Multi-mode communication control method for intelligent driving automobile

Abstract

The invention provides a multi-mode communication control method for an intelligent driving automobile, which comprises the following steps of: acquiring key vehicle dynamic parameters, including steering, braking and power system state parameters, through a vehicle-mounted sensing terminal; the parameters are transmitted to a control center and are stored and analyzed by using a cloud database. The control center processes the data by using a machine learning algorithm and calculates a real-time risk coefficient Rc. And the safety threshold Rth is determined through a steering dynamic coefficient, a braking dynamic coefficient, a power system risk coefficient and a scene adaptive weight. When the risk coefficient calculated in real time exceeds a safety threshold value, the system triggers risk early warning, gives an alarm to a driver through the HUD module, and meanwhile generates a vehicle control instruction to be displayed in the HUD module so as to reduce the risk. According to the method, the safety and reliability of the intelligent automobile under various driving conditions are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent driving vehicles, and in particular to a multimodal communication control method for intelligent driving vehicles. Background Art

[0002] With the rapid development of artificial intelligence and sensor technology, intelligent driving vehicles have gradually become an important development direction in the automotive industry. During the intelligent driving process, the vehicle needs to continuously perceive its own state and external environment information, and make accurate decisions and controls based on this information to ensure driving safety and efficiency. Therefore, an efficient and reliable communication control method is crucial for intelligent driving vehicles.

[0003] Most of the existing communication control methods for intelligent driving vehicles focus on single-modal information processing, such as only relying on the vehicle operation parameters collected by on-vehicle sensors, or only making decisions based on external environment perception data. This single-modal control method has obvious deficiencies and cannot comprehensively evaluate the driving state of the vehicle, making it difficult to handle complex and changeable actual driving scenarios. For example, when abnormalities occur in the vehicle steering, braking, or power system, due to the lack of a comprehensive analysis and dynamic evaluation mechanism for multiple parameters, it is often impossible to accurately judge risks in a timely manner, which is likely to lead to accidents.

[0004] Although some multimodal control methods consider multiple information sources, in the process of data processing and risk assessment, they lack an accurate grasp of the dynamic characteristics of the vehicle in different scenarios. The set safety thresholds are usually fixed or only adjusted according to simple rules, and cannot be adaptively optimized according to actual scenarios such as road types and weather conditions, resulting in low accuracy and reliability of risk assessment. At the same time, during the vehicle operation, it is also difficult to dynamically learn and optimize the safety thresholds based on actual feedback, resulting in insufficient adaptability and stability of the system.

[0005] Therefore, there is an urgent need for a communication control method for intelligent driving vehicles that can fuse multimodal information, adaptively adjust the safety thresholds according to different scenarios, and continuously optimize the control strategy through online learning to improve the safety and reliability of intelligent driving vehicles. Summary of the Invention

[0006] In view of this, the present invention provides a multimodal communication control method for intelligent driving vehicles. The vehicle dynamic parameters are collected by an on-vehicle sensing terminal and stored in the cloud. The control center calculates the relevant coefficients of the steering, braking, and power systems based on this, and sets the safety threshold Rth. The vehicle dynamic parameters are collected in real time to calculate the risk index Rc, which is compared with the threshold to judge the driving state. When the threshold is exceeded, an alarm is issued and the threshold is optimized, realizing multimodal data fusion and dynamic risk control, and improving the safety and reliability of intelligent driving, effectively solving the problems raised in the background art.

[0007] To achieve the above object, the present invention provides the following technical solutions: S1. The vehicle-mounted sensing terminal collects the vehicle dynamic parameters of the same series of vehicles and stores these vehicle dynamic parameter data in the cloud database module. The vehicle dynamic parameter data of each vehicle constructs an independent data subset, forming a data set; S2. The control center extracts the above vehicle dynamic parameters in a single data subset from the data set as the training set. The vehicle dynamic parameters include steering dynamic parameters, braking dynamic parameters, and power system state parameters; S3. The control center processes the extracted steering dynamic parameters to obtain the steering dynamic coefficient, processes the braking dynamic parameters to obtain the braking dynamic coefficient, processes the power system state parameters to obtain the power system risk coefficient, and sets the safety threshold Rth through scene adaptive weight calculation; S4. The vehicle-mounted sensing terminal collects the current vehicle dynamic parameters in real time and transmits the current vehicle dynamic parameter data to the control center in real time through the real-time communication module. After receiving the current vehicle dynamic parameter data, the control center performs preprocessing and conducts a comprehensive evaluation based on multiple risk factors to calculate the risk index Rc; S5. The control center compares the calculated risk index Rc with the preset safety threshold Rth to determine whether the driving state of the vehicle exceeds the safety threshold Rth; S6. If the driving state of the vehicle does not exceed the safety threshold Rth, the current driving state is maintained without intervention; S7. If the driving state of the vehicle exceeds the safety threshold Rth, while the vehicle control system sends a warning message to the HUD module and emits a sound to remind the vehicle owner according to the control instruction, the vehicle control system records and feeds back the execution result to the control center. The control center dynamically optimizes the safety threshold Rth through online learning based on these recorded and fed-back data, and at the same time updates the warning prompt of the HUD.

[0008] The technical effects and advantages of the present invention: By real-time monitoring of vehicle dynamic parameters such as steering dynamic parameters, braking dynamic parameters, and power system state parameters, and conducting risk assessment based on these parameters, the present invention can significantly improve the driving safety of intelligent driving vehicles. By accurately calculating the risk coefficient Rc and comparing it with the safety threshold Rth, the system can timely identify potential risks and take corresponding measures to avoid accidents; The present invention can not only identify and respond to the current risk situation, but also learn from historical data through machine learning algorithms, continuously optimize the risk assessment model, improve the accuracy and adaptability of risk management. By dynamically adjusting the safety threshold Rth and model parameters, the system can adapt to different driving conditions and environments, achieving more refined risk control; Through real-time risk assessment and control, the present invention can reduce unnecessary emergency braking or lane changes, provide a smoother driving experience, and enhance the comfort of passengers. At the same time, through the HUD module, risk information and control instructions are timely fed back to the driver, enhancing the driver's understanding of the vehicle state and improving driving confidence and satisfaction. Through precise risk assessment and control, the present invention helps to reduce vehicle maintenance and repair costs caused by accidents or wrong decisions. The optimized control strategy can also reduce fuel consumption and vehicle wear, thereby reducing long-term operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a schematic diagram of the overall structure of the present invention.

[0010] Figure 2 It is a flowchart of the overall structure of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

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

[0012] As shown in the attached Figure 1 A multimodal communication control method for an intelligent driving vehicle, including: an in-vehicle perception terminal, a vehicle control system, a control center, a cloud database module, and a HUD module; The in-vehicle perception terminal, as the core device for an intelligent driving vehicle to achieve environmental perception and data collection, integrates a variety of high-precision sensors including meteorological sensors, in-vehicle cameras, etc., and real-time collects the vehicle's own dynamic parameters and external environment information, and processes the data into a standardized format, laying a foundation data support for the entire multimodal communication control system; The vehicle control system is the execution center for an intelligent driving vehicle to achieve precise control and safe driving. It strictly implements real-time and precise control of key systems such as vehicle power output, braking response, and steering operation according to the instructions issued by the control center, ensuring that the vehicle always runs stably according to the predetermined strategy, and at the same time has a perfect fault emergency handling mechanism, providing a solid guarantee for driving safety; The control center is the core hub of the multimodal communication control system of an intelligent driving vehicle, integrating functions such as data processing, analysis and decision-making, and instruction scheduling. Through in-depth integration and efficient operation of various types of information, it outputs scientific decision-making support for the safe and efficient operation of the vehicle, and coordinates the orderly cooperation of each terminal and module; As the data foundation of the multi-modal communication control system of intelligent driving vehicles, the cloud database module undertakes the tasks of storing, managing, invoking, and sharing massive amounts of data, providing rich data support for the control center to build a risk assessment model, realize real-time vehicle risk monitoring, and dynamic decision-making, and ensuring the stable and efficient operation of the entire system; The HUD module is a key device for the intelligent driving vehicle to achieve efficient human-machine interaction. It clearly projects important driving information in front of the driver's line of sight, avoiding potential safety hazards caused by line-of-sight transfer, and assisting the driver in decision-making in an intuitive and clear visual way. At the same time, this module shoulders the heavy responsibility of risk warning, which can significantly improve driving safety and driving experience.

[0013] As shown in the appendix Figure 2 A multi-modal communication control method for intelligent driving vehicles is shown as follows. The specific implementation steps include the following: S1. The on-vehicle perception terminal collects the vehicle dynamic parameters of the same series of vehicles and stores these vehicle dynamic parameter data in the cloud database module. The vehicle dynamic parameter data of each vehicle constructs an independent data subset, forming a data set.

[0014] In this embodiment, it is necessary to specifically explain the specific application of vehicle dynamic parameter data in multi-modal communication control. Vehicle dynamic parameter data represents the specific situation of the vehicle. Based on these situations, the vehicles are classified. The vehicle classification can be based on different ranges of vehicle dynamic parameters. For example, the steering command response delay can be divided into categories such as less than 0.5 seconds, 0.5 to 1 second, and more than 1 second.

[0015] It should be further noted that the on-vehicle perception terminal reaches the control center through the real-time communication module, supporting the real-time transmission of data to the cloud database module. Through the communication setting function integration unit of the on-vehicle terminal, such as the "data upload" or "synchronization" function, the on-vehicle perception terminal is guided to upload dynamic parameter data, where the uploaded data includes real-time collected vehicle dynamic parameters and historical collected data. It should be explained that before collecting vehicle dynamic parameters, it is necessary to clearly inform the data collection scope and purpose in the on-vehicle system. For example, an authorization pop-up window is displayed through the in-vehicle infotainment system to request the driver's consent for data collection, so as to reduce the risk of privacy leakage and improve privacy protection.

[0016] S2. The control center extracts the above vehicle dynamic parameters from a single data subset in the data set as a training set, and the vehicle dynamic parameters include steering dynamic parameters, braking dynamic parameters, and power system state parameters.

[0017] In this embodiment, it should be specifically explained that the control center extracts vehicle dynamic parameter data from the cloud database module as the training set. The steering dynamic parameters include: the time interval tr from when the driver issues a steering command to when the vehicle actually executes the steering action. This parameter directly reflects the response sensitivity of the steering system to the operation command. The shorter tr is, the more timely the steering system can respond to the driving intention, ensuring the timeliness and safety of operations such as lane changes and turning. The time period Te for the vehicle to complete a full steering action reflects the working efficiency of the steering system. The shorter Te is, the more obvious the handling advantage of the vehicle when steering on a narrow road or making an emergency avoidance. The braking dynamic parameters include: the time interval td from when the brake pedal is pressed to when the braking system actually responds. Its length directly affects the starting time of emergency braking. The smaller td is, the faster the braking can be started at a dangerous moment, shortening the braking distance. The brake pedal pressure build-up time tp reflects how quickly the braking system generates effective braking force. The shorter tp is, the more quickly sufficient braking force can be established, improving the braking efficiency. The efficiency η of the braking system under actual driving conditions is the ratio of the actual braking distance to the theoretical braking distance calculated based on vehicle speed, road surface friction coefficient, and slope. The higher η is, the more fully the braking system can perform its function under actual road conditions, ensuring safe stopping. The power system state parameters include: the engine output efficiency ϵ, which represents the efficiency level of the engine in converting fuel energy into mechanical energy. The higher ϵ is, the more efficient the fuel utilization, and the better the vehicle's power performance and economy. The engine coolant temperature T is a key indicator for measuring the working thermal state of the engine. An appropriate T can ensure that the engine is in the optimal working temperature range, avoiding failures caused by overheating and maintaining the stability and reliability of power output.

[0018] S3. The control center processes the extracted steering dynamic parameters to obtain the steering dynamic coefficient, processes the braking dynamic parameters to obtain the braking dynamic coefficient, processes the power system state parameters to obtain the power system risk coefficient, and combines the above coefficients to set the safety threshold Rth through scene adaptive weight calculation.

[0019] In this embodiment, the complete process of the control center setting the safety threshold Rth needs to be explained. For the steering dynamic parameters, according to the formula , the steering dynamic coefficient is calculated. c1 is the steering dynamic coefficient, trmax is the maximum historical steering delay time of the same vehicle model, and Temax is the maximum steering period allowed by the vehicle design. This formula converts the parameters related to the steering response time into the steering dynamic coefficient. The larger the coefficient, the more sluggish the steering system response and the higher the potential risk. For the braking dynamic parameters, the formula , c2 is the braking dynamic coefficient, tdnominal and tpnominal are nominal values, η is the ratio of actual braking distance to theoretical braking distance, w1, w2, w3 are dynamic weights dynamically assigned through the correlation analysis between real-time vehicle speed and braking pressure. This formula comprehensively considers the braking response time and braking efficiency. The higher the coefficient, the worse the braking system performance. For the power system state parameters, according to the formula , calculate the power system risk coefficient c3, and finally, the control center combines the above three coefficients through the formula , calculate the safety threshold Rth, where α, β, and γ are scene adaptive weight coefficients. In highway scenarios, due to the high vehicle speed, the weights of the steering dynamic coefficient and the braking dynamic coefficient will be increased; in rainy days, considering the impact of slippery road surface on braking, the weight of the braking dynamic coefficient will be increased to ensure that the safety threshold meets the safety requirements of different scenarios.

[0020] S4. The vehicle-mounted sensing terminal collects the current vehicle dynamic parameters in real time, and transmits the current vehicle dynamic parameter data to the control center in real time through the real-time communication module. After receiving the current vehicle dynamic parameter data, the control center performs preprocessing and performs a comprehensive evaluation based on multiple risk factors to calculate the risk index Rc; In this embodiment, the dynamic parameters collected by the vehicle-mounted sensing terminal include, but are not limited to, the time interval tr from the driver issuing a steering command to the vehicle actually starting the steering action, the time period Te required for the vehicle to complete a complete steering action, the time interval td from the vehicle's brake pedal being stepped on to the brake system responding, the brake pedal pressure build-up time tp, the efficiency η of the brake system under actual driving conditions, the engine output efficiency ϵ, the temperature T of the engine coolant, etc. These parameters are sampled by the vehicle-mounted sensing terminal and packaged into a standardized data format.

[0021] It is necessary to explain the process of data transmission and preprocessing to the control center. The vehicle-mounted sensing terminal transmits the collected data to the control center quickly and stably through the real-time communication module. After receiving the data, the control center first performs data verification, and ensures that the data has not been tampered with during the transmission process through hash verification and other methods; then performs outlier detection to eliminate erroneous data caused by sensor failure and other reasons; finally, the data is normalized, and different parameters are uniformly mapped to the appropriate numerical range for subsequent analysis and calculation.

[0022] What needs to be emphasized is the application and subsequent processing of the real-time risk assessment results. The control center compares the calculated risk coefficient Rc with the pre-set safety threshold Rth. When Rc ≥ Rth, the system immediately initiates the corresponding instruction measures, such as reminding the driver to adjust the power output to limit the speed, issuing early warning prompts to the driver, etc. In addition, the control center will archive the data and processing results of each real-time risk assessment, conduct in-depth analysis of these data on a regular basis, explore potential risk patterns and laws, dynamically adjust and optimize the safety threshold Rth, and continuously improve the accuracy and effectiveness of real-time risk assessment to better ensure vehicle driving safety.

[0023] S5. The control center compares the calculated risk index Rc with the preset safety threshold Rth to determine whether the driving state of the vehicle exceeds the safety threshold Rth.

[0024] In this embodiment, it is necessary to specifically explain the complete process of the control center based on the comparison between the risk coefficient and the safety threshold: the control center, as the decision-making core of the intelligent driving car, receives the risk coefficient Rc calculated from S4 in real time and compares it with the pre-set safety threshold Rth. Based on this comparison result, the control center will conduct risk judgment and generate corresponding control instructions to ensure the driving safety and stability of the vehicle.

[0025] What needs to be further explained is the specific method and decision logic for comparing the risk coefficient with the safety threshold. The risk coefficient Rc is calculated based on multi-dimensional risk factors such as the real-time dynamic parameters of the vehicle and the driving environment, and accurately reflects the current risk level of the vehicle; the safety threshold Rth is the safety baseline set by the control center based on a large amount of historical data, combined with factors such as vehicle type, road conditions, and weather conditions, after optimization through machine learning algorithms. The control center directly compares Rc and Rth by comparing the numerical values. This simple and intuitive comparison method provides a clear and definite basis for judgment for subsequent risk management decisions.

[0026] S6. If the vehicle's driving state does not exceed the safety threshold Rth, the current driving state is maintained without intervention. In this embodiment, it should be specifically explained that the control center, as the core decision-making unit of the intelligent driving car, completes the comparison and judgment of the risk index Rc and the safety threshold Rth. If it concludes that Rc≤Rth, that is, it determines that the current vehicle driving state is within a safe and controllable range, then the "non-intervention maintenance" strategy is triggered: the control center sends a "maintain current state" instruction to the vehicle control system, and the vehicle continues to operate stably according to the existing power output, steering angle, driving speed and other parameters, while maintaining continuous monitoring of real-time data.

[0027] It should be further noted that the safety threshold Rth is the safety benchmark dynamically adjusted by the control center based on historical data training and combined with real-time scenarios. When Rc ≤ Rth, it indicates that the comprehensive evaluation result of the vehicle's real-time dynamic parameters and environmental risk factors does not exceed the safety boundary, and the collaborative work of the vehicle's steering, braking, and power systems is within the design expectation range. Without additional intervention, driving safety can be ensured. This decision depends on the scenario adaptive safety model constructed in step S3 and the risk index calculated in real time in step S4. The two form a precise quantitative evaluation of the driving state through multi-modal data fusion, providing a scientific basis for the "no intervention" decision.

[0028] It should be explained that "no intervention" does not mean completely stopping monitoring. The in-vehicle sensing terminal still collects data in real time and transmits it to the control center. The control center updates the risk index Rc at a constant frequency to ensure the instant response ability to sudden risks, forming a dynamic balance between safety and efficiency.

[0029] S7. If the driving state of the vehicle exceeds the safety threshold Rth, while the vehicle control system sends a warning message to the HUD module according to the control instruction and emits a sound to remind the vehicle owner, the vehicle control system records and feeds back the execution result to the control center. The control center dynamically optimizes the safety threshold Rth through online learning based on the recorded and fed-back data, and at the same time updates the warning prompt of the HUD. In this embodiment, it should be specifically noted that the specific process and logic of control execution and feedback optimization are as follows: The vehicle control system receives control instructions from the control center, including operations such as deceleration, braking, lane change, or parking, and displays them in the HUD module to remind the driver; after the driver executes the control instruction, the execution result is recorded, and the vehicle control system feeds back the execution result to the control center for further analysis and optimization; the control center uses the fed-back data to dynamically optimize the safety threshold Rth through an online learning algorithm; the control center updates the warning prompt displayed on the HUD according to the optimized safety threshold Rth to provide the driver with the latest risk assessment result; It should be noted that through this step, the intelligent driving vehicle can dynamically generate and execute control instructions according to the real-time evaluated risk level, ensuring the safety and reliability of vehicle driving. At the same time, through the feedback and online learning mechanisms, the system can continuously optimize the control strategy, improve the performance of intelligent driving. The optimized weight parameters of the logistic regression model and the safety threshold Rth will be applied to subsequent real-time risk assessments, making the system's risk judgment more accurate and the control instruction generation more reasonable.

[0030] A multi-modal communication control medium for an intelligent driving vehicle, which stores executable program code that, when executed by a processor, implements the above-mentioned multi-modal communication control method for an intelligent driving vehicle.

[0031] Second: In the attached drawings of the disclosed embodiments of the present invention, only the structures related to the disclosed embodiments are involved. For other structures, reference may be made to the general design. Without conflict, the same embodiment and different embodiments of the present invention can be combined with each other; Finally: The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A multimodal communication control method for an intelligent driving car, characterized in that: The method comprises a vehicle-mounted sensing terminal, a vehicle control system, a control center, a cloud database module and a HUD module, wherein the vehicle-mounted sensing terminal, the vehicle control system, the control center and the HUD module are all arranged on the smart driving vehicle, and the cloud database module is located in the cloud. The method specifically comprises the following steps: S1. The vehicle sensing terminal collects vehicle dynamic parameters of the same series of vehicles and stores these vehicle dynamic parameter data in the cloud database module. The vehicle dynamic parameter data of each vehicle constructs an independent data subset to form a data set; S2. The control center extracts the above-mentioned vehicle dynamic parameters in a single data subset from the data set as a training set, wherein the vehicle dynamic parameters include steering dynamic parameters, braking dynamic parameters and power system state parameters; S3. The control center processes the extracted steering dynamic parameters to obtain a steering dynamic coefficient, processes the braking dynamic parameters to obtain a braking dynamic coefficient, processes the power system state parameters to obtain a power system risk coefficient, and sets a safety threshold Rth through scenario adaptive weight calculation; S4. The vehicle-mounted sensing terminal collects the current vehicle dynamic parameters in real time, and transmits the current vehicle dynamic parameter data to the control center in real time through the real-time communication module. After receiving the current vehicle dynamic parameter data, the control center performs preprocessing and performs a comprehensive evaluation based on multiple risk factors to calculate the risk index Rc; S5. The control center compares the calculated risk index Rc with the preset safety threshold Rth to determine whether the driving state of the vehicle exceeds the safety threshold Rth; S6. If the driving state of the vehicle does not exceed the safety threshold Rth, the current driving state is maintained without intervention; S7. If the driving state of the vehicle exceeds the safety threshold Rth, the vehicle control system sends the warning information to the HUD module according to the control command and issues a sound to remind the owner. At the same time, the vehicle control system records and feeds back the execution results to the control center. The control center dynamically optimizes the safety threshold Rth through online learning based on these recorded and fed back data, and updates the HUD warning prompt at the same time.

2. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 1, characterized in that: The vehicle dynamic parameters include steering dynamic parameters, braking dynamic parameters and power system status parameters. The steering dynamic parameters include the time interval tr from the driver issuing a steering command to the start of the vehicle's actual steering action, and the time period Te required for the vehicle to complete a complete steering action; the braking dynamic parameters include the time interval td from the vehicle's brake pedal being pressed to the brake system responding, the brake pedal pressure buildup time tp, and the efficiency η of the braking system under actual driving conditions; the power system status parameters include the engine output efficiency ϵ and the temperature of the engine coolant, denoted as T.

3. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 2, characterized in that: The efficiency η of the braking system under actual driving conditions is the ratio of the actual braking distance to the theoretical braking distance. The theoretical braking distance is calculated based on the real-time vehicle speed, the real-time road friction coefficient and the slope, and the real-time road friction coefficient and slope information are obtained through the vehicle-mounted sensing terminal.

4. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 1, characterized in that: The control center processes the extracted steering dynamic parameters to obtain the steering dynamic coefficient. The specific processing method is: calculating the steering dynamic coefficient by a standardized method. The calculation method is as follows: , c1 is the steering dynamic coefficient, trmax is the maximum historical steering delay time of the same model, and Temax is the maximum steering cycle allowed by the vehicle design.

5. The multimodal communication control method for an intelligent driving vehicle according to claim 1, characterized in that: The control center processes the extracted braking dynamic parameters to obtain the braking dynamic coefficient. The specific processing method is: calculating the braking dynamic coefficient by weighted fusion. The calculation method is as follows: ， c2 is the braking dynamic coefficient, tdnominal and tpnominal are nominal values, η is the ratio of actual braking distance to theoretical braking distance, and w1, w2, and w3 are dynamic weights dynamically allocated through correlation analysis between real-time vehicle speed and braking pressure.

6. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 1, characterized in that: The control center processes the extracted power system state parameters to obtain the power system risk coefficient. The specific processing method is: calculating the power system risk coefficient by threshold segmentation, denoted as c3, and the calculation method is as follows: 。 7. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 1, characterized in that: The safety threshold integrates the steering dynamic coefficient, the braking dynamic coefficient, the power system state parameter and calculates the safety threshold Rth through scene adaptive weighting. The calculation method is as follows: ， α, β, and γ are scene adaptive weight coefficients. The vehicle-mounted camera in the vehicle perception terminal identifies the road type and the meteorological sensor obtains weather condition information, and is dynamically adjusted according to the road type and weather conditions.

8. The multimodal communication control method for an intelligent driving vehicle according to claim 1, characterized in that: The risk index Rc is calculated by integrating the steering dynamic coefficient c1, the braking dynamic coefficient c2 and the power system risk coefficient c3, and the specific weight is dynamically allocated according to the real-time driving scenario.

9. A multimodal communication control method for an intelligent driving vehicle as claimed in claim 1, characterized in that: The optimization process of the control center dynamically optimizing the safety threshold Rth through online learning includes: the control center counts the feedback frequency of the vehicle driving status exceeding the safety threshold Rth, and the statistical cycle is once per hour; if the feedback frequency is continuously higher than the preset threshold for three consecutive statistical cycles, it indicates that the current safety threshold Rth is high, and the control center automatically reduces the values ​​of the scene adaptive weight coefficients α, β, and γ to narrow the range of the safety threshold Rth; if the feedback frequency is continuously lower than the preset threshold for three consecutive statistical cycles, it indicates that the current safety threshold Rth is low, and the control center automatically increases the values ​​of the scene adaptive weight coefficients α, β, and γ to expand the range of the safety threshold Rth.

Images