Vehicle anti-rollover stability control system and method

Abstract

The invention discloses a vehicle anti-rollover stability control system and method, and relates to the technical field of intelligent control, and the system comprises a driving behavior dynamic monitoring module which collects the behavior data of a driver in real time through a multi-sensor fusion module; the driving environment dynamic monitoring module is used for real-time driving environment data; the driving risk prediction module is used for performing dual-channel risk prediction; the rollover intervention judgment module is used for generating an intervention instruction; and the front rollover intervention module is used for implementing risk intervention according to the intervention instruction and feeding back the operation of the driver in real time to carry out dynamic cooperative control. The technical problem that the rollover risk cannot be timely and accurately controlled under the complex driving condition due to the lack of comprehensive monitoring of the driver behavior and the environment information in the prior art is solved, and the purposes of performing double-channel risk prediction and prospective dynamic intervention by fusing the driving behavior and the environment data and improving the rollover risk control accuracy are achieved. And the technical effects of robustness and real-time performance of vehicle control are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control, and particularly to a vehicle rollover prevention and stability control system and method. Background Art

[0002] With the development of new energy technologies, three-wheeled vehicles driven by new energy have been widely used in fields such as urban logistics and last-mile delivery. However, due to its front single-wheel and rear double-wheel structural characteristics, the vehicle is prone to rollover during high-speed turning, emergency obstacle avoidance, or complex road conditions, posing a significant safety hazard. Existing rollover prevention control methods are mostly applied to four-wheeled vehicles and mainly rely on vehicle dynamic parameters for control. It is difficult to adapt to three-wheeled vehicles with asymmetric structures and poor stability, and lack of comprehensive perception and prediction of driver behavior and environmental changes, resulting in insufficient robustness, leading to system response lag and insufficient accuracy, and unable to perform real-time and effective vehicle control and safety guarantee during actual driving. Summary of the Invention

[0003] This application provides a vehicle rollover prevention and stability control system and method, which are used to solve the technical problem that the existing rollover prevention control of three-wheeled vehicles lacks comprehensive monitoring of driver behavior and environmental information, resulting in the inability to timely and accurately control the rollover risk under complex driving conditions.

[0004] In the first aspect of this application, a vehicle rollover prevention and stability control system is provided. The system includes: a driving behavior dynamic monitoring module, which is used to collect driver behavior data in real time based on a multi-sensor fusion module to obtain driving behavior dynamic information; a driving environment dynamic monitoring module, which is used to collect driving environment information in real time according to a multi-source environment sensor module built in the vehicle, and the driving environment information includes weather environment data, ground state data, and obstacle detection data; a driving risk prediction module, which is used to perform dual-channel independent risk prediction using the driving behavior dynamic information and the driving environment information respectively, and perform risk fusion according to the independent prediction results to generate a driving risk prediction result; a rollover intervention judgment module, which is used to judge whether there is a rollover risk based on the driving risk prediction result. If so, a pre-driving intervention instruction is generated; a pre-rollover intervention module, which is used to perform pre-rollover risk intervention according to the pre-driving intervention instruction, and during the intervention execution process, feedback the driver's operation behavior in real time for dynamic collaborative control.

[0005] In the second aspect of the present application, a vehicle rollover prevention and stability control method is provided. The method includes: based on a multi-sensor fusion module, collecting driver behavior data in real time to obtain driving behavior dynamic information; according to a multi-source environment sensing module built in the vehicle, collecting driving environment information in real time, where the driving environment information includes weather environment data, ground condition data, and obstacle detection data; using the driving behavior dynamic information and the driving environment information to perform dual-channel independent risk prediction respectively, and performing risk fusion according to the independent prediction results to generate a driving risk prediction result; based on the driving risk prediction result, determining whether there is a rollover risk, and if so, generating a pre-driving intervention instruction; according to the pre-driving intervention instruction, performing pre-rollover risk intervention, and during the intervention execution process, providing real-time feedback on the driver's operation behavior to perform dynamic collaborative control.

[0006] One or more technical solutions provided in the present application have at least the following technical effects or advantages: The vehicle rollover prevention and stability control system and method provided in the present application relate to the field of intelligent control technology. By using a multi-sensor fusion module to collect driver behavior data and environmental information in real time, performing dual-channel independent risk prediction and fusing the results to generate a driving risk prediction, determining whether there is a rollover risk based on the risk prediction, and generating a pre-intervention instruction, risk control is achieved through pre-intervention. At the same time, real-time feedback on the driver's operation behavior is provided to perform dynamic collaborative control, solving the technical problem that the existing anti-rollover control of a three-wheeled vehicle lacks comprehensive monitoring of driver behavior and environmental information, resulting in the inability to timely and accurately control the rollover risk under complex driving conditions. The technical effect of providing accurate rollover risk warning and real-time vehicle control by fusing driving behavior and environmental data, performing dual-channel risk prediction and forward-looking dynamic intervention, and improving the robustness of vehicle control and driving safety is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0008] Figure 1 It is a schematic structural diagram of a vehicle rollover prevention and stability control system provided by an embodiment of the present application; Figure 2 It is a schematic flowchart of a vehicle rollover prevention and stability control method provided by an embodiment of the present application.

[0009] Description of reference numerals: Driving behavior dynamic monitoring module 10, driving environment dynamic monitoring module 20, driving risk prediction module 30, rollover intervention judgment module 40, pre-rollover intervention module 50. Detailed implementation manners

[0010] The present application provides a vehicle rollover prevention and stability control system and method, which are used to solve the technical problem that the existing anti-rollover control of a right three-wheeled vehicle lacks comprehensive monitoring of driver behavior and environmental information, resulting in the inability to timely and accurately control the rollover risk under complex driving conditions.

[0011] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0012] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned accompanying drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server including a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0013] Embodiment 1, as Figure 1 shown, the present application provides a vehicle rollover prevention and stability control system, and the system includes: A driving behavior dynamic monitoring module 10, which is used to collect driver behavior data in real time based on a multi-sensor fusion module to obtain driving behavior dynamic information.

[0014] Specifically, the driving behavior dynamic monitoring module 10 of the present application is used to monitor and collect data on the behavioral characteristics of the three-wheeled vehicle driver during driving in real time. Specifically, the driving behavior dynamic monitoring module 10 implements information collection based on a multi-sensor fusion module, wherein the multi-sensor fusion module includes but is not limited to an inertial measurement unit (IMU, InertialMeasurement Unit), a steering wheel angle sensor, a throttle opening sensor, a brake force sensor, and a seat posture detection sensor. The above-mentioned multiple sensors organically integrate various types of original sensor signals through fusion technology, that is, through time synchronization and data weighted fusion methods, so that the consistent extraction and dynamic follow-up of driver behavior data can be achieved under different sampling frequencies and measurement accuracies.

[0015] Through the multi-sensor fusion module, the driving behavior dynamic monitoring module 10 can collect key driving operation parameters such as steering angle change rate, throttle opening and closing degree change, brake application force, vehicle body tilt angle change, abnormal driving posture change, etc. in real time, and then form driving behavior dynamic information. The driving behavior dynamic information not only includes static state data, but also covers dynamic behavior characteristics such as acceleration, deceleration, sharp turns, and sudden braking, which are used for subsequent driving risk assessment and intervention decision support.

[0016] Preferably, the driving behavior dynamic monitoring module 10 is also equipped with a data preprocessing unit to filter, denoise and remove outliers on the collected raw sensor data to improve the accuracy and stability of driving behavior feature extraction. In this way, this module can accurately and timely reflect the changes in driver operation behavior under different road conditions and different driving styles, ensuring that the input data for subsequent rollover risk prediction and intervention judgment has high reliability and real-time performance.

[0017] The driving environment dynamic monitoring module 20 is used to collect driving environment information in real time based on the multi-source environment sensing module built into the vehicle. The driving environment information includes weather environment data, ground state data and obstacle detection data.

[0018] Optionally, the driving environment dynamic monitoring module 20 of the present application is used to perform real-time perception and data collection of the driving environment of the three-wheeled vehicle to provide environmental information support for subsequent risk assessment and intervention decisions.

[0019] Specifically, the driving environment dynamic monitoring module 20 operates relying on the multi-source environment sensing module built into the vehicle. The multi-source environment sensing module includes, but is not limited to, temperature and humidity sensors, light intensity sensors, rain detection sensors, road condition identification sensors, millimeter wave radars, lidars (LiDAR), and environmental cameras, etc. Through complementary sensing technology, that is, synchronizing the information collected by different types of sensors in time, extracting features, and performing multi-modal data fusion processing on the information collected by different types of sensors, the above multi-source environment sensing module can achieve high-precision and multi-dimensional monitoring of the external environmental state. Through this module, the system can collect in real time driving environment information including weather environment data (such as temperature, humidity, rainfall, light intensity, etc.), ground state data (such as the dryness and wetness of the road surface, icing conditions, muddy conditions, friction coefficient estimation, etc.), and obstacle detection data (such as the position, type, relative speed, and distance of obstacles ahead, etc.).

[0020] Preferably, the driving environment dynamic monitoring module 20 is also integrated with an environmental perception intelligent processing unit to extract features, detect anomalies, and filter signals from the original environmental data, so as to improve the accuracy and stability of the perception data. Through the above method, this module can reflect the changes in the vehicle's surrounding environment in real time and accurately under various complex meteorological conditions, bad road conditions, and dynamic obstacle environments, ensuring that the driving risk prediction module can carry out subsequent analysis based on comprehensive and highly reliable environmental input data.

[0021] The driving risk prediction module 30 is used to perform dual-channel independent risk prediction using the driving behavior dynamic information and the driving environment information respectively, and perform risk fusion according to the independent prediction results to generate a driving risk prediction result.

[0022] Furthermore, when performing dual-channel independent risk prediction, the driving risk prediction module 30 of the present application is also used to perform the following steps: P31: Build a rollover risk prediction network, which includes a driving behavior analysis and prediction channel and a driving environment risk prediction channel; P32: Based on the driving behavior analysis and prediction channel, perform driver behavior analysis and prediction on the driving behavior dynamic information to generate a predicted driving behavior risk; P33: Based on the driving environment risk prediction channel, perform environmental risk prediction on the driving environment information to generate a predicted driving environment risk; P34: Combine the predicted driving behavior risk and the predicted driving environment risk to perform dynamic risk coupling prediction to generate the driving risk prediction result.

[0023] It should be understood that the driving risk prediction module 30 of the present application is used to perform dual-channel independent risk prediction based on the driving behavior dynamic information and the driving environment information respectively, and perform risk fusion on this basis to generate a comprehensive driving risk prediction result.

[0024] Specifically, the driving risk prediction module 30 first constructs a rollover risk prediction network, which includes two parallel and independent risk prediction channels, namely, the driving behavior analysis and prediction channel and the driving environment risk prediction channel. Moreover, the rollover risk prediction network is a neural network structure that takes driving safety events (especially rollover risks) as the prediction target, fuses multi-modal data inputs, and performs feature learning and classification outputs. Preferably, this network can be implemented based on deep learning algorithms or traditional machine learning algorithms.

[0025] In the driving behavior analysis and prediction channel, I perform driver behavior analysis and prediction based on driving behavior dynamic information. By dynamically modeling and trend predicting feature quantities such as the driver's acceleration behavior, braking behavior, steering angle change rate, and vehicle body tilt trend, a predicted driving behavior risk of instability tendency caused by driving operations is generated. In addition, the predicted driving behavior risk refers to the risk level of the vehicle instability state (such as rollover caused by sharp turns) that the driver may trigger based on historical and current operation behaviors.

[0026] Meanwhile, in the driving environment risk prediction channel, the system performs environment risk prediction based on driving environment information. By analyzing weather conditions, ground friction status, the distribution of front obstacles, and their dynamic change characteristics, a predicted driving environment risk is generated, that is, an assessment of the possibility of vehicle out-of-control or rollover independently caused by environmental factors. For example, a slippery road surface leads to an extended braking distance, thus increasing the risk of losing control during turning.

[0027] After completing the two independent predictions, the driving risk prediction module 30 further combines the predicted driving behavior risk and the predicted driving environment risk, performs dynamic risk coupling prediction, and generates the final driving risk prediction result. Dynamic risk coupling prediction refers to dynamically evaluating the comprehensive rollover risk level according to the interaction and influence relationship between behavior risk and environment risk, through weighted fusion, feature cross-analysis, or other data fusion strategies, to ensure that the risk prediction result can comprehensively reflect the combined effect of the driver's operation intention and external environment changes.

[0028] In the above manner, the driving risk prediction module 30 can not only achieve the independent identification and prediction of driving behaviors and environmental factors, but also improve the accuracy and forward-looking of overall risk identification through dynamic fusion, providing a highly reliable decision-making basis for subsequent rollover intervention judgment and intervention execution.

[0029] Furthermore, when the driving risk prediction module 30 of the present application performs driver behavior analysis and prediction, it is also used to perform the following steps: P32-1: Extract key operation features based on the dynamic driving behavior information; P32-2: Perform time series modeling of operation data according to the key operation features to predict and generate a driving behavior trend portrait; P32-3: Identify abnormal driving behavior patterns and determine the driving behavior risk level according to the driving behavior trend portrait.

[0030] Optionally, when performing driver behavior analysis and prediction, the driving risk prediction module 30 in this application is further configured to perform the following refinement steps to improve the accuracy and timeliness of driving behavior risk assessment.

[0031] First, extract key operation features based on the dynamic driving behavior information. The key operation features refer to the driving action data parameters closely related to vehicle stability during driving, preferably including the acceleration change rate, braking intensity, steering wheel rotation angular velocity, vehicle body yaw rate, pitch angle change, roll angle change, etc. By extracting these key operation parameters from the dynamic driving behavior information, the characteristic changes of the driver during normal driving, sudden acceleration, sudden braking, sharp turning and other actions can be effectively captured, providing a basis for subsequent behavior trend modeling.

[0032] Next, perform time series modeling of operation data according to the key operation features to predict and generate a driving behavior trend portrait. Time series modeling refers to establishing a driving behavior evolution model based on the evolution law of key features in the time dimension, using statistical learning or deep sequence modeling techniques (such as long short-term memory network LSTM). Through time series modeling, the operation trend of the driver in the current and future periods can be predicted to generate a driving behavior trend portrait. The driving behavior trend portrait refers to a graphical and multi-dimensional parameter correlation method to describe the dynamic change trajectory and pattern of driving behavior, intuitively reflecting the continuity, stability and potential abnormal features of the driver's operation.

[0033] Furthermore, identify abnormal driving behavior patterns and determine the driving behavior risk level according to the driving behavior trend portrait. Abnormal driving behavior patterns refer to a set of operation behaviors that deviate from the normal driving behavior standard, such as frequent sudden acceleration, abnormal change in sharp turning angle, excessive braking or severe tilting operation, etc. By detecting the abnormality of the driving behavior trend portrait, it can be determined whether the driver has a high-risk operation tendency. Further, the system comprehensively evaluates and determines the driving behavior risk level according to factors such as the type, severity and occurrence frequency of the abnormal pattern, preferably divided into multiple levels of low risk, medium risk and high risk, to quantitatively characterize the potential impact of the driver's operation on vehicle stability.

[0034] Through the above refinement steps, the driving risk prediction module 30 can characterize the changing trend of driver behavior at a finer granularity, timely detect potential dangerous operations, and provide a more accurate basis for subsequent risk fusion and rollover intervention in terms of driving behavior risk input.

[0035] Furthermore, when the driving risk prediction module 30 of the present application performs dynamic risk coupling prediction by combining the predicted driving behavior risk and the predicted driving environment risk, it is further configured to perform the following steps: P34-1: Establish a risk coupling model, which is embedded with a risk coupling function for driving behavior risk and driving environment risk, and at the same time, is associated with a reinforcement learning component; P34-2: Through the reinforcement learning component, dynamically adjust the model parameters of the risk coupling model according to real-time driving behavior data and driving environment data; P34-3: Based on the adjusted risk coupling model, couple the predicted driving behavior risk and the predicted driving environment risk to generate a comprehensive risk prediction value; P34-4: According to the comprehensive risk prediction value, generate a driving risk prediction result and update the risk level in real time.

[0036] Specifically, when the driving risk prediction module 30 in the present application combines the predicted driving behavior risk and the predicted driving environment risk to perform dynamic risk coupling prediction, it further includes the following refinement steps to improve the accuracy and adaptability of comprehensive risk assessment.

[0037] First, establish a risk coupling model. The risk coupling model is a mathematical model used to describe the interaction relationship between driving behavior risk and driving environment risk. Preferably, the risk coupling model is embedded with a risk coupling function for driving behavior risk and driving environment risk. The risk coupling function defines the mechanism of mutual influence, superposition or non-linear change of different risk factors during the coupling process. Further, to improve the adaptive ability of the model, a reinforcement learning component is associated with the risk coupling model. The reinforcement learning component refers to a learning module that continuously optimizes the decision-making strategy and improves the prediction performance by interacting with the environment. Preferably, it can be implemented based on the Deep Reinforcement Learning framework.

[0038] Next, through the reinforcement learning component, dynamically adjust the model parameters of the risk coupling model according to the real-time collected driving behavior data and driving environment data. Specifically, the reinforcement learning component continuously corrects the weight coefficients, non-linear coupling terms and threshold parameters of the risk coupling function according to the driving operation characteristics, environmental state changes and historical prediction error feedback at the current moment, so that the risk coupling model can adapt to different driving styles, environmental conditions and real-time changing driving situations, and improve the accuracy and robustness of risk prediction.

[0039] Further, based on the dynamically adjusted risk coupling model, the system couples the predicted driving behavior risk and the predicted driving environment risk to generate a comprehensive risk prediction value. The comprehensive risk prediction value refers to the result of quantitatively evaluating the current and short-term future driving situation after integrating the driving operation risk and the environment risk, and can comprehensively reflect the instability or rollover possibility of the right three-wheeled vehicle under the comprehensive influence.

[0040] Finally, according to the comprehensive risk prediction value, a driving risk prediction result is generated and the risk level is updated in real time. Specifically, the system classifies the comprehensive risk prediction value according to a preset risk level classification standard (such as low risk, medium risk, high risk, extremely high risk), dynamically adjusts the risk level corresponding to the current driving state, and provides timely and accurate risk input for the subsequent rollover intervention judgment module to ensure that the anti-rollover control system can achieve pre-warning and intervention response.

[0041] Through the above refinement steps, the present invention can achieve efficient fusion prediction of driving behavior risk and environment risk, has real-time adaptive adjustment ability, significantly improves the accuracy, flexibility and robustness of driving risk prediction, and provides a solid foundation support for the anti-rollover stability control of the right three-wheeled vehicle.

[0042] Further, when performing dynamic risk coupling prediction, the driving risk prediction module 30 of the present application is further configured to perform the following steps: Introduce a time decay factor to weight the predicted values of the predicted driving behavior risk and the predicted driving environment risk, and preferentially extract high-impact features in the near future for risk coupling.

[0043] In a possible embodiment of the present application, when the driving risk prediction module 30 in the present application performs dynamic risk coupling prediction, the following steps are further included to optimize the response sensitivity to features of different time scales during the risk fusion process.

[0044] First, during the dynamic risk coupling process, the system introduces a time decay factor to weight the predicted values of the predicted driving behavior risk and the predicted driving environment risk. Specifically, the time decay factor refers to a weight factor used to characterize the decreasing law of the influence of historical data on the current decision-making. Preferably, it can be implemented by an exponential decay function, a sliding window decay function or an adaptive time decay strategy. By introducing the time decay factor, different weights can be assigned to the driving behavior risk values and environment risk values collected at each moment, and the weights gradually decrease as the time interval of the data increases, so as to preferentially highlight the feature data collected recently, which has high sensitivity and high correlation to the change of the current driving state.

[0045] In a specific implementation, the system gradually accumulates the risk prediction values over a past period of time based on a set time decay function, and uses the weighted recent high-impact features as the main input to participate in the calculation of the risk coupling model and the generation of the comprehensive risk prediction value. Through the above method, it can effectively avoid the interference of long-term data on the current risk assessment, and improve the sensitivity and response speed of the system to short-term high-risk events such as sudden changes in driver operations and rapid environmental changes.

[0046] In summary, by introducing a time decay factor and performing weighted processing, this application can further enhance the real-time performance and accuracy of dynamic risk coupling prediction, ensuring that the system can timely identify high-risk situations in a rapidly changing driving environment, and significantly improving the overall warning and protection effects of the vehicle rollover prevention and stability control system.

[0047] The rollover intervention judgment module 40 is used to judge whether there is a rollover risk based on the driving risk prediction result. If so, a pre-driving intervention instruction is generated.

[0048] Optionally, the rollover intervention judgment module 40 of this application is used to dynamically evaluate the stable state of the right three-wheeled vehicle based on the driving risk prediction result, and judge whether there is a rollover risk accordingly.

[0049] In a specific implementation process, the rollover intervention judgment module 40 performs discriminant analysis on the driving risk prediction result according to a set risk level threshold. Preferably, the system can preset multiple risk level intervals in advance, such as low risk, medium risk, high risk and extremely high risk intervals, and each level corresponds to a different intervention response strategy. When the driving risk prediction result reaches or exceeds the high risk and above levels, the rollover intervention judgment module 40 determines that there is a rollover risk currently and triggers the intervention decision-making process.

[0050] When it is detected that there is a rollover risk, the rollover intervention judgment module 40 generates a pre-driving intervention instruction. The pre-driving intervention instruction refers to a control instruction actively sent to the vehicle control system or the driver before the rollover risk actually occurs, including but not limited to intervention measures such as slowing down the acceleration response, moderately applying brakes, adjusting the vehicle body attitude control, restricting the steering angle change rate or providing driver warning prompts. The pre-driving intervention instruction aims to reduce the probability of the vehicle becoming unstable, skidding or overturning through early intervention, and achieve the control goal of preventing problems before they occur.

[0051] Through the above method, the rollover intervention judgment module 40 can achieve efficient identification and timely response to the vehicle rollover risk, ensure that the entire rollover prevention and stability control system has forward-looking and proactive intervention capabilities, and significantly improve the driving safety and stability of the right three-wheeled vehicle in a complex dynamic environment.

[0052] The pre - rollover intervention module 50 is used to perform pre - rollover risk intervention according to the pre - driving intervention instruction, and during the intervention execution process, it provides real - time feedback on the driver's operation behavior for dynamic collaborative control.

[0053] Furthermore, when the pre - rollover intervention module 50 of the present application performs pre - rollover risk intervention, it is also used to execute the following steps: P51: Identify the intervention level based on the pre - driving intervention instruction; P52: When the intervention level is low - risk, perform voice warning intervention; P53: When the intervention level is medium - risk, perform vehicle body self - stability adjustment intervention; P54: When the intervention level is high - risk, perform pre - vehicle braking intervention.

[0054] It should be understood that the pre - rollover intervention module 50 of the present application is used to perform pre - rollover risk intervention according to the pre - driving intervention instruction, and during the intervention execution process, it provides real - time feedback on the driver's actual operation behavior, thereby achieving dynamic collaborative control. Specifically, the pre - rollover intervention module 50 not only actively adjusts the vehicle control state, but also continuously collects the driver's response actions (such as steering correction, acceleration and deceleration adjustment) during the intervention for real - time data update and strategy coordination, ensuring an effective cooperation between the intervention measures and the driver's autonomous operation, so as to improve the smoothness and safety of the intervention.

[0055] Exemplarily, first, the intervention level is identified based on the pre - driving intervention instruction. Intervention level identification refers to classifying intervention actions into different levels of low - risk, medium - risk and high - risk according to the risk level corresponding to the driving risk prediction result, so as to match appropriate intervention means. Preferably, the intervention level identification module can be dynamically adjusted by combining multiple factors such as risk value range, change rate and environmental complexity.

[0056] When the intervention level is determined to be low - risk, the pre - rollover intervention module 50 performs voice warning intervention. Voice warning intervention means sending risk prompts to the driver through the in - vehicle voice broadcast system, such as prompting "Pay attention to lateral tilt" "Slow down and drive carefully", etc., to enhance the driver's perception of the current potential dangerous state and prompt him to take corrective measures actively, so as to reduce the risk without interfering with the vehicle control right.

[0057] When the intervention level is determined to be medium - risk, the pre - rollover intervention module 50 performs vehicle body self - stability adjustment intervention. Vehicle body self - stability adjustment means, on the basis of maintaining the driver's dominant control right, moderately intervening in the vehicle body dynamic posture by fine - tuning the electronic control system (such as Electronic Stability Control ESC), active tilt control system or rear - wheel differential control system to slow down the lateral tilt trend and improve the vehicle stability in the critical state.

[0058] When the intervention level is determined to be high risk, the pre-tipover intervention module 50 implements pre-vehicle braking intervention. Pre-vehicle braking intervention means that when an extremely high rollover risk is detected, the vehicle braking system is actively triggered. Preferably, braking can be applied only to specific wheels to achieve direction control, or the overall driving speed can be gradually reduced on the premise of ensuring driving safety, so as to quickly suppress the vehicle's dynamic instability trend and effectively prevent the occurrence of rollover accidents.

[0059] Through the above refinement steps, the pre-tipover intervention module 50 can accurately match the intervention strategy according to the risk level, and realize the dynamic feedback and collaborative adjustment of the driver's operation behavior during the intervention process, significantly improving the anti-rollover ability and active safety protection level of the three-wheeled vehicle under variable road conditions.

[0060] Further, when the intervention level is high risk, the pre-tipover intervention module 50 of the present application, when performing pre-vehicle braking intervention, is further used to perform the following steps: P54-1: Extract the comprehensive risk prediction result based on the pre-driving intervention instruction; P54-2: Use the comprehensive risk prediction result to perform an intervention rehearsal through the digital twin model to generate a rehearsal intervention strategy; P54-3: Generate a pre-driving intervention instruction according to the rehearsal intervention strategy, and the pre-driving intervention instruction includes multi-dimensional collaborative control indicators; P54-4: Based on the pre-driving intervention instruction, perform pre-vehicle braking intervention.

[0061] Specifically, when the intervention level is determined to be high risk, the pre-tipover intervention module 50 in the present application, when performing pre-vehicle braking intervention, further includes the following refinement steps to achieve the precision and adaptive optimization of the intervention action.

[0062] First, extract the comprehensive risk prediction result based on the pre-driving intervention instruction. The comprehensive risk prediction result refers to the quantitative risk assessment value formed by the combination of the driving behavior dynamic information and the driving environment information output by the driving risk prediction module 30 after risk coupling processing. Preferably, it includes multi-dimensional risk indicators such as the overall rollover probability, the lateral force change trend, and the vehicle center of gravity offset rate. By extracting the comprehensive risk prediction result, a basic decision-making basis can be provided for the subsequent formulation of the intervention strategy.

[0063] Next, using the comprehensive risk prediction results, intervention rehearsals are carried out through the digital twin model to generate rehearsal intervention strategies. The digital twin model refers to a simulation model that is built in real-time synchronization in a virtual environment and highly corresponds to the actual vehicle's dynamic behavior. Preferably, it includes a vehicle dynamics model, an environmental interaction model, and a driver operation simulation model. By performing pre-intervention action simulations and effect predictions based on the current risk situation in the digital twin environment, the system can generate multiple alternative intervention plans and screen out the rehearsal intervention strategy with the best effect according to the rehearsal results, thereby improving the rationality and accuracy of the intervention actions.

[0064] Furthermore, according to the rehearsal intervention strategy, a new pre-driving intervention instruction is generated, and the pre-driving intervention instruction includes multi-dimensional collaborative control indicators. The multi-dimensional collaborative control indicators refer to multiple control dimension parameters set to achieve efficient and stable intervention. Preferably, they include the braking application timing, braking force distribution ratio, wheel independent braking logic, steering correction assistance amplitude, and vehicle body stability adjustment parameters, etc. By setting multi-dimensional control indicators, it is possible to coordinate the linkage responses of braking, steering, and vehicle body attitude adjustment, ensuring that the intervention process can not only effectively suppress the rollover tendency but also minimize the negative impact on driving smoothness and safety to the greatest extent.

[0065] Finally, based on the updated pre-driving intervention instruction, a pre-emptive vehicle braking intervention is executed. Specifically, the system controls the braking execution unit (such as the ABS system, ESC system) to precisely apply the braking action according to the braking force distribution and timing scheme defined by the intervention instruction. Preferably, by independently braking specific wheels, the vehicle's yaw moment is controlled to quickly restore or maintain the vehicle's lateral stability and prevent rollover accidents.

[0066] Through the above refinement steps, the present application can dynamically optimize the pre-emptive vehicle braking intervention response through digital twin simulation rehearsals and multi-dimensional collaborative control strategies in high-risk situations, significantly improving the intelligence, accuracy, and stability of the intervention actions, thereby providing a higher level of active safety guarantee for the vehicle rollover prevention and stability control system.

[0067] Furthermore, the digital twin model in the pre-emptive rollover intervention module 50 of the present application includes: P54-21a: Vehicle dynamics simulation layer. Based on Trucksim, a vehicle multi-body dynamics model is constructed, tire parameters and mass distribution data are imported in real-time, and the dynamic response of the vehicle under different working conditions is simulated; P54-22a: Intervention strategy optimization layer. Perform 0.1-second-level simulations on differential braking, torque distribution, and active suspension adjustment instructions, and select the instruction with the largest reduction in roll angle as the output.

[0068] Optionally, the digital twin model in the pre - rollover intervention module 50 described in this application specifically includes the following two functional layers to achieve efficient and real - time pre - rehearsal of intervention strategies and optimization decisions: First, a vehicle dynamics simulation layer is set up. The vehicle dynamics simulation layer is mainly constructed based on the Trucksim simulation platform. Trucksim is a widely used vehicle dynamics simulation software in the industry, which can highly reproduce the multi - body dynamics characteristics of actual vehicles. In this application, the vehicle dynamics simulation layer constructs a multi - body dynamics model for a three - wheeled vehicle through Trucksim. The multi - body dynamics model includes key structural modules such as the frame, front and rear suspension systems, steering system, and double - rear - wheel assemblies, and real - time imports tire parameters (such as tire longitudinal / lateral force characteristics, friction coefficient curves) and vehicle mass distribution data (such as vehicle mass center position, load change distribution). Based on this simulation model, the system can simulate the vehicle's dynamic response behavior corresponding to different working conditions (such as different speeds, steering radii, and road adhesion coefficient changes) of the three - wheeled vehicle, especially the real - time simulation prediction of the body roll angle change, lateral acceleration change, and tire load transfer, so as to provide high - precision vehicle dynamic data support for the evaluation of intervention strategies.

[0069] Subsequently, an intervention strategy optimization layer is set up. The intervention strategy optimization layer performs rapid simulation and optimization evaluation for various combinations of intervention instructions on the basis of the vehicle dynamics simulation layer. Specifically, the intervention strategy optimization layer performs simulations with an ultra - short time step of 0.1 seconds for differential braking instructions (i.e., adjusting the yaw moment by the difference in braking forces between the left and right wheels), torque distribution instructions (i.e., adjusting the torque distribution ratio between the driving wheels), and active suspension adjustment instructions (i.e., adjusting the suspension stiffness or damping to suppress body roll). By quickly simulating and comparing the effects of reducing the vehicle roll angle of each combination of intervention instructions in a short time, the system automatically selects the intervention instruction with the largest reduction in roll angle within the simulation window as the output. Preferably, the reduction in roll angle refers to the amplitude of the decrease in the body roll angle per unit time, which is used as the core index to evaluate the quality of the intervention effect, so as to ensure that the intervention action can quickly suppress the vehicle overturning trend to the greatest extent.

[0070] Through the above structural design, the digital twin model of this application can complete vehicle dynamic simulation and intervention strategy optimization in an extremely short time, not only ensuring the real - time and accuracy of intervention decisions, but also greatly improving the consistency and executability of the intervention prediction results and the actual vehicle dynamic response by introducing a multi - body dynamics basic model with strong physical consistency, thus providing a solid and reliable technical support for pre - rollover intervention in high - risk rollover situations.

[0071] Furthermore, the system of this application also includes a periodic channel optimization module 60, which is used to perform the following steps: P61: Continuously monitor and record the driver behavior data and vehicle driving environment data, and extract the driver's driving habit characteristics and common driving environment characteristics; P62: Based on the driver's driving habit characteristics and common driving environment characteristics, perform periodic dual-channel parameter optimization.

[0072] In a possible embodiment of the present application, the system of the present application further includes a periodic channel optimization module 60, which is used to dynamically adjust and optimize the dual channels (i.e., the driving behavior analysis and prediction channel and the driving environment risk prediction channel) of the driving risk prediction module 30 based on the actual driving data, so as to continuously improve the overall prediction accuracy and adaptability of the system.

[0073] First, continuously monitor and record the driver behavior data and vehicle driving environment data. The driver behavior data includes, but is not limited to, acceleration mode, braking frequency, steering amplitude, number of sharp change operations, etc., which reflect the driver's operation habits and style characteristics; the vehicle driving environment data includes road type (such as urban road, highway, rural road), meteorological conditions (such as sunny day, rain, snow weather), change of road surface adhesion coefficient, etc., which reflect the environmental characteristics of the vehicle's normal driving. Through long-term accumulation and analysis of the above multi-source data, the system extracts the driver's driving habit characteristics and common driving environment characteristics. Driving habit characteristics refer to the stable behavior patterns formed by the driver during long-term driving, such as preferring aggressive driving or gentle driving; common driving environment characteristics refer to the environmental background where the vehicle is usually located, such as often driving on low adhesion coefficient roads or urban congestion sections.

[0074] Next, based on the extracted driver's driving habit characteristics and common driving environment characteristics, perform periodic dual-channel parameter optimization. Specifically, for the driving behavior analysis and prediction channel, the system dynamically adjusts the behavior feature extraction weight, abnormal behavior determination threshold, and time series modeling parameters according to the driver's habit characteristics, so that the driving behavior risk assessment is more in line with the individual driving style characteristics; for the driving environment risk prediction channel, the system adjusts the feature extraction strategy, risk mapping model, and fusion weighting coefficient of the environment perception model according to the common driving environment characteristics, so as to improve the regional adaptability and situation sensitivity of the environment risk identification. Preferably, the periodic optimization can be set with a fixed time period (such as every 500 kilometers of driving or once every two weeks) or event-driven (such as triggering optimization after detecting a significant change in driving style).

[0075] Through the above method, the periodic channel optimization module 60 can dynamically fine-tune and adaptively optimize the risk prediction model according to the actual usage data, ensuring that the system continuously maintains a high level of risk prediction accuracy and response sensitivity during long-term operation, thereby further improving the intelligent level and practicality of the vehicle rollover prevention and stability control system.

[0076] In summary, the embodiments of the present application have at least the following technical effects: In the present application, the driving behavior dynamic monitoring module collects driver behavior data in real time, the driving environment dynamic monitoring module collects weather, ground condition, and obstacle data, and the driving risk prediction module performs dual-channel independent risk prediction and fusion to generate a comprehensive risk prediction result. Based on this result, the rollover intervention judgment module determines whether there is a rollover risk. If so, a pre-intervention instruction is generated, and finally, the pre-rollover intervention module executes risk intervention and provides real-time feedback on the driver's operations for dynamic collaborative control.

[0077] It achieves the technical effect of providing accurate rollover risk warnings and real-time interventions by fusing driving behavior and environmental data for dual-channel risk prediction and prospective dynamic intervention, improving the robustness of vehicle control, and effectively enhancing the driving safety of the vehicle.

[0078] Embodiment 2 is based on the same inventive concept as the vehicle rollover prevention and stability control system in the foregoing embodiment. As Figure 2 shown, the present application provides a vehicle rollover prevention and stability control system. The method in the embodiments of the present application and the system embodiments are based on the same inventive concept. Among them, the method includes: Based on a multi-sensor fusion module, collect driver behavior data in real time to obtain driving behavior dynamic information; according to the multi-source environment sensor module built in the vehicle, collect driving environment information in real time, where the driving environment information includes weather environment data, ground condition data, and obstacle detection data; use the driving behavior dynamic information and the driving environment information to perform dual-channel independent risk prediction respectively, and perform risk fusion according to the independent prediction results to generate a driving risk prediction result; based on the driving risk prediction result, determine whether there is a rollover risk. If so, generate a pre-driving intervention instruction; according to the pre-driving intervention instruction, perform pre-rollover risk intervention, and during the intervention execution process, provide real-time feedback on the driver's operation behavior for dynamic collaborative control.

[0079] Further, using the driving behavior dynamic information and the driving environment information to perform dual-channel independent risk prediction respectively includes: Build a rollover risk prediction network, which includes a driving behavior analysis and prediction channel and a driving environment risk prediction channel; based on the driving behavior analysis and prediction channel, perform driver behavior analysis and prediction on the driving behavior dynamic information to generate a predicted driving behavior risk; based on the driving environment risk prediction channel, perform environmental risk prediction on the driving environment information to generate a predicted driving environment risk; combine the predicted driving behavior risk and the predicted driving environment risk for dynamic risk coupling prediction to generate the driving risk prediction result.

[0080] Further, based on the driving behavior analysis and prediction channel, driver behavior analysis and prediction are performed on the driving behavior dynamic information to generate a predicted driving behavior risk, including: Based on the driving behavior dynamic information, key operation features are extracted; according to the key operation features, time series modeling of operation data is performed to predict and generate a driving behavior trend portrait; according to the driving behavior trend portrait, abnormal driving behavior patterns are identified, and the driving behavior risk level is determined.

[0081] Further, combined with the predicted driving behavior risk and the predicted driving environment risk, dynamic risk coupling prediction is performed to generate the driving risk prediction result, including: A risk coupling model is established. The risk coupling model is embedded with a risk coupling function for driving behavior risk and driving environment risk, and at the same time, a reinforcement learning component is associated; through the reinforcement learning component, the model parameters of the risk coupling model are dynamically adjusted according to real-time driving behavior data and driving environment data; based on the adjusted risk coupling model, the predicted driving behavior risk and the predicted driving environment risk are coupled to generate a comprehensive risk prediction value; according to the comprehensive risk prediction value, a driving risk prediction result is generated, and the risk level is updated in real time. Further, a time decay factor is introduced to perform weighted processing on the predicted values of the predicted driving behavior risk and the predicted driving environment risk, and high-impact features in the near future are preferentially extracted for risk coupling.

[0082] Further, according to the pre-driving intervention instruction, pre-tipping risk intervention is performed, including: Based on the pre-driving intervention instruction, intervention level identification is performed; when the intervention level is low risk, voice warning intervention is performed; when the intervention level is medium risk, vehicle body self-stabilization adjustment intervention is performed; when the intervention level is high risk, pre-vehicle braking intervention is performed.

[0083] Further, when the intervention level is high risk, pre-vehicle braking intervention is performed, including: Based on the pre-driving intervention instruction, a comprehensive risk prediction result is extracted; using the comprehensive risk prediction result, intervention pre-play is performed through a digital twin model to generate a pre-play intervention strategy; according to the pre-play intervention strategy, a pre-driving intervention instruction is generated, and the pre-driving intervention instruction includes multi-dimensional collaborative control indicators; based on the pre-driving intervention instruction, pre-vehicle braking intervention is executed.

[0084] Further, the digital twin model includes: The vehicle dynamics simulation layer constructs a vehicle multi-body dynamics model based on Trucksim, imports tire parameters and mass distribution data in real time, and simulates the dynamic response of the vehicle under different working conditions; the intervention strategy optimization layer performs a simulation at the 0.1-second level on differential braking, torque distribution, and active suspension adjustment commands, and selects the command with the largest reduction in roll angle as the output.

[0085] Further, the method further includes: Continuously monitor and record driver behavior data and vehicle driving environment data, and extract driver driving habit characteristics and common driving environment characteristics; based on the driver driving habit characteristics and common driving environment characteristics, perform periodic two-channel parameter optimization.

[0086] It should be noted that the above order of the embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. Additionally, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0087] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0088] This specification and the drawings are only exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A vehicle rollover prevention and stability control system, characterized in that, The system includes: A driving behavior dynamic monitoring module, which is used to collect driver behavior data in real time based on a multi-sensor fusion module to obtain driving behavior dynamic information; A driving environment dynamic monitoring module, which is used to collect driving environment information in real time according to a multi-source environment sensor module built in the vehicle, and the driving environment information includes weather environment data, ground state data, and obstacle detection data; A driving risk prediction module, which is used to perform dual-channel independent risk prediction using the driving behavior dynamic information and the driving environment information respectively, and perform risk fusion based on the independent prediction results to generate a driving risk prediction result; A rollover intervention judgment module, which is used to judge whether there is a rollover risk based on the driving risk prediction result. If so, a pre-driving intervention instruction is generated; A pre-rollover intervention module, which is used to perform pre-rollover risk intervention according to the pre-driving intervention instruction, and during the intervention execution process, feedback the driver's operation behavior in real time for dynamic collaborative control.

2. The vehicle rollover prevention and stability control system according to claim 1, wherein When performing dual-channel independent risk prediction, the driving risk prediction module is further used for: Building a rollover risk prediction network, which includes a driving behavior analysis and prediction channel and a driving environment risk prediction channel; Based on the driving behavior analysis and prediction channel, analyzing and predicting the driver's behavior for the driving behavior dynamic information to generate a predicted driving behavior risk; Based on the driving environment risk prediction channel, predicting the environmental risk for the driving environment information to generate a predicted driving environment risk; Combining the predicted driving behavior risk and the predicted driving environment risk for dynamic risk coupling prediction to generate the driving risk prediction result.

3. The vehicle rollover prevention and stability control system according to claim 2, characterized in that, When performing driver behavior analysis and prediction, the driving risk prediction module is further used for: Extracting key operation features based on the driving behavior dynamic information; Performing time series modeling of operation data according to the key operation features to predict and generate a driving behavior trend portrait; Identifying abnormal driving behavior patterns according to the driving behavior trend portrait and determining the driving behavior risk level.

4. The vehicle rollover prevention and stability control system according to claim 2, wherein, When combining the predicted driving behavior risk and the predicted driving environment risk for dynamic risk coupling prediction, the driving risk prediction module is further used for: Establishing a risk coupling model, which is embedded with a risk coupling function for driving behavior risk and driving environment risk, and at the same time, is associated with a reinforcement learning component; Through the reinforcement learning component, dynamically adjusting the model parameters of the risk coupling model according to real-time driving behavior data and driving environment data; Based on the adjusted risk coupling model, coupling the predicted driving behavior risk and the predicted driving environment risk to generate a comprehensive risk prediction value; Generating a driving risk prediction result according to the comprehensive risk prediction value and updating the risk level in real time.

5. The vehicle rollover prevention and stability control system according to claim 4, characterized in that, When performing dynamic risk coupling prediction, the driving risk prediction module is further used for: Introducing a time decay factor to perform weighted processing on the predicted values of the predicted driving behavior risk and the predicted driving environment risk, and preferentially extracting recent high-impact features for risk coupling.

6. The vehicle rollover prevention and stability control system according to claim 1, characterized in that, When performing pre-rollover risk intervention, the pre-rollover intervention module is further configured to: Identify the intervention level based on the pre-driving intervention instruction; When the intervention level is low risk, perform voice warning intervention; When the intervention level is medium risk, perform vehicle self-stabilization adjustment intervention; When the intervention level is high risk, perform pre-vehicle braking intervention.

7. The vehicle rollover prevention and stability control system according to claim 6, characterized in that, When the intervention level is high risk, when the pre-rollover intervention module performs pre-vehicle braking intervention, it is further configured to: Extract the comprehensive risk prediction result based on the pre-driving intervention instruction; Use the comprehensive risk prediction result to perform intervention rehearsal through the digital twin model to generate a rehearsal intervention strategy; Generate a pre-driving intervention instruction according to the rehearsal intervention strategy, and the pre-driving intervention instruction includes multi-dimensional collaborative control indicators; Based on the pre-driving intervention instruction, perform pre-vehicle braking intervention.

8. The vehicle rollover prevention and stability control system according to claim 7, wherein, The digital twin model in the pre-rollover intervention module includes: The vehicle dynamics simulation layer constructs a vehicle multi-body dynamics model based on Trucksim, imports tire parameters and mass distribution data in real time, and simulates the dynamic response of the vehicle under different working conditions; The intervention strategy optimization layer performs 0.1-second-level simulation on differential braking, torque distribution, and active suspension adjustment instructions, and selects the instruction with the largest reduction in roll angle as the output.

9. The vehicle rollover prevention and stability control system according to claim 1, characterized in that, The system further includes a periodic channel optimization module for: Continuously monitor and record driver behavior data and vehicle driving environment data, and extract driver driving habit characteristics and common driving environment characteristics; Based on the driver driving habit characteristics and common driving environment characteristics, perform periodic two-channel parameter optimization.

10. A vehicle rollover prevention and stability control method, characterized in that, The method is implemented by the vehicle rollover prevention and stability control system according to any one of claims 1-9, and the method includes: Based on the multi-sensor fusion module, collect driver behavior data in real time to obtain driving behavior dynamic information; According to the multi-source environment sensing module built in the vehicle, collect driving environment information in real time, and the driving environment information includes weather environment data, ground state data, and obstacle detection data; Use the driving behavior dynamic information and the driving environment information to perform two-channel independent risk prediction respectively, and perform risk fusion according to the independent prediction results to generate a driving risk prediction result; Based on the driving risk prediction result, determine whether there is a rollover risk. If so, generate a pre-driving intervention instruction; According to the pre-driving intervention instruction, perform pre-rollover risk intervention, and during the intervention execution process, feedback the driver's operation behavior in real time to perform dynamic collaborative control.

Images