary feature parameter database.

8. The method for determining the stability of an electric vehicle based on tire side slip characteristics and phase plane as described in claim 1, characterized in that: When defining core indicators in S5, the distance from the vehicle state point to the instability boundary is taken as the core to define vehicle stability evaluation indicators. The indicator values ​​are positively correlated with the stability margin. When optimizing and correcting the indicators in S5, dynamic weighting coefficients are introduced to optimize the static distance calculation logic and improve the accuracy of the indicators, taking into account the vehicle state point movement trend, i.e. the convergence and divergence directions of the phase trajectory.

9. The method for determining the stability of an electric vehicle based on tire side slip characteristics and phase plane as described in claim 1, characterized in that: In S6, when conducting full-condition verification tests, it is necessary to carry out full-condition coverage verification tests based on the joint simulation platform and compare the judgment results with the vehicle stability status measured by the sensors. In step S6, when performing error source analysis, the error sources in the phase plane feature value extraction accuracy and tire side slip characteristic partitioning threshold are investigated.

10. The method for determining the stability of an electric vehicle based on tire side slip characteristics and phase plane according to claim 1, characterized in that: In step S6, when performing parameter iterative optimization, it is necessary to adjust the mapping relationship between the region boundary parameters and critical parameters, and to optimize the determination method to ensure its accuracy and robustness.

Images