Vehicle control method and device based on Internet of Vehicles and electronic equipment

Abstract

The invention discloses a vehicle control method and device based on the Internet of Vehicles and electronic equipment, and the method comprises the steps: obtaining the real-time driving state information of a vehicle, and the driving state information comprises at least one of a vehicle state, an environment state and a network state; processing the real-time driving state information of the vehicle through a pre-trained Internet of Vehicles prediction model to obtain a driving strategy and a network resource allocation strategy of the vehicle at a future moment; wherein the Internet of Vehicles prediction model performs iterative training based on historical driving state information, a strategy corresponding to the historical driving state information and an award corresponding to the strategy until the performance of the Internet of Vehicles prediction model meets a preset condition or the number of training times reaches a preset threshold value; and controlling the vehicle based on the driving strategy and the network resource allocation strategy.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of Internet of Vehicles, and particularly relates to a vehicle control method and device based on Internet of Vehicles and electronic equipment. BACKGROUND

[0002] At present, Internet of Vehicles technology has become an important part of intelligent transportation systems, but the traditional Internet of Vehicles network faces problems such as insufficient bandwidth, high delay, unstable wireless environment, and limited number of connected devices, which to some extent limits the wide application and development of Internet of Vehicles. At present, most Internet of Vehicles relies on 4G network, and its communication capability cannot fully meet the needs of high-precision automatic driving and large-scale device connection, resulting in low real-time data transmission efficiency and long data processing delay. In addition, the security, reliability and scalability of the Internet of Vehicles network still face great challenges, especially when dealing with the increasing number of connected devices and complex and variable traffic environment.

[0003] In the prior art, the 5G Quality of Service Identifier (5QI) based on the subscription of Internet of Vehicles users configures differentiated guarantee parameters such as priority, delay and bandwidth for Internet of Vehicles users, and provides service guarantee for Internet of Vehicles users. However, it depends on static traffic model and rule-based decision mechanism, and there are problems that path planning / vehicle control and network resource allocation are difficult to adapt to dynamically changing traffic environment, resulting in inaccurate path planning, unintelligent vehicle speed control, low network resource allocation efficiency and lack of comprehensive optimization, etc., which cannot provide better strategies for vehicle driving. SUMMARY

[0004] The purpose of the embodiments of the application is to provide a vehicle control method and device based on Internet of Vehicles and electronic equipment, which can solve the problem that in the related art, the static traffic model and rule-based decision mechanism are relied on, and the path planning / vehicle control and network resource allocation are difficult to adapt to the dynamically changing traffic environment, resulting in inaccurate path planning, unintelligent vehicle speed control, low network resource allocation efficiency and lack of comprehensive optimization, etc., which cannot provide better strategies for vehicle driving.

[0005] In a first aspect, an embodiment of the present application provides a vehicle control method based on vehicle networking, the method comprising: obtaining real-time driving state information of a vehicle, wherein the driving state information comprises at least one of a vehicle state, an environment state, and a network state; processing the real-time driving state information of the vehicle through a pre-trained vehicle networking prediction model to obtain a driving strategy and a network resource allocation strategy of the vehicle at a future time; wherein the vehicle networking prediction model is iteratively trained based on historical driving state information, a strategy corresponding to the historical driving state information, and a reward corresponding to the strategy, until a performance of the vehicle networking prediction model meets a preset condition or a training frequency reaches a preset threshold; and controlling the vehicle based on the driving strategy and the network resource allocation strategy.

[0006] In a second aspect, an embodiment of the present application provides a vehicle control device based on vehicle networking, the device comprising: an obtaining module configured to obtain real-time driving state information of a vehicle, wherein the driving state information comprises at least one of a vehicle state, an environment state, and a network state; a prediction module configured to process the real-time driving state information of the vehicle through a pre-trained vehicle networking prediction model to obtain a driving strategy and a network resource allocation strategy of the vehicle at a future time; wherein the vehicle networking prediction model is iteratively trained based on historical driving state information, a strategy corresponding to the historical driving state information, and a reward corresponding to the strategy, until a performance of the vehicle networking prediction model meets a preset condition or a training frequency reaches a preset threshold; and a control module configured to control the vehicle based on the driving strategy and the network resource allocation strategy.

[0007] In a third aspect, an embodiment of the present application provides an electronic device, which comprises a processor, a memory, and a program or instructions stored in the memory and executable on the processor, and the program or instructions, when executed by the processor, implement the steps of the method according to the first aspect.

[0008] In a fourth aspect, an embodiment of the present application provides a readable storage medium, which stores a program or instructions, and the program or instructions, when executed by a processor, implement the steps of the method according to the first aspect.

[0009] In a fifth aspect, an embodiment of the present application provides a chip, which comprises a processor and a communication interface, the communication interface and the processor being coupled, and the processor being configured to execute a program or instructions to implement the steps of the method according to the first aspect.

[0010] In a sixth aspect, an embodiment of the present application provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising programs or instructions, the programs or instructions being executed to implement the steps of the method according to the first aspect.

[0011] In the embodiment of the present application, by acquiring the real-time driving state information of the vehicle, the real-time driving state information including at least one of the vehicle state, the environment state and the network state, a rich and reliable data basis is provided for subsequent prediction strategy, so that the vehicle networking prediction model can more accurately know the current vehicle, environment and network state. The real-time driving state information of the vehicle is processed through the pre-trained vehicle networking prediction model to obtain the driving strategy and the network resource allocation strategy of the vehicle at the future time; wherein the vehicle networking prediction model is iteratively trained based on the historical driving state information, the strategy corresponding to the historical driving state information and the reward corresponding to the strategy until the performance of the vehicle networking prediction model meets a preset condition or the training times reaches a preset threshold; the pre-trained vehicle networking prediction model has a better performance after training, can provide a better driving strategy and network resource allocation strategy for the vehicle at the future time, and controls the vehicle based on the driving strategy and the network resource allocation strategy, so that the vehicle drives more safely and intelligently. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is a flowchart of a vehicle control method based on vehicle networking provided by an embodiment of the present application; Figure 2 is a flowchart of a training method of a vehicle networking prediction model provided by an embodiment of the present application; Figure 3 is a schematic diagram of a component of a reward provided by an embodiment of the present application; Figure 4a is a flowchart of another vehicle control method based on vehicle networking provided by an embodiment of the present application; Figure 4b is a flowchart of another vehicle control method based on vehicle networking provided by an embodiment of the present application; Figure 5 is a structural schematic diagram of a vehicle control device based on vehicle networking provided by an embodiment of the present application; Figure 6 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0013] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly described. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art are within the scope of the present application.

[0014] The terms "first", "second", and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally a category and do not limit the number of objects, for example, the first object can be one or more. In addition, "and / or" in the specification and claims indicates at least one of the connected objects, and the character " / ", generally indicates that the front and rear associated objects are in a "or" relationship.

[0015] The vehicle control method, device and electronic device based on vehicle networking provided by the embodiments of the present application will be described in detail below in combination with the drawings, through specific embodiments and application scenarios.

[0016] Figure 1 A flowchart of a vehicle control method based on vehicle networking provided by an embodiment of the present application is shown, which can be executed by an electronic device. Referring to Figure 1 , the method can include the following steps.

[0017] Step 102, acquiring real-time driving state information of the vehicle, wherein the driving state information includes at least one of the following: vehicle state, environment state and network state.

[0018] Among them, the vehicle state can include but is not limited to: position, speed, acceleration, vehicle experience, etc., the environment state can include but is not limited to: road congestion, road obstacles, traffic signs, pedestrians, other traffic participants, etc., the vehicle state and the environment state can be collected in real time based on the 5th Generation Core network (5th Generation Core network, 5GC). The network state can include but is not limited to: network load, network performance, network traffic, network service demand, etc., which can be collected in real time based on wireless operation, management and maintenance (Operation, Administration and Maintenance).

[0019] In the 5G Vehicle to Everything (V2X) system, multi-dimensional sensors establish an efficient information collection mechanism between vehicles and the environment, enabling precise and real-time data flow and analysis. The collection of vehicle location information relies on high-precision Global Navigation Satellite System (GNSS) combined with vehicle-mounted Inertial Measurement Unit (IMU), improving positioning accuracy through Real Time Kinematic (RTK) positioning technology to ensure that vehicle dynamic trajectory data can achieve centimeter-level accuracy. At the same time, through multi-band radar sensors and LiDAR, the surrounding environment is monitored in real time to identify road obstacles, traffic signs, pedestrians and other road users. The collection of vehicle experience information involves the vehicle sensor network, which records the driver's physiological state, driving behavior, vehicle ride comfort and other parameters through vehicle-mounted sensors. Physiological information includes driver heart rate, breathing rate, eye tracking, etc. These data are transmitted in real time to the cloud through vehicle-mounted sensors and external monitoring equipment for processing to provide diversified personalized services for the vehicle networking system. In addition, the collection of environmental information covers key parameters such as air quality, weather data (such as temperature and humidity, wind speed, etc.), road slipperiness, and traffic flow. Environmental data is transmitted through intelligent sensor networks and Low Power Wide Area Network (LPWAN) protocols, combined with the high-bandwidth characteristics of 5G networks to ensure low-latency, mass-concurrent, and fast data transmission. These multi-dimensional sensors collect and transmit data in real time through vehicle-mounted computing units (ECU) and vehicle communication modules (V2X communication) through the 5G vehicle networking architecture, edge computing and cloud platform collaboration to process and optimize data in real time, accurately identify and intelligently adjust vehicle driving status, traffic conditions and driver behavior. The core data fields collected by the sensors include Global Position System (GPS) coordinates, speed, acceleration, tilt angle, physiological data read by in-vehicle sensors, external environmental temperature, humidity, air quality, etc.

[0020] Step 104, processing the real-time driving state information of the vehicle through the pre-trained vehicle networking prediction model to obtain the driving strategy and network resource allocation strategy of the vehicle at the future time.

[0021] The vehicle networking prediction model is iteratively trained based on historical driving state information, a strategy corresponding to the historical driving state information, and a reward corresponding to the strategy until the performance of the vehicle networking prediction model meets a preset condition or the number of training reaches a preset threshold.

[0022] The vehicle networking prediction model is constructed based on a deep learning algorithm, and the structure thereof includes a fusion architecture of a convolutional neural network (CNN) and a long short-term memory network (LSTM). The real-time driving state information input is subjected to time series data analysis and feature extraction, and future actions are predicted, including a combination of driving strategies and network resource allocation strategies. In addition, before the data is input into the vehicle networking prediction model, the data is compressed and transmitted, and a dynamic coding technology is used to reduce the packet loss rate of the data during transmission.

[0023] Step 106: Controlling the vehicle based on the driving strategy and the network resource allocation strategy.

[0024] The vehicle is controlled based on the driving strategy and the network resource allocation strategy predicted by the pre-trained vehicle networking prediction model, which enhances the adaptability of the vehicle to traffic conditions and improves the intelligent level and safety of traffic management.

[0025] In the embodiment of the present application, by acquiring real-time driving state information of the vehicle, the driving state information includes at least one of the following: vehicle state, environmental state and network state, which provides a rich and reliable data basis for subsequent prediction strategy, so that the vehicle networking prediction model can more accurately know the current vehicle, environmental and network state. The real-time driving state information of the vehicle is processed through the pre-trained vehicle networking prediction model to obtain the driving strategy and network resource allocation strategy of the vehicle at a future time; wherein the vehicle networking prediction model is iteratively trained based on historical driving state information, a strategy corresponding to the historical driving state information, and a reward corresponding to the strategy until the performance of the vehicle networking prediction model meets a preset condition or the number of training reaches a preset threshold; the pre-trained vehicle networking prediction model has better performance after training, which can provide better driving strategy and network resource allocation strategy for the vehicle at a future time, and the vehicle is controlled based on the driving strategy and the network resource allocation strategy, so that the vehicle is safer and more intelligent.

[0026] In one implementation, the driving strategy includes at least one planned path and condition information of each planned path; and step 106 of controlling the vehicle based on the driving strategy and the network resource allocation strategy can include the following steps.

[0027] Step 1061, determining a target path from the at least one planning path according to the condition information of each planning path.

[0028] The target path can be a path that avoids hot spot areas. It can also be a path that meets other requirements, such as few traffic lights, short distance, high safety, etc.

[0029] Step 1062, in response to the distance between the real-time position of the vehicle and the hot spot area in the target path being less than a preset distance threshold, optimizing the network resource allocation strategy according to the service level agreement (SLA) of the vehicle in the hot spot area, to determine a target network resource allocation strategy; wherein the hot spot area is an area with a vehicle density greater than or equal to a preset density threshold.

[0030] When the distance between the real-time position of the vehicle and the hot spot area in the target path is less than the preset distance threshold, it means that the vehicle is about to travel to the hot spot area. In order to ensure the smooth network of the vehicle, the real-time situation of the vehicle is judged whether it meets the service level agreement (SLA) of the hot spot area, so as to optimize the predicted network resource allocation strategy in real time, and fine-tune it again on the basis of prediction, so that the network resource allocation strategy can meet the SLA and ensure the smooth network of the vehicle in the hot spot area.

[0031] Step 1063, controlling the vehicle based on the target path and the target network resource allocation strategy.

[0032] The target path and the target network resource allocation strategy are fine-tuned again according to the real-time traffic conditions on the basis of the prediction of the Internet of Vehicles prediction model, so as to further improve the adaptability of the target path and the target network resource allocation strategy to the real-time traffic conditions.

[0033] In the embodiments of the present application, the more optimal target path is selected by analyzing the condition information of each planned path, and the network resource allocation strategy is further optimized based on the conditions of each region passed by the target path. In response to the distance between the real-time position of the vehicle and the hot spot region in the target path being less than a preset distance threshold, the network resource allocation strategy is optimized according to the service level agreement (SLA) of the vehicle in the hot spot region, and a target network resource allocation strategy is determined, thereby realizing dynamic network resource allocation. According to the real-time traffic conditions and network state, the network resources are dynamically adjusted to ensure that the performance and response time of key applications in the Internet of Vehicles are always in the optimal state. Through continuous monitoring of network traffic, user demand and quality of service parameters, the predicted network resource allocation strategy is further optimized in real time to adapt to the burst traffic caused by vehicle-intensive areas or specific events. For example, when a large public event or a major traffic accident occurs, the network resources of adjacent regions, such as spectrum allocation and power control, can be adjusted in real time to ensure that the network communication of the vehicle is not affected, as shown in Table 1, which shows an exemplary effect of dynamic resource allocation under different traffic and network conditions. Through real-time dynamic optimization of the predicted planned path and network resource allocation strategy, a more adaptive target path and target network resource allocation strategy to the real-time traffic conditions are provided for vehicle driving.

[0034] Table 1.

[0035] In the table, it is reflected how to adjust the spectrum resources and power settings to optimize the quality of service and reduce network delay according to the number of vehicles and the type of event under different scenarios. Dynamic resource allocation not only improves the efficiency of network resource use, but also ensures the continuity and reliability of Internet of Vehicles applications such as emergency response systems and real-time navigation services.

[0036] In one implementation, the step 106 of controlling the vehicle based on the driving strategy and the network resource allocation strategy can include the following steps.

[0037] Step 1064, in response to receiving the abnormal warning information sent by the network side device, obtaining the information of the abnormal region; wherein the abnormal warning information is used to indicate that there is an abnormal region in the driving strategy that occurs an abnormal situation.

[0038] Wherein, the abnormal detection by the network side device includes the detection of network state, the detection of vehicle state, and the detection of environment state, for example, there is an abnormal region that occurs a traffic accident or a network failure. The network side device sends abnormal warning information to the vehicle side, and the vehicle side optimizes the driving strategy and the network resource allocation strategy according to the information of the abnormal region indicated by the abnormal warning information.

[0039] Step 1065: Based on the information of the abnormal area, optimize the driving strategy or the network resource allocation strategy to determine the target driving strategy or the target network resource allocation strategy.

[0040] The information on abnormal areas can include the location of the abnormal area, the abnormal events that occurred in the abnormal area, and the severity of the abnormal events, providing reference information for optimizing driving strategies and network resource allocation strategies.

[0041] In one alternative implementation, information about abnormal regions can be input into the pre-trained vehicle-to-everything (V2X) prediction model and used in conjunction with real-time driving information for prediction. Alternatively, the model can be optimized after it outputs its prediction results.

[0042] For example, for emergency communication needs, such as high-priority communication services for emergency vehicles (like ambulances or fire trucks), network resource priorities can be adjusted in real time to ensure that communication links for these critical services are guaranteed under any circumstances.

[0043] Step 1066: Control the vehicle based on the target driving strategy and the target network resource allocation strategy.

[0044] In this embodiment, the vehicle's driving strategy and network resource allocation strategy for future moments are predicted based on the vehicle's real-time driving information using a vehicle-to-everything (V2X) prediction model. The method may also include responding to anomaly alerts sent by network-side devices, acquiring information about abnormal areas, and then optimizing the predicted driving strategy and network resource allocation strategy based on this information to determine target driving strategies and target network resource allocation strategies. This is to address abnormal situations such as network failures or traffic accidents, thereby ensuring vehicle driving safety and network stability.

[0045] In one implementation, the vehicle-to-everything (V2X) prediction model is trained before using a pre-trained V2X prediction model to predict the vehicle's driving strategy and network resource allocation strategy at future moments. Figure 2 This document illustrates a flowchart of a training method for a vehicle-to-everything (V2X) prediction model according to an embodiment of this application. (See attached diagram.) Figure 2 The training method may include the following steps.

[0046] Step 202: Obtain the training dataset of the vehicle-to-everything (V2X) prediction model; wherein the training dataset includes the historical driving state information, the strategy corresponding to the historical driving state information, and the reward corresponding to the strategy. The vehicle-to-everything (V2X) prediction model is built based on deep learning algorithms. Its training data consists of states, actions, and rewards. In this embodiment, the states are historical driving state information, which may include: Vehicle location: the precise coordinates of the vehicle on a map obtained through GPS or onboard sensors; Speed: the vehicle's current speed; Acceleration: the vehicle's current acceleration, used to assess the vehicle's acceleration or deceleration; Road congestion: information on the degree of road congestion obtained through road condition sensors or real-time traffic data; Traffic light status: the current status of the traffic lights ahead of the vehicle (red, green, yellow); Communication network signal strength and latency parameters: 5G network signal strength and latency information provided by the onboard communication system's state sensors. The actions are strategies corresponding to the historical driving state information, namely driving strategies and network resource allocation strategies, which may include: Vehicle acceleration: increasing the vehicle's speed; Deceleration: decreasing the vehicle's speed; Steering: changing the vehicle's direction of travel; Lane changing: changing lanes on the road; Maintaining speed: maintaining the current speed. Network resource allocation strategy: dynamically adjusting the allocation of network resources based on the vehicle and network states to optimize communication performance. Rewards are measured by multiple metrics, which may include: time efficiency rewards, energy consumption penalties, environmental response rewards, and network resource utilization rewards.

[0047] Step 204: Process the training dataset through the vehicle-to-everything (V2X) prediction model to obtain the prediction strategy and expected reward corresponding to the historical driving status information.

[0048] The training data used in the training can be a portion of the training dataset randomly selected to improve training efficiency. Based on deep learning algorithms, the vehicle-to-everything (V2X) prediction model can include a policy network and a value network. The prediction policy is output by the policy network, and the expected reward is output by the value network. This expected reward is the cumulative reward for each prediction policy.

[0049] Step 206 uses the mean squared error (MSE) algorithm to determine the loss function based on the prediction strategy and the expected reward.

[0050] The loss function can be determined using the Mean Square Error (MSE) algorithm, as shown below: Where n is the amount of training data. The predicted reward is determined based on the prediction strategy. This is the expected reward.

[0051] Step 208: Update the parameters of the vehicle-to-everything (V2X) prediction model based on the loss function and gradient descent algorithm.

[0052] The parameters of the vehicle-to-everything (V2X) prediction model are updated based on the loss function and gradient descent algorithm as follows: .in, These are the parameters of the current vehicle-to-everything (V2X) prediction model. These are the parameters for the updated vehicle-to-everything (V2X) prediction model. The learning rate is used to control the step size for parameter updates. For loss function about The gradient.

[0053] In this embodiment, historical driving state information, corresponding strategies, and rewards are input into a vehicle-to-everything (V2X) prediction model for processing. The model outputs a prediction strategy and expected reward based on the historical driving state information. The prediction strategy and expected reward are then used to determine a loss function using the MSE algorithm. The parameters of the V2X prediction model are updated based on the loss function and gradient descent algorithm. This process is repeated multiple times to minimize the loss function, iterating the model until convergence. Convergence conditions may include the V2X prediction model meeting preset performance conditions or reaching a preset threshold for the number of training iterations. By updating the parameters of the V2X prediction model, its performance is continuously optimized, enabling the trained model to provide better driving strategies and network resource allocation strategies for vehicle operation.

[0054] In one implementation, after updating the parameters of the vehicle-to-everything (V2X) prediction model based on the loss function and gradient descent algorithm in step 208 above, the method may further include the following steps.

[0055] Step 210: Obtain the actual driving status information and actual reward after executing the prediction strategy.

[0056] Based on the historical driving state information used in each training round, the prediction strategy output by the vehicle-to-everything (V2X) prediction model is executed, and the results after execution, namely the actual driving state information and the actual reward, are analyzed. The performance of the V2X prediction model is evaluated using the actual driving state information and the actual reward to guide the parameter updates of the V2X prediction model.

[0057] Step 212: Determine the performance of the vehicle-to-everything (V2X) prediction model based on the actual driving status information and the actual reward.

[0058] The performance of the vehicle-to-everything (V2X) prediction model can be determined based on actual driving status information and actual rewards from the following aspects: the coordination and smoothness of vehicle driving behavior, adaptability to the environment, and communication quality.

[0059] Step 214: If the performance of the vehicle-to-everything (V2X) prediction model does not meet the preset conditions, the actual driving state information and the actual reward are added to the training dataset to retrain the V2X prediction model.

[0060] The preset conditions may include: the accuracy of path planning is greater than a preset threshold, and the efficiency of network resource allocation is greater than a preset threshold. If the performance of the vehicle-to-everything (V2X) prediction model does not meet the preset conditions, it indicates that the performance of the V2X prediction model needs to be improved. In this case, actual driving state information and actual rewards are added to the training dataset to enrich the training dataset, and the V2X prediction model is iteratively trained again.

[0061] In this embodiment, the performance of the vehicle-to-everything (V2X) prediction model is evaluated by acquiring the actual driving state information and actual rewards after executing the prediction strategy. The evaluation determines whether the model's performance meets preset conditions. If the model's performance does not meet these conditions, it indicates poor performance and requires further training. The acquired actual driving state information and rewards are then added to the training dataset to enrich it. If the model's performance meets the preset conditions, it indicates that the model's performance meets the requirements and can be used for real-time processing of driving state information to predict future driving strategies and network resource allocation strategies. Alternatively, training can be stopped when the number of training iterations reaches a preset threshold to avoid endless training and resource consumption. This indicates the model's performance has reached a certain level and can also be used for real-time processing of driving state information to predict future driving strategies and network resource allocation strategies. This application provides another way to evaluate whether the parameters of the vehicle-to-everything (V2X) prediction model should continue to be updated by assessing the performance of the V2X prediction model. It focuses on the performance of the V2X prediction model so that the converged V2X prediction model can provide better driving strategies and network resource allocation strategies for vehicle driving.

[0062] In one implementation, the actual reward mentioned above, that is, the immediate reward, may include at least one of the following: time efficiency reward, energy consumption penalty, environmental response reward, and network resource utilization reward.

[0063] Among them, see Figure 3 , Figure 3The diagram illustrates the components of a reward provided in an embodiment of this application. The time efficiency reward is determined based on the vehicle's estimated arrival time in the historical driving status information and the vehicle's actual arrival time in the actual driving status information; the energy consumption penalty is determined based on the energy actually consumed by the vehicle; the environmental response reward is determined based on the vehicle's rate of change of acceleration and the distance between the vehicle and obstacles; and the network resource utilization reward is determined based on network throughput.

[0064] The actual reward provided in this application embodiment, that is, the immediate reward, is measured by a combination of multiple indicators. Firstly, it includes the outcome of the vehicle's driving behavior, such as whether it automatically adjusts its speed to match surrounding vehicles while maintaining a safe distance, reducing unnecessary acceleration and braking, and effectively alleviating traffic congestion; and secondly, it includes the vehicle's response to the environment, including weather conditions, road conditions, and traffic restrictions. Through advanced sensors and 5G communication technology, the vehicle can receive and process environmental data in real time, adjusting its driving strategy accordingly, such as reducing speed in rainy weather and increasing following distance on icy roads. Higher reward scores are assigned to behaviors that effectively adapt to environmental changes and ensure driving safety. This directly quantifies the adaptability of the vehicle's control behavior to the environment and feeds it back to aggressive adjustments in control behavior, thereby improving vehicle driving safety and environmental adaptability.

[0065] Figure 4a and Figure 4b The illustration shows a flowchart of another vehicle control method based on the Internet of Vehicles (IoV) provided in an embodiment of this application. This method can be executed by an electronic device. The method may include the following steps.

[0066] Step 401: Obtain the training dataset, which includes: historical driving status information of the vehicle, the strategy corresponding to the historical driving status information, and the reward corresponding to the strategy.

[0067] The strategies include driving strategies and network resource allocation strategies.

[0068] Step 402: Initialize the parameters of the constructed vehicle-to-everything (V2X) prediction model.

[0069] The vehicle-to-everything (V2X) prediction model is built based on deep learning algorithms and includes a policy network and a value network.

[0070] Step 403: Extract a portion of the training data from the training dataset and input it into the vehicle-to-everything (V2X) prediction model for processing to obtain the prediction strategy and expected reward.

[0071] Step 404: Determine the loss function using the mean squared error (MSE) algorithm based on the prediction strategy and expected reward.

[0072] Step 405: Update the parameters of the vehicle-to-everything (V2X) prediction model based on the loss function and gradient descent algorithm.

[0073] Step 406: Obtain the actual driving status information and actual reward after executing the prediction strategy, and determine the performance of the vehicle-to-everything (V2X) prediction model based on the actual driving status information and actual reward.

[0074] Step 407: Determine whether (1) the performance of the vehicle-to-everything (V2X) prediction model meets the preset conditions and (2) the number of training iterations exceeds the preset threshold. If (1) yes, proceed to step 408. If (1) no and (2) no, proceed to step 409. If (1) no and (2) yes, proceed to step 408.

[0075] Step 408: Training ends, the vehicle-to-everything (V2X) prediction model converges, and the trained V2X prediction model is obtained.

[0076] Step 409: Add the actual driving status information and actual rewards to the training dataset, then return to step 403.

[0077] Step 410: Collect real-time driving status information of the vehicle, input the real-time driving status information into the trained vehicle-to-everything (V2X) prediction model for processing, and obtain the vehicle's driving strategy and network resource allocation strategy for future time moments.

[0078] The driving strategy includes at least one planned route and information about the condition of that route.

[0079] Step 411: Determine the target path from at least one planned path based on the status information of each planned path.

[0080] Step 412: In response to the vehicle about to enter a hotspot area in the target path, optimize the network resource allocation strategy according to the Service Level Agreement (SLA) of the vehicle in the hotspot area.

[0081] In some alternative embodiments, step 412 described above may include the following steps.

[0082] Step 412a: In cases where network quality is poor in hotspot areas, establish a Guaranteed Bit Rate (GBR) guarantee.

[0083] Step 412b: If the Service Level Agreement (SLA) is low in the hotspot area, reduce the speed in the driving strategy.

[0084] Step 412c: If the minimum service level agreement (SLA) cannot be met in a hotspot area, the backend vehicle control platform is alerted and instructed to take over vehicle control as needed.

[0085] In one implementation, the training method for the vehicle-to-everything (V2X) prediction model provided in this application embodiment can also evaluate the performance of the V2X prediction model through simulation experiments in order to optimize the V2X prediction model.

[0086] This includes building a simulation environment capable of simulating the dynamic driving states of vehicles, including speed, position, and acceleration; vehicle-to-vehicle (V2V) and V2I (V2I) communication interactions; real-time changes in road conditions, such as traffic congestion and road construction; fluctuations in network signal strength, considering factors such as base station distribution, signal attenuation, and interference; and potential cybersecurity threats, such as cyberattacks and data breaches. Through meticulous parameter settings and model calibration, the simulation environment is ensured to realistically reflect the complexity and dynamism of 5G vehicle-to-everything (V2X) networks.

[0087] This also includes setting up different case studies, which cover vehicle communication performance tests under different traffic densities to evaluate the performance of vehicle-to-everything (V2X) prediction models in high-density and low-density traffic scenarios; data transmission efficiency tests under different network topologies, such as urban, suburban, and highway scenarios; network stability tests under extreme weather conditions, such as the impact of heavy rain and fog on communication quality; and defense capability tests against specific security threats, such as network attacks and data tampering scenarios.

[0088] This also includes generating simulation data. Based on the simulation environment and cases, simulation data is generated, including vehicle status information, environmental status information, and network status information, such as vehicle driving trajectory, communication latency, packet loss rate, network throughput, security event records, etc.

[0089] This process includes inputting vehicle status information, environmental status information, and network status information into a vehicle-to-everything (V2X) prediction model for processing, and outputting a predicted first driving strategy and a first network resource allocation strategy. Simultaneously, under the same conditions, vehicle status information, environmental status information, and network status information are processed using traditional V2X prediction methods to determine a second driving strategy and a second network resource allocation strategy. Vehicle control is then performed based on the obtained first driving strategy and first network resource allocation strategy, and second driving strategy and second network resource allocation strategy, respectively, and actual rewards are obtained to evaluate the effectiveness of the two methods. Evaluation metrics may include: network throughput, communication latency, packet loss rate, and security incident occurrence rate.

[0090] This also includes optimizing the vehicle-to-everything (V2X) prediction model based on the above evaluation results, further improving the performance of the V2X prediction model.

[0091] Furthermore, representative simulation environments and cases are selected, and the corresponding simulation data is applied to real-world environments to deploy the vehicle-to-everything (V2X) prediction model for long-term operation and monitoring. Real-world data is then acquired to evaluate the V2X prediction model, further optimizing its performance to provide better strategies for vehicle operation.

[0092] Through the embodiments of this application, multi-dimensional real-time data is collected to provide a reliable data foundation for the prediction strategy. The vehicle-to-everything (V2X) prediction model used for the prediction strategy is built based on a deep learning algorithm. Through multiple iterations of historical data, the parameters of the V2X prediction model are updated, and its performance is continuously optimized. This enables the trained V2X prediction model to provide better strategies for vehicle driving in future moments, adaptively adjusting driving paths, vehicle behavior, and network resource allocation to address potential or existing security risks and network anomalies, thereby improving vehicle driving safety.

[0093] It should be noted that the vehicle control method based on the Internet of Vehicles (IoV) provided in this application embodiment can be executed by an IoV-based vehicle control device, or a control module within that IoV-based vehicle control device for executing the IoV-based vehicle control method. This application embodiment uses an IoV-based vehicle control device executing the method as an example to illustrate the IoV-based vehicle control device provided in this application embodiment.

[0094] Figure 5 This illustration shows a structural schematic diagram of a vehicle control device based on the Internet of Vehicles (IoV) according to an embodiment of this application. (See also...) Figure 5 The device 500 may include: an acquisition module 51, a prediction module 52, and a control module 53.

[0095] The system includes: an acquisition module 51 for acquiring real-time driving status information of the vehicle, wherein the driving status information includes at least one of the following: vehicle status, environmental status, and network status; a prediction module 52 for processing the real-time driving status information of the vehicle through a pre-trained vehicle-to-everything (V2X) prediction model to obtain the vehicle's driving strategy and network resource allocation strategy at future times; wherein the V2X prediction model is iteratively trained based on historical driving status information, the strategies corresponding to the historical driving status information, and the rewards corresponding to the strategies, until the performance of the V2X prediction model meets preset conditions or the number of training iterations reaches a preset threshold; and a control module 53 for controlling the vehicle based on the driving strategy and the network resource allocation strategy.

[0096] In one implementation, the driving strategy includes at least one planned path and condition information for each planned path; the control module 53 can be used to determine a target path from the at least one planned path based on the condition information of each planned path; in response to the distance between the real-time location of the vehicle and a hotspot area in the target path being less than a preset distance threshold, the network resource allocation strategy is optimized according to the Service Level Agreement (SLA) of the vehicle in the hotspot area to determine a target network resource allocation strategy; wherein, the hotspot area is an area where the vehicle density is greater than or equal to a preset density threshold; the vehicle is controlled based on the target path and the target network resource allocation strategy.

[0097] In one implementation, the control module 53 can be used to obtain information about an abnormal area in response to receiving an abnormal warning message sent by a network-side device; wherein the abnormal warning message is used to indicate that an abnormal situation has occurred in the abnormal area in the driving strategy; based on the information about the abnormal area, the driving strategy or the network resource allocation strategy is optimized to determine a target driving strategy or a target network resource allocation strategy; and the vehicle is controlled based on the target driving strategy and the target network resource allocation strategy.

[0098] In one implementation, the aforementioned apparatus 500 may further include a training module for acquiring a training dataset for the vehicle-to-everything (V2X) prediction model; wherein the training dataset includes the historical driving state information, the strategy corresponding to the historical driving state information, and the reward corresponding to the strategy; the training dataset is processed by the V2X prediction model to obtain the prediction strategy and expected reward corresponding to the historical driving state information; the prediction strategy and the expected reward are used to determine a loss function using a mean squared error (MSE) algorithm; and the parameters of the V2X prediction model are updated based on the loss function and a gradient descent algorithm.

[0099] In one implementation, the training module described above can also be used to obtain the actual driving state information and actual reward after executing the prediction strategy; determine the performance of the vehicle-to-everything (V2X) prediction model based on the actual driving state information and the actual reward; and, if the performance of the V2X prediction model does not meet the preset conditions, add the actual driving state information and the actual reward to the training dataset to retrain the V2X prediction model.

[0100] In one implementation, the actual reward includes at least one of the following: a time efficiency reward, an energy consumption penalty, an environmental response reward, and a network resource utilization reward; wherein the time efficiency reward is determined based on the vehicle's estimated arrival time in the historical driving status information and the vehicle's actual arrival time in the actual driving status information; the energy consumption penalty is determined based on the energy actually consumed by the vehicle; the environmental response reward is determined based on the vehicle's rate of change of acceleration and the distance between the vehicle and obstacles; and the network resource utilization reward is determined based on network throughput.

[0101] The vehicle control device based on the Internet of Vehicles (IoV) in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the specific type of device.

[0102] The vehicle control device based on the Internet of Vehicles (IoV) in this application embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit it.

[0103] The vehicle control device based on the Internet of Vehicles provided in this application embodiment can achieve... Figure 1 , Figure 2 , Figure 4a and Figure 4b The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0104] Based on the same technical concept, embodiments of this application also provide an electronic device for executing the aforementioned vehicle control method based on the Internet of Vehicles. Figure 6This is a schematic diagram of the structure of an electronic device to implement the various embodiments of this application. The electronic device can vary significantly due to differences in configuration or performance, and may include a processor 601, a communications interface 602, a memory 603, and a communication bus 604. The processor 601, communications interface 602, and memory 603 communicate with each other via the communication bus 604. The processor 601 can call a computer program stored in the memory 603 and executable on the processor 601 to perform the various steps of the above-described vehicle control method embodiments based on the Internet of Vehicles, achieving the same technical effects. To avoid repetition, further details are omitted here.

[0105] It should be noted that the electronic devices in the embodiments of this application include servers, terminals, or other devices besides terminals. For example, automobiles, robots, and handheld devices.

[0106] The above electronic device structure does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or arrange them differently. For example, an input unit may include a Graphics Processing Unit (GPU) and a microphone, and a display unit may use a liquid crystal display (LCD), organic light-emitting diode (OLED), or other similar display panels. User input units include at least one of a touch panel and other input devices. A touch panel is also called a touchscreen. Other input devices may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be elaborated further here.

[0107] Memory can be used to store software programs and various data. Memory can primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area can store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, memory can include volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus RAM (DRRAM).

[0108] The processor may include one or more processing units; optionally, the processor integrates an application processor and a modem processor, wherein the application processor mainly handles operations related to the operating system, user interface, and applications, while the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into the processor.

[0109] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described vehicle control method embodiments based on the Internet of Vehicles and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0110] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0111] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described vehicle control method embodiment based on the Internet of Vehicles, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0112] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0113] This application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes a program or instructions. When the program or instructions are executed, they implement the various processes of the above-described vehicle control method embodiments based on the Internet of Vehicles and achieve the same technical effects. To avoid repetition, they will not be described again here.

[0114] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0115] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0116] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A vehicle control method based on vehicle-to-everything (V2X) communication, characterized in that, include: Obtain real-time driving status information of the vehicle, wherein the driving status information includes at least one of the following: vehicle status, environmental status, and network status; The real-time driving status information of the vehicle is processed by a pre-trained vehicle-to-everything (V2X) prediction model to obtain the vehicle's driving strategy and network resource allocation strategy at future times. The V2X prediction model is iteratively trained based on historical driving status information, the strategies corresponding to the historical driving status information, and the rewards corresponding to the strategies, until the performance of the V2X prediction model meets the preset conditions or the number of training times reaches the preset threshold. The vehicle is controlled based on the driving strategy and the network resource allocation strategy.

2. The method according to claim 1, characterized in that, The driving strategy includes at least one planned route and condition information for each planned route; The control of the vehicle based on the driving strategy and the network resource allocation strategy includes: Based on the status information of each of the planned paths, a target path is determined from the at least one planned path; In response to the real-time location of the vehicle being less than a preset distance threshold between it and a hotspot area in the target path, the network resource allocation strategy is optimized based on the Service Level Agreement (SLA) of the vehicle in the hotspot area to determine the target network resource allocation strategy; wherein, the hotspot area is an area where the vehicle density is greater than or equal to a preset density threshold. The vehicle is controlled based on the target path and the target network resource allocation strategy.

3. The method according to claim 1, characterized in that, The control of the vehicle based on the driving strategy and the network resource allocation strategy includes: In response to receiving an abnormal warning message from a network-side device, information about the abnormal area is obtained; wherein, the abnormal warning message is used to indicate that there is an abnormal situation in the abnormal area in the driving strategy; Based on the information of the abnormal area, the driving strategy or the network resource allocation strategy is optimized to determine the target driving strategy or the target network resource allocation strategy. The vehicle is controlled based on the target driving strategy and the target network resource allocation strategy.

4. The method according to claim 1, characterized in that, The method further includes: Obtain the training dataset of the vehicle-to-everything (V2X) prediction model; wherein the training dataset includes the historical driving state information, the strategy corresponding to the historical driving state information, and the reward corresponding to the strategy; The training dataset is processed by the vehicle-to-everything (V2X) prediction model to obtain the prediction strategy and expected reward corresponding to the historical driving status information. The loss function is determined by the mean squared error (MSE) algorithm using the prediction strategy and the expected reward. The parameters of the vehicle-to-everything (V2X) prediction model are updated based on the loss function and gradient descent algorithm.

5. The method according to claim 4, characterized in that, After updating the parameters of the vehicle-to-everything (V2X) prediction model based on the loss function and gradient descent algorithm, the method further includes: Obtain the actual driving status information and actual reward after executing the prediction strategy; The performance of the vehicle-to-everything (V2X) prediction model is determined based on the actual driving status information and the actual reward. If the performance of the vehicle-to-everything (V2X) prediction model does not meet the preset conditions, the actual driving state information and the actual reward are added to the training dataset to retrain the V2X prediction model.

6. The method according to claim 5, characterized in that, The actual rewards include at least one of the following: time efficiency rewards, energy consumption penalties, environmental response rewards, and network resource utilization rewards; The time efficiency reward is determined based on the vehicle's estimated arrival time in the historical driving status information and the vehicle's actual arrival time in the actual driving status information. The energy consumption penalty is determined based on the actual energy consumed by the vehicle; The environmental response reward is determined based on the vehicle's rate of change of acceleration and the distance between the vehicle and the obstacle; The network resource utilization reward is determined based on network throughput.

7. A vehicle control device based on the Internet of Vehicles, characterized in that, include: The acquisition module is used to acquire real-time driving status information of the vehicle, wherein the driving status information includes at least one of the following: vehicle status, environmental status, and network status. The prediction module is used to process the real-time driving status information of the vehicle through a pre-trained vehicle-to-everything (V2X) prediction model to obtain the vehicle's driving strategy and network resource allocation strategy at future times. The V2X prediction model is iteratively trained based on historical driving status information, the strategies corresponding to the historical driving status information, and the rewards corresponding to the strategies, until the performance of the V2X prediction model meets preset conditions or the number of training iterations reaches a preset threshold. The control module is used to control the vehicle based on the driving strategy and the network resource allocation strategy.

8. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the vehicle control method based on the Internet of Vehicles as described in any one of claims 1 to 6.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the vehicle control method based on the Internet of Vehicles as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including programs or instructions that, when executed, implement the steps of the vehicle control method based on the Internet of Vehicles as described in any one of claims 1 to 6.

Images