Communication resource allocation method and on-board unit

By acquiring vehicle status and communication load data and optimizing resource allocation using a communication quality prediction model, the problem of low efficiency and stability in vehicle-to-everything (V2X) communication is solved, and efficient communication in complex scenarios is achieved.

WO2026232858A1PCT designated stage Publication Date: 2026-11-12CHERY AUTOMOBILE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2025/148273
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-08-14
Filing Date
2025-12-31
Publication Date
2026-11-12

AI Technical Summary

Technical Problem

In vehicle-to-everything (V2X) communication, existing technologies struggle to achieve efficient and stable allocation of communication resources in complex scenarios such as high-speed vehicle movement, rapid topology changes, and non-line-of-sight occlusion, resulting in low communication efficiency and stability.

Method used

By acquiring vehicle status data and communication load data, the target quality index of future sub-channels is predicted using a communication quality prediction model. This allows for dynamic adjustment of the resource pool, optimization of communication resource allocation, avoidance of channel congestion and interference, and ensures the dynamic adaptability of the communication link.

Benefits of technology

It improves the efficiency and stability of vehicle-to-everything (V2X) communication, reduces communication latency and the probability of resource collisions, enables rapid response to changes in vehicle status and environment, and enhances the dynamic adaptability of the communication link.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2025148273_12112026_PF_FP_ABST
    Figure CN2025148273_12112026_PF_FP_ABST
Patent Text Reader

Abstract

Provided in the embodiments of the present disclosure are a communication resource allocation method and an on-board unit. The method is applied to an on-board unit on a vehicle, and comprises: in response to a resource allocation request being received, acquiring state data of a vehicle and communication load data of an area where the vehicle is located, wherein the state data comprises at least one of the following: a position, a vehicle speed, a heading angle, and information of target obstacles around the vehicle, the information of the target obstacles being information of objects around the vehicle that block the transmission of communication signals, and the communication load data being used for indicating occupancy states of a plurality of subchannels; on the basis of the state data, determining target quality indices of the plurality of subchannels within a preset future time period, wherein the target quality indices are used for characterizing the carrying capacities of the plurality of subchannels; and on the basis of the target quality indices and the communication load data, allocating to the vehicle a communication sub-resource corresponding to a target subchannel. The present disclosure solves the technical problem in the related art of the relatively low communication efficiency and stability of Internet-of-Vehicles communication performed by a vehicle.
Need to check novelty before this filing date? Find Prior Art

Description

Communication resource allocation method and vehicle-mounted unit Technical Field

[0001] This disclosure relates to vehicle technology and the field of vehicle networking technology, and more specifically, to a communication resource allocation method and an on-board unit. Background Technology

[0002] With the increasing prevalence of intelligent transportation systems, the efficient and stable operation of vehicle-to-everything (V2X) communication technology has become a crucial requirement. Currently, V2X communication faces significant technical challenges in complex scenarios such as high-speed vehicle movement, rapid topology changes, and non-line-of-sight occlusion. Traditional communication resource allocation mechanisms, such as centralized scheduling and distributed contention architectures, exhibit significant limitations in these scenarios. These mechanisms struggle to allocate resources accurately and in real-time and coordinate interference, effectively respond to dynamic changes such as rapid vehicle movement or sudden lane changes, significantly increase the probability of communication resource collisions, and fail to meet the low-latency requirements of real-time communication services, resulting in low communication efficiency and stability for V2X communication.

[0003] There is currently no good solution to the above problems. Summary of the Invention

[0004] This disclosure provides a communication resource allocation method and an on-board unit to at least solve the technical problems of low communication efficiency and stability of vehicles conducting vehicle-to-everything (V2X) communication in related technologies.

[0005] According to one aspect of the present disclosure, a communication resource allocation method is provided, applied to an on-board unit in a vehicle, comprising: in response to receiving a resource allocation request, acquiring vehicle status data and communication load data of the area where the vehicle is located, wherein the status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle, the target obstruction information being information about objects around the vehicle that obstruct the transmission of communication signals, and the communication load data being used to represent the occupancy status of multiple sub-channels; based on the status data, determining a target quality index for multiple sub-channels within a future preset time period, wherein the target quality index is used to characterize the carrying capacity of multiple sub-channels; and based on the target quality index and the communication load data, allocating sub-communication resources corresponding to the target sub-channels to the vehicle.

[0006] In the above embodiments of this disclosure, determining the target quality index of multiple sub-channels within a future preset time period based on state data includes: obtaining the historical quality index of multiple sub-channels; fusing the state data and historical quality value data using a communication quality prediction model to obtain spatiotemporal features; and predicting the target quality index of multiple sub-channels within a future preset time period using the communication quality prediction model and the spatiotemporal features.

[0007] In the embodiments described above, the communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; the communication quality prediction model is used to fuse state data and historical quality value data to obtain spatiotemporal features, including: encoding the state data and historical quality indices using the encoding module to obtain input features; extracting the spatial positional relationships between the input features using the spatial feature extraction module to obtain spatial features; and extracting the temporal dependencies between the spatial features using the temporal feature extraction module to obtain spatiotemporal features; preferably, the communication quality prediction model further includes: an attention module and an output module; the communication quality prediction model and the spatiotemporal features are used to predict a target quality index, including: performing attention processing on the spatiotemporal features using the attention module to obtain attention features; and performing Q-value learning on the attention features using the output module to obtain the target quality index.

[0008] In the above embodiments of this disclosure, allocating sub-communication resources corresponding to a target sub-channel to a vehicle based on a target quality index and communication load data includes: determining the number of candidate resources in a target resource pool based on communication load data, wherein the target resource pool contains sub-communication resources corresponding to the target sub-channel; adjusting the sub-communication resources corresponding to multiple sub-channels contained in the original resource pool based on the number of candidate resources to obtain a target resource pool; and allocating sub-channel resources corresponding to the target sub-channel to the vehicle from the target resource pool based on the target quality index and status data; preferably, the original resource pool is a resource pool that matches the target communication priority of the vehicle among multiple resource pools, wherein different resource pools correspond to different communication priorities.

[0009] In the above embodiments of this disclosure, allocating sub-channel resources corresponding to a target sub-channel to a vehicle from the target resource pool based on the target quality index and state data includes: determining an initial resource allocation strategy for the vehicle from the target resource pool based on the target quality index; updating the initial resource allocation strategy based on the state data to obtain a target resource allocation strategy; and allocating sub-channel resources corresponding to a target sub-channel to the vehicle based on the target resource allocation strategy.

[0010] In the above embodiments of this disclosure, updating an initial resource allocation strategy based on state data to obtain a target resource allocation strategy includes: determining a first utility value corresponding to the initial resource allocation strategy based on state data, wherein the first utility value is used to characterize the benefit obtained by allocating resources to vehicles based on the initial resource allocation strategy; updating the initial resource allocation strategy to obtain an updated resource allocation strategy; determining a second utility value corresponding to the updated resource allocation strategy based on state data; determining a utility difference between the first utility value and the second utility value; in response to a utility difference greater than or equal to a preset value, using the updated resource allocation strategy as the initial resource allocation strategy, and repeatedly executing the steps of determining the first utility value corresponding to the initial resource allocation strategy based on state data, updating the initial resource allocation strategy to obtain the updated resource allocation strategy, and determining the second utility value corresponding to the updated resource allocation strategy based on state data, until the utility difference is less than the preset value; in response to a utility difference less than the preset value, determining the updated resource allocation strategy as the target resource allocation strategy.

[0011] In the above embodiments of this disclosure, the method further includes: obtaining the occupancy status of the target sub-channel; determining a backoff time in response to the occupancy status indicating that the target sub-channel is occupied, wherein the backoff time is a waiting time for re-allocating communication resources to the vehicle; and regenerating a resource allocation request in response to the arrival of the backoff time. Preferably, obtaining the occupancy status of the target sub-channel includes: obtaining the reference signal received power of the target sub-channel; determining that the occupancy status indicates that the target sub-channel is idle in response to the reference signal received power being less than a preset power; and determining that the occupancy status indicates that the target sub-channel is occupied in response to the reference signal received power being greater than or equal to a preset power.

[0012] In the above embodiments of this disclosure, the method further includes: determining a signal power compensation value based on target obstruction information; determining the transmit power of the target sub-channel based on the target quality index and the signal power compensation value; and determining the modulation scheme and coding rate of the target sub-channel based on the target quality index.

[0013] In the above embodiments of this disclosure, obtaining target occlusion information includes: acquiring image data collected by a visual sensor and point cloud data collected by radar; identifying occlusions around the vehicle based on the point cloud data to obtain first occlusion information; identifying occlusions around the vehicle based on the image data to obtain second occlusion information; and fusing the first occlusion information and the second occlusion information to obtain target occlusion information. Preferably, fusing the first occlusion information and the second occlusion information to obtain target occlusion information includes: determining the first occlusion based on position, vehicle speed, and heading angle. The system calculates a first initial weight value for object information and a second initial weight value for second occluded object information; performs semantic understanding on the first occluded object information to obtain a first semantic understanding result, and performs semantic understanding on the second occluded object information to obtain a second semantic understanding result; adjusts the first initial weight value based on the first semantic understanding result to obtain a first target weight value, and adjusts the second initial weight value based on the second semantic understanding result to obtain a second target weight value; and performs weighted processing on the first occluded object information and the second occluded object information based on the first target weight value and the second target weight value to obtain target occluded object information.

[0014] According to another aspect of the embodiments of this disclosure, a communication resource allocation apparatus is also provided, applied to an on-board unit in a vehicle, comprising: an acquisition module configured to acquire vehicle status data and communication load data of the area where the vehicle is located in response to receiving a resource allocation request, wherein the status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle, the target obstruction information being information of objects around the vehicle that obstruct the transmission of communication signals, and the communication load data being used to represent the occupancy status of multiple sub-channels; a determination module configured to determine a target quality index of multiple sub-channels within a future preset time period based on the status data, wherein the target quality index is used to characterize the carrying capacity of multiple sub-channels; and an allocation module configured to allocate sub-communication resources corresponding to the target sub-channels to the vehicle based on the target quality index and the communication load data.

[0015] According to another aspect of the present disclosure, an in-vehicle unit is also provided, including: a memory storing an executable program; and a processor configured to run the program, wherein the program executes the methods in the various embodiments of the present disclosure when it runs.

[0016] According to another aspect of the embodiments of the present disclosure, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device where the computer-readable storage medium is located to perform the methods of the various embodiments of the present disclosure.

[0017] According to another aspect of the embodiments of this disclosure, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this disclosure.

[0018] According to another aspect of the embodiments of this disclosure, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods of various embodiments of this disclosure.

[0019] According to another aspect of the embodiments of this disclosure, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this disclosure.

[0020] In this embodiment, an on-board unit configured on a vehicle, in response to receiving a resource allocation request, can acquire vehicle status data and communication load data of the area where the vehicle is located. The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information refers to information about objects around the vehicle that obstruct communication signal transmission. The communication load data is used to represent the occupancy status of multiple sub-channels. Then, based on the status data, a target quality index for multiple sub-channels within a preset future time period can be determined. The target quality index characterizes the carrying capacity of multiple sub-channels. Next, based on the target quality index and communication load data, sub-communication resources corresponding to the target sub-channels can be allocated to the vehicle. The communication resource allocation method provided in this disclosure, by acquiring vehicle status data and communication load information, obtains a dynamic and comprehensive communication environment profile, and then determines the target quality index, which can proactively assess the future carrying capacity of each sub-channel. The target quality index can comprehensively reflect the quality and availability of each sub-channel, providing a quantitative basis for resource allocation. Based on the target quality index and communication load data, a suitable target sub-channel is actively selected for the vehicle, and corresponding sub-communication resources are allocated. By estimating the target quality index, potential channel congestion and interference can be predicted in advance. This allows for the avoidance of inefficient channels during resource allocation, effective interference coordination, and reduced communication latency. Resource allocation can be dynamically adjusted based on the vehicle's real-time status and environmental changes. Furthermore, it enables rapid response when vehicles are moving at high speeds, suddenly changing lanes, or entering obstructed areas, reducing the probability of resource collisions and improving the dynamic adaptability of the communication link. This addresses the technical challenges of low communication efficiency and stability in vehicle-to-everything (V2X) communication in related technologies. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of this disclosure and form part of this disclosure, illustrate exemplary embodiments of the present disclosure and are used to explain the disclosure, but do not constitute an undue limitation of the disclosure. In the drawings:

[0022] Figure 1 is a flowchart of a communication resource allocation method according to an embodiment of the present invention;

[0023] Figure 2 is a schematic diagram of a communication system architecture according to an embodiment of the present invention;

[0024] Figure 3 is a schematic diagram of a communication resource allocation device according to an embodiment of the present invention. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present disclosure, the technical solutions of the present disclosure will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present disclosure, and not all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present disclosure.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] According to embodiments of this disclosure, an embodiment of a communication resource allocation method is provided. The steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in a different order than that shown here.

[0028] This embodiment provides a communication resource allocation method applied to an on-board unit in a vehicle. Figure 1 is a flowchart of the communication resource allocation method according to an embodiment of this disclosure. As shown in Figure 1, the process includes the following steps:

[0029] The aforementioned vehicles can refer to transportation vehicles equipped with communication capabilities, enabling them to communicate autonomously and intelligently with other vehicles or roadside infrastructure on the road. In high-dynamic scenarios, such as highways and complex urban road conditions, vehicles need to communicate with other entities in real time and efficiently to ensure traffic safety, improve traffic efficiency, and support functions such as autonomous driving.

[0030] The aforementioned vehicle-mounted unit can refer to an electronic device installed in a vehicle. The vehicle-mounted unit can integrate a series of electronic components and software algorithms, enabling the vehicle to communicate and make corresponding communication decisions based on vehicle dynamics and environmental changes.

[0031] Step S102: In response to receiving a resource allocation request, obtain the vehicle's status data and the communication load data of the area where the vehicle is located.

[0032] The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information refers to the information of objects around the vehicle that obstruct the transmission of communication signals. The communication load data is used to represent the occupancy status of multiple sub-channels.

[0033] The aforementioned resource allocation request can refer to a communication resource demand signal issued by an on-board unit in the vehicle when it needs to communicate. In response to the on-board unit's need to send or receive information, it can send a resource allocation request to the vehicle system, indicating that it needs to use communication resources for data exchange.

[0034] The aforementioned status data can refer to various information data about the vehicle's real-time operation, including but not limited to the vehicle's position, speed, acceleration, heading angle, target obstruction information, and vehicle type. Status data constitutes the vehicle's operational status at a specific point in time, helping to predict the communication environment, forecast communication quality, and improve resource allocation.

[0035] The aforementioned location refers to the vehicle's specific coordinates in space, which can be obtained through methods such as navigation satellite systems and real-time dynamic differential positioning technology. Accurate location information helps vehicles identify their communication distance with surrounding vehicles and infrastructure, as well as whether there are obstructions affecting the communication path.

[0036] The vehicle speed mentioned above refers to the vehicle's current speed. Vehicle speed information is crucial for the allocation of communication resources in high-speed moving scenarios. Vehicle speed information helps predict the vehicle's future trajectory and position, thereby effectively adjusting the allocation of communication resources and ensuring a continuous and stable communication link.

[0037] The heading angle mentioned above can refer to the angle of the vehicle's driving direction, such as the angle between the vehicle's actual driving direction and the direction of the vehicle's longitudinal centerline. The heading angle helps to predict the relative position changes of the vehicle with other vehicles or communication nodes, and improve the strategy for establishing and maintaining communication links.

[0038] The aforementioned target obstruction information refers to information about obstacles around the vehicle that may affect communication signal transmission. This information can include the obstacle's location, size, shape, and scattering characteristics on the communication signal. Target obstruction information can be obtained through the fusion of millimeter-wave radar point cloud data and visual sensor image information. It can be used to compensate for path loss in non-line-of-sight scenarios and improve communication parameters. The specific target obstruction information can be determined according to actual needs.

[0039] The aforementioned communication load data can refer to data reflecting the occupancy status of multiple sub-channels in the current communication system, including but not limited to information such as sub-channel usage frequency and bandwidth utilization. Communication load data can be used to assess the overall communication load in order to dynamically adjust resource allocation strategies and avoid resource conflicts.

[0040] The aforementioned sub-channels can refer to frequency band units configured for data transmission in a communication system. Each sub-channel can have a specific frequency band and bandwidth, and multiple sub-channels can collectively form a communication resource pool available for vehicles. Dividing the spectrum into sub-channels helps to achieve efficient utilization of spectrum resources.

[0041] As an optional implementation, the on-board unit can collect and process data in real time to obtain the vehicle's current status data and the communication load data of the area. Specifically, the on-board unit can respond to a received resource allocation request and activate the data acquisition process. The on-board unit can utilize various sensors on the vehicle, such as positioning systems, LiDAR, millimeter-wave radar, and vision sensors, to collect vehicle status data, which may include motion information such as position, speed, and heading angle, as well as millimeter-wave radar point cloud data and vision sensor image data. Real-time acquisition of status data ensures the timeliness of communication resource allocation decisions and helps to promptly address communication obstacles caused by high-speed vehicle movement.

[0042] Information on target obstructions can be obtained in the following ways: Radar can be used to process point cloud data through clustering algorithms to identify potential obstacles around the vehicle, generating an obstruction heatmap that accurately reflects the obstructed area. Visual sensors can also be used for image semantic segmentation to identify obstacle types, such as other vehicles, concrete structures, and green belts, assigning compensation values ​​to each obstruction type to quantify the impact of obstacles on signal attenuation. By fusing the obstruction heatmap from radar point clouds and the obstruction type identification results from visual sensors, target obstruction information can be generated. The onboard unit can also collect communication load data for each sub-channel within the vehicle's area, including sub-channel occupancy status, historical communication quality indicators, and network load rate. This data can be obtained through collaborative communication with roadside units and neighboring vehicles, providing a comprehensive reflection of the current communication environment's load status and offering dynamic reference for resource allocation decisions.

[0043] In the above process, by accurately acquiring vehicle status data, the vehicle's surrounding environment can be perceived in real time and accurately. In scenarios such as high-speed movement, obstruction events can be effectively predicted, and measures can be taken in advance to improve the communication link. Acquiring communication load data of the area where the vehicle is located enables the onboard unit to assess the resource occupancy status of the current communication environment, providing a basis for subsequent resource allocation decisions.

[0044] Step S104: Based on the status data, determine the target quality index of multiple sub-channels within a future preset time period.

[0045] The target quality index is used to characterize the carrying capacity of multiple sub-channels.

[0046] The aforementioned future preset time period can refer to the time period considered when making communication quality predictions and resource allocation decisions. For example, it is possible to predict the quality changes of the communication channel within the next 300ms, so as to make resource allocation decisions in advance and provide better channel conditions for vehicle communication. The specific future preset time period can be determined according to actual needs.

[0047] The aforementioned target quality index can be a quantitative indicator describing the carrying capacity of a sub-channel. The target quality index reflects the level of communication quality that a sub-channel can support under specific conditions and can be used to guide resource allocation and parameter improvement.

[0048] As an optional implementation, the channel quality of multiple sub-channels within a preset future time period can be predicted based on state data, and the carrying capacity of different sub-channels within the preset future time period can be quantified to obtain a target quality index. This process can utilize vehicle state data and can also combine historical communication quality data to perform spatiotemporal joint prediction through a communication quality prediction model. The communication quality prediction model can encode and process the state data into machine-readable input features, and can integrate vehicle dynamic characteristics and environmental occlusion to predict spatiotemporal features. An attention mechanism can also be used, allowing the communication quality prediction model to pay more attention to certain spatiotemporal features during the prediction process, thereby improving the accuracy of the prediction. A learning mechanism can also be applied to dynamically adjust the target quality index based on the predicted spatiotemporal features.

[0049] In the above process, by predicting the communication quality of multiple sub-channels within a preset time period, resource allocation can be proactive, avoiding resource mismatch or conflict caused by rapid vehicle movement. Dynamically adjusted target quality indices can guide onboard units to more rationally select and use sub-channels, avoiding excessive competition during resource-constrained periods, and improving channel utilization and the overall performance of the communication system.

[0050] Step S106: Based on the target quality index and communication load data, allocate sub-communication resources corresponding to the target sub-channel to the vehicle.

[0051] The aforementioned target sub-channel can refer to a sub-channel that is specifically assigned to a vehicle. The determination of the target sub-channel can be based on factors such as predicted communication quality, current load status, and vehicle communication priority.

[0052] The aforementioned sub-communication resources may refer to the specific communication resources of the target sub-channel allocated to the vehicle, which may include, but are not limited to, frequency bands, bandwidth, time and space resources.

[0053] As an optional implementation, sub-communication resources corresponding to target sub-channels can be allocated to vehicles based on the target quality indices of multiple sub-channels within a predetermined future time period and communication load data. This process may include dynamic resource pool management, conflict resolution strategies, and utility maximization decisions. Dynamic resource pool management can adaptively adjust the size of the resource pool according to the current network load rate. The size of the resource pool can affect the number of candidate resources, thereby affecting the flexibility and efficiency of resource allocation. It can ensure that resources are available for allocation in high-load scenarios. The dynamic adjustment strategy can match the size of the resource pool with communication needs, avoiding waste caused by over-configuration when resources are abundant, while ensuring sufficient communication capacity when resources are scarce.

[0054] In the above process, the resource allocation strategy based on the target quality index and communication load data can maximize the utilization of high-quality sub-channels while avoiding wasting resources on sub-channels with poor signal quality, thereby improving the overall quality and stability of the communication link. This achieves intelligent and efficient allocation of communication resources, significantly enhancing communication performance in highly dynamic scenarios.

[0055] In this embodiment of the disclosure, the on-board unit applied to a vehicle, in response to receiving a resource allocation request, can acquire the vehicle's status data and the communication load data of the area where the vehicle is located. The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information is information about objects around the vehicle that obstruct the transmission of communication signals. The communication load data is used to represent the occupancy status of multiple sub-channels. Then, based on the status data, the target quality index of multiple sub-channels can be determined within a future preset time period. The target quality index is used to characterize the carrying capacity of multiple sub-channels. Then, based on the target quality index and the communication load data, sub-communication resources corresponding to the target sub-channels can be allocated to the vehicle. It is noteworthy that the communication resource allocation method disclosed herein obtains a dynamic and comprehensive communication environment profile by acquiring vehicle status data and communication load information, and then determines a target quality index. This allows for a forward-looking assessment of the future carrying capacity of each sub-channel. The target quality index comprehensively reflects the quality and availability of each sub-channel, providing a quantitative basis for resource allocation. Based on the target quality index and communication load data, the system proactively selects suitable target sub-channels for the vehicle and allocates corresponding sub-communication resources. By estimating the target quality index, the system can predict potential channel congestion and interference in advance, thereby avoiding these inefficient channels during resource allocation, achieving effective interference coordination, reducing communication latency, and dynamically adjusting resource allocation according to the vehicle's real-time status and environmental changes. It can also respond quickly when the vehicle is moving at high speed, suddenly changing lanes, or entering an obstructed area, reducing the probability of resource collisions and improving the dynamic adaptability of the communication link. This solves the technical problem of low communication efficiency and stability in vehicle-to-everything (V2X) communication in related technologies.

[0056] In the above embodiments of this disclosure, determining the target quality index of multiple sub-channels within a future preset time period based on state data includes: obtaining the historical quality index of multiple sub-channels; fusing the state data and historical quality value data using a communication quality prediction model to obtain spatiotemporal features; and predicting the target quality index of multiple sub-channels within a future preset time period using the communication quality prediction model and the spatiotemporal features.

[0057] The aforementioned historical quality index can refer to the communication performance indicators of each sub-channel in a communication system over a past period of time, including but not limited to reference signal received power, signal-to-noise ratio, and packet error rate. Historical quality indices can serve as input information for communication quality prediction models, helping them understand the past performance of each sub-channel and thus predict future quality trends.

[0058] The aforementioned communication quality prediction model can refer to a hybrid architecture based on deep learning techniques, such as multilayer perceptrons and long short-term memory networks. By learning the complex relationships between vehicle state data, including position, speed, heading angle, and target obstruction information, and historical communication quality data, the model can predict the quality index of sub-channels in future communication environments. The communication quality prediction model can also incorporate attention mechanisms to highlight spatiotemporal features that are more important to communication quality prediction, thereby improving the accuracy and reliability of the prediction.

[0059] The aforementioned spatiotemporal characteristics refer to communication quality-related features that combine spatial and temporal dimensions. Spatial characteristics can include the vehicle's position relative to the target channel, the distribution of obstructions, etc., while temporal characteristics can reflect the trend of communication quality changes over time. By fusing state data with historical quality indices, the communication quality prediction model can capture the spatiotemporal factors affecting communication quality, such as channel changes caused by vehicle movement and the evolution of the impact of obstructions on the signal over time.

[0060] As an optional implementation, the vehicle-mounted unit can acquire historical quality indices for multiple sub-channels from the communication system. These historical quality indices reflect the performance of each sub-channel in previous communication services. The vehicle-mounted unit can also collect current vehicle status data, including dynamic information such as position, speed, and heading angle, as well as the type and location of obstructions. Next, the status data and historical quality indices can be input into a communication quality prediction model for encoding. The spatial feature extraction module in the communication quality prediction model can analyze the relative spatial relationship between the vehicle and obstructions, extracting location information related to communication quality. The temporal feature extraction module in the communication quality prediction model can analyze the changing patterns of communication quality over time, capturing the dynamic characteristics of the channel state. By fusing information from both spatial and temporal dimensions, the communication quality prediction model can generate a representation containing spatiotemporal features to reflect the complexity of the vehicle's communication environment, including the impact of obstructions and the effect of vehicle speed changes on the channel. The output module of the communication quality prediction model can apply a learning mechanism to predict the target quality index of each sub-channel within a preset time period. By learning the relationship between vehicle status and communication environment changes and communication quality, the learning mechanism can dynamically adjust the parameters of the prediction model to output a more accurate target quality index prediction.

[0061] In the process described above, by integrating state data with historical quality indices, the communication quality prediction model can accurately capture the multidimensional factors affecting communication quality, providing more accurate communication quality predictions and helping to allocate resources in complex and ever-changing environments.

[0062] In the embodiments described above, the communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; the communication quality prediction model is used to fuse state data and historical quality value data to obtain spatiotemporal features, including: encoding the state data and historical quality indices using the encoding module to obtain input features; extracting the spatial positional relationships between the input features using the spatial feature extraction module to obtain spatial features; and extracting the temporal dependencies between the spatial features using the temporal feature extraction module to obtain spatiotemporal features; preferably, the communication quality prediction model further includes: an attention module and an output module; the communication quality prediction model and the spatiotemporal features are used to predict a target quality index, including: performing attention processing on the spatiotemporal features using the attention module to obtain attention features; and performing Q-value learning on the attention features using the output module to obtain the target quality index.

[0063] The aforementioned encoding module refers to the module in the communication quality prediction model that encodes the input data. The encoding module can convert state data and historical quality indices into numerical forms or feature vectors that the communication quality prediction model can process, inputting them into the next layer of the model in a unified data format. For example, vehicle speed, obstruction information, etc., need to be converted into numerical or vector representations through a specific encoding mechanism to facilitate analysis and processing by subsequent modules.

[0064] The aforementioned spatial feature extraction module can refer to a model configured in a communication quality prediction model to analyze the spatial relationships between input features. By utilizing techniques such as convolutional neural networks, the spatial feature extraction module can identify spatial patterns and structures that affect communication quality.

[0065] The aforementioned time-series feature extraction module refers to the module configured in the communication quality prediction model to extract time-series data, i.e., the changing trend of historical quality indices over time. The time-series feature extraction module can use recurrent neural network techniques such as Long Short-Term Memory networks or gated recurrent units to capture the dynamic changes in communication quality and provide time-dependent information for prediction.

[0066] The aforementioned spatial features can refer to the feature vectors output by the spatial feature extraction module, which reflect the spatially related attributes in the input features. Spatial features can describe the spatial structure of the environment in which the vehicle is located.

[0067] The aforementioned spatiotemporal features can refer to feature data that combines spatial features with temporal features. Spatiotemporal features comprehensively consider the combined impact of spatial configuration and temporal changes on communication quality.

[0068] The aforementioned attention module refers to an attention module that uses an attention mechanism to identify key information fragments from spatiotemporal features. The attention module can dynamically adjust weights based on the importance of spatiotemporal features, enabling the communication quality prediction model to pay more attention to factors that have a more significant impact on communication quality during the prediction process.

[0069] The aforementioned output module can refer to a communication quality prediction model configured to transform attention features into specific target quality indices. The output module can employ a Q-value learning mechanism, learning and improving strategies to output the target quality index of the sub-channel within a preset future time period.

[0070] The attention features mentioned above refer to the weighted features resulting from the attention module's processing, which incorporates information on the importance of spatiotemporal features. Attention features focus more on the key factors in predicting communication quality, thus helping to improve prediction accuracy.

[0071] The Q-score learning mentioned above refers to a reinforcement learning algorithm that iteratively learns and evaluates the expected reward (Q-score) of taking different actions under different states, thereby improving the decision-making strategy. In communication quality prediction, Q-score learning can be used to determine the predicted quality values ​​of different sub-channels at a future point in time, as well as the corresponding communication strategies.

[0072] As an optional implementation, the encoding module in the communication quality prediction model can be used to encode state data and historical quality values ​​into numerical or vector representations to obtain input features. Subsequently, the spatial feature extraction module can be used to analyze the spatial relationships between input features, extracting spatial features that reflect the spatial structure of the vehicle's environment. Next, the temporal feature extraction module can be used to identify the patterns of change in input features over time, capturing the temporal dependence of communication quality and obtaining temporal features. Then, spatial and temporal features can be fused to generate spatiotemporal features. An attention module can further process the spatiotemporal features, highlighting features with a significant impact on communication quality through a weight adjustment mechanism, resulting in attention features. This allows for focus on more critical aspects of the prediction, improving prediction quality. Finally, the output module can use a Q-value learning algorithm to process the attention features and generate a target quality index. Q-value learning, through iterative updates, continuously improves the predictive capability of the communication quality prediction model, ensuring that the output target quality index accurately reflects the carrying capacity of each sub-channel within a preset future time period.

[0073] In the above process, the combination of the encoding module, spatial feature extraction module, and temporal feature extraction module can capture multi-dimensional and complex information affecting communication quality, improving prediction accuracy. The application of the attention module enables the communication quality prediction model to automatically filter out the most critical information for communication quality prediction when processing large amounts of data, effectively avoiding information redundancy and overfitting. Through dynamic adjustment via Q-score learning, the communication quality prediction model can adapt to rapid changes in the communication environment, ensuring real-time effectiveness of prediction results even when vehicles are moving at high speeds or obstructions are frequently changing.

[0074] In the above embodiments of this disclosure, allocating sub-communication resources corresponding to a target sub-channel to a vehicle based on a target quality index and communication load data includes: determining the number of candidate resources in a target resource pool based on communication load data, wherein the target resource pool contains sub-communication resources corresponding to the target sub-channel; adjusting the sub-communication resources corresponding to multiple sub-channels contained in the original resource pool based on the number of candidate resources to obtain a target resource pool; and allocating sub-channel resources corresponding to the target sub-channel to the vehicle from the target resource pool based on the target quality index and status data; preferably, the original resource pool is a resource pool that matches the target communication priority of the vehicle among multiple resource pools, wherein different resource pools correspond to different communication priorities.

[0075] The aforementioned target resource pool can refer to a dynamically generated set containing specific sub-channel resources based on current communication needs and environmental conditions. The target resource pool provides optimal resource allocation for specific communication tasks, such as secure message transmission, ensuring high efficiency and reliability of communication.

[0076] The aforementioned number of candidate resources can refer to the number of sub-channel resources available in the target resource pool. The number of candidate resources can be dynamically adjusted based on communication load data and expected communication needs to ensure that resources that meet critical communication tasks can be reasonably allocated when resources are limited.

[0077] The aforementioned initial resource pool can refer to a collection of multiple sub-channel resources initially configured in a communication system. The initial resource pool may contain available communication resources, but these have not yet been improved or adjusted according to the priorities or requirements of a specific scenario.

[0078] The communication priority mentioned above refers to the priority obtained by prioritizing different resource pools based on the importance and urgency of the communication content. In highly dynamic scenarios, safety-related communications, such as emergency braking warnings, have higher priority and require priority resource allocation.

[0079] As an optional implementation, the division and adjustment of the dynamic resource pool can be based on communication load data and vehicle status data. Specifically, the original resource pool can be determined based on the vehicle's basic communication needs and the frequency band it operates in. Due to differences in communication needs under different scenarios, such as in environments with dense traffic or severe obstruction, communication priorities and resource requirements will change. The original resource pool can be adjusted to generate a target resource pool that better meets the current needs. Specifically, the vehicle unit can determine the number of candidate resources for the target resource pool based on real-time communication load data, including network load rate and channel occupancy status. For example, in response to a low network load rate, the number of candidate resources is small to prevent resource waste. As the network load rate increases, the number of candidate resources increases to cope with higher communication demands. Based on the number of candidate resources, the vehicle unit can adjust the sub-channel resources in the original resource pool. For example, in high-load scenarios, the number of high-frequency sub-channels used for secure communication in the original resource pool can be increased, while low-frequency resources can be reduced to ensure priority transmission of critical information. This process can be performed automatically by an algorithm to ensure dynamic matching between the resource pool and actual communication needs. Finally, based on the target quality index obtained from the communication quality prediction model and vehicle status data, the on-board unit selects sub-channels with better quality and that better meet current communication needs from the target resource pool for resource allocation, ensuring that communication resources can meet the priority of communication tasks. It can also make adjustments in advance based on predicted changes in the future communication environment, achieving efficient resource utilization and improved communication performance.

[0080] The dynamic adjustment of the number of candidate resources and the target resource pool in the above process enables flexible and efficient allocation of communication resources based on actual communication needs and load conditions, avoiding resource waste or shortages under static resource allocation. Resource pool adjustment based on target communication priorities ensures the transmission of safety-critical information is prioritized, reducing latency and error rates in high-priority communications, thus improving driving safety and traffic efficiency.

[0081] In the above embodiments of this disclosure, allocating sub-channel resources corresponding to a target sub-channel to a vehicle from the target resource pool based on the target quality index and state data includes: determining an initial resource allocation strategy for the vehicle from the target resource pool based on the target quality index; updating the initial resource allocation strategy based on the state data to obtain a target resource allocation strategy; and allocating sub-channel resources corresponding to a target sub-channel to the vehicle based on the target resource allocation strategy.

[0082] The aforementioned initial resource allocation strategy can refer to a communication resource allocation strategy based on a target quality index, without considering current specific state information. The initial resource allocation strategy can be a relatively coarse allocation based on predictions of the communication quality of each sub-channel over a future period.

[0083] The aforementioned target resource allocation strategy refers to a resource allocation strategy that improves and adjusts the initial resource allocation strategy after analyzing vehicle status data and combining it with real-time communication requirements. The target resource allocation strategy can be viewed as a more refined and personalized communication resource configuration scheme. The generation of the target resource allocation strategy can be a dynamic process, continuously updated based on real-time changes in vehicle status data such as speed, position, heading angle, and obstruction information.

[0084] As an optional implementation, the formulation and updating of the resource allocation strategy can be based on continuous improvement of vehicle status data and communication quality prediction results. Specifically, when the on-board unit receives a resource allocation request, it can determine a preliminary set of sub-channel resources from the target resource pool based on the target quality index given by the communication quality prediction model, serving as the vehicle's initial communication link. Next, the on-board unit can update the initial resource allocation strategy using real-time vehicle status data, including position, speed, heading angle, and obstruction information. Utility calculation can be performed during this process, comparing communication performance indicators under different resource allocation schemes and selecting the strategy that maximizes communication quality and efficiency as the target resource allocation strategy. The on-board unit can calculate the utility value under the current strategy and dynamically adjust the resource allocation based on the difference between the current utility value and the utility value of the updated strategy until a strategy that maximizes utility or meets predetermined convergence conditions is found. Finally, the on-board unit can select specific target sub-channels from the target resource pool for resource allocation according to the target resource allocation strategy, ensuring an optimal match between vehicle communication resources and the current status and predicted communication environment. This approach takes into account communication quality predictions and incorporates real-time vehicle status information, making resource allocation more precise and personalized.

[0085] In the above process, the update strategy based on vehicle status data ensures that communication resource allocation is more aligned with the vehicle's current driving status and communication environment, enabling personalized resource improvements and enhancing communication quality and efficiency. The dynamically adjusted target resource allocation strategy can quickly respond to changes in vehicle speed and environment, preventing communication interruptions or performance degradation due to improper resource allocation when communication quality deteriorates or obstructions appear, thus ensuring the continuity and stability of inter-vehicle communication.

[0086] In the above embodiments of this disclosure, updating an initial resource allocation strategy based on state data to obtain a target resource allocation strategy includes: determining a first utility value corresponding to the initial resource allocation strategy based on state data, wherein the first utility value is used to characterize the benefit obtained by allocating resources to vehicles based on the initial resource allocation strategy; updating the initial resource allocation strategy to obtain an updated resource allocation strategy; determining a second utility value corresponding to the updated resource allocation strategy based on state data; determining a utility difference between the first utility value and the second utility value; in response to a utility difference greater than or equal to a preset value, using the updated resource allocation strategy as the initial resource allocation strategy, and repeatedly executing the steps of determining the first utility value corresponding to the initial resource allocation strategy based on state data, updating the initial resource allocation strategy to obtain the updated resource allocation strategy, and determining the second utility value corresponding to the updated resource allocation strategy based on state data, until the utility difference is less than the preset value; in response to a utility difference less than the preset value, determining the updated resource allocation strategy as the target resource allocation strategy.

[0087] The aforementioned first utility value can refer to the value used to quantitatively evaluate the communication performance benefit that specific communication resources allocated to a vehicle can bring under the current state data. The first utility value can comprehensively consider multiple factors, such as communication quality, resource utilization, vehicle speed and direction, and express the effectiveness of the resource allocation strategy in numerical form.

[0088] The aforementioned second utility value can refer to the utility value calculated based on the new strategy and current state data after adjusting or updating the initial resource allocation strategy. The second utility value can reflect the expected communication performance of the updated resource allocation strategy in the current environment and is used to evaluate the potential performance improvement after strategy adjustment.

[0089] The aforementioned utility difference can refer to the difference between the second utility value and the first utility value. By comparing the utility difference with the preset value, it can be determined whether the current strategy needs further improvement and whether the degree of improvement is significant.

[0090] The aforementioned preset value can refer to a pre-defined threshold for a smaller improvement in utility, which can be used to determine whether the resource allocation strategy update has achieved the expected goals. If the utility difference exceeds the preset value, it indicates that the adjustment of the resource allocation strategy has brought about a significant improvement in communication performance, and further improvement is needed. If the utility difference is below the preset value, it can be considered that the resource allocation strategy has stabilized relatively, and the improvement process can be stopped.

[0091] As an optional implementation, improving the resource allocation strategy can be an iterative process, including evaluating and adjusting the current strategy, and re-evaluating the new strategy. Specifically, the on-board unit can calculate the communication performance gain, i.e., the first utility value, under the current initial resource allocation strategy based on the vehicle's current state data, such as position, speed, heading angle, obstruction information, and communication quality prediction. Then, the on-board unit can adjust the current resource allocation strategy using improved algorithms, such as gradient descent or genetic algorithms, based on changes in vehicle state and communication environment, generating an updated resource allocation strategy. This may include resource reallocation, transmission power adjustment, or changes in modulation and coding schemes. After implementing the updated resource allocation strategy, the on-board unit can calculate the communication performance gain under the updated strategy, i.e., the second utility value, based on real-time state data, to evaluate the performance improvement effect of the updated strategy. Then, the difference between the first and second utility values, i.e., the utility difference, can be calculated and compared with a preset value. If the utility difference is greater than or equal to the preset value, it can be determined that the updated strategy brings a significant performance improvement. The on-board unit can use the updated resource allocation strategy as the new initial strategy and restart the above process to further explore performance improvement potential. If the utility difference is less than the preset value, it can be determined that the resource allocation strategy has reached a better or sub-better state than the current state, the improvement process ends, and the updated strategy can be determined as the target resource allocation strategy.

[0092] In the above process, by calculating and comparing utility values, resource allocation strategies can be continuously iterated and improved, ensuring that the resource allocation strategies can meet real-time status data and communication needs, thereby improving the accuracy and efficiency of resource allocation. The evaluation and maximization of utility values ​​help maintain or improve communication performance in dynamically changing communication environments, reduce data transmission error rates, and ensure the accurate and timely transmission of information.

[0093] In the above embodiments of this disclosure, the method further includes: obtaining the occupancy status of the target sub-channel; determining a backoff time in response to the occupancy status indicating that the target sub-channel is occupied, wherein the backoff time is a waiting time for re-allocating communication resources to the vehicle; and regenerating a resource allocation request in response to the arrival of the backoff time. Preferably, obtaining the occupancy status of the target sub-channel includes: obtaining the reference signal received power of the target sub-channel; determining that the occupancy status indicates that the target sub-channel is idle in response to the reference signal received power being less than a preset power; and determining that the occupancy status indicates that the target sub-channel is occupied in response to the reference signal received power being greater than or equal to a preset power.

[0094] The backoff time mentioned above refers to a pre-set period of time during which the current vehicle must wait to retry resource allocation in response to the detection that the target subchannel has been occupied by other vehicles or equipment. The backoff time can be set based on communication technology standards and policies to avoid resource collisions and conflicts, ensuring the stable operation of the communication system.

[0095] The aforementioned reference signal received power refers to the received power of the reference signal received by the terminal device, which can be used as an indicator to measure the communication quality of the target sub-channel. A higher reference signal received power indicates better channel quality and a more stable signal. By monitoring the reference signal received power of the target sub-channel, the vehicle-mounted unit can determine the current occupancy status and quality level of the channel, providing a basis for resource allocation decisions.

[0096] The aforementioned preset power can refer to a power threshold pre-set in the communication system, which can be used to distinguish whether a channel is idle. If the received power of the reference signal received by the vehicle-mounted unit is lower than the preset power, it can be determined that the target sub-channel is idle. If the received power of the reference signal is greater than or equal to the preset power, it can be determined that the target sub-channel is occupied. The preset power can be adjusted according to communication technology and environmental factors to ensure the accuracy of the judgment.

[0097] As an optional implementation, the vehicle-mounted unit can continuously monitor the reference signal received power of the target sub-channel. The reference signal received power reflects the current communication quality and occupancy status of the channel. By setting a reasonable monitoring period, channel status information can be updated in real time, providing immediate reference for resource allocation decisions. If the vehicle-mounted unit detects that the reference signal received power of the target sub-channel is less than a preset power, it can determine that the target sub-channel is idle and resource allocation can be performed immediately. If the reference signal received power is greater than or equal to the preset power, it can determine that the target sub-channel is occupied, and the vehicle-mounted unit will not allocate resources immediately but will enter a backoff process. When the target sub-channel is occupied, the vehicle-mounted unit can enter a backoff mechanism and calculate the backoff time. The backoff time can be randomly generated based on a suitable algorithm, such as a binary exponential backoff algorithm, to ensure that when multiple vehicles compete for the same channel, resource allocation requests between vehicles avoid synchronous execution, thereby reducing the probability of resource collisions and improving communication efficiency. After the backoff time expires, the on-board unit can retry the resource allocation request, re-monitor the reference signal received power of the target sub-channel, and make resource allocation decisions based on the new status data. This mechanism can ensure that in an environment of high resource contention, the vehicle has the opportunity to regain communication resources and maintain the continuity and stability of communication.

[0098] In the above process, the backoff time and channel occupancy status detection mechanism effectively reduces resource collisions caused by multiple vehicles simultaneously competing for the same sub-channel, ensuring the rational allocation of communication resources and the stable operation of the communication system. Rapid resource allocation when the channel is idle reduces waiting time and improves communication efficiency. By monitoring channel occupancy status and allocating resources in a timely manner, the continuity and reliability of communication are guaranteed, preventing the loss of important information due to channel occupancy.

[0099] In the above embodiments of this disclosure, the method further includes: determining a signal power compensation value based on target obstruction information; determining the transmit power of the target sub-channel based on the target quality index and the signal power compensation value; and determining the modulation scheme and coding rate of the target sub-channel based on the target quality index.

[0100] The aforementioned signal power compensation value refers to the quantified value of the increased signal transmission power required to compensate for or offset the attenuation or interference caused by obstructions. In communications, obstructions such as buildings and vehicles can significantly affect the propagation of communication signals, leading to a weakening of signal strength. The determination of the signal power compensation value can be based on the type and distance of the obstruction, as well as an estimate of the degree of signal attenuation.

[0101] The modulation schemes mentioned above refer to the technical solutions used in digital communication to convert digital signals into analog signals suitable for transmission over physical channels. The choice of modulation scheme in communication can affect the spectral efficiency and interference resistance of signal transmission.

[0102] The coding rate mentioned above refers to the ratio of raw data to redundant data during data encoding in digital communication, reflecting the efficiency of data encoding and the reliability of communication. In communication, the choice of coding rate can find a balance between spectral efficiency and error correction capability.

[0103] As an optional implementation, in response to the onboard unit identifying obstruction information through visual sensors and radar point cloud data, a signal power compensation value can be calculated using preset compensation rules based on the type of obstruction, its distance, and the degree of signal attenuation. For example, for a metal truck within a distance of less than 30m, the signal power compensation value can be 15dB. For a concrete building within a distance of less than 50m, the signal power compensation value can be 8dB. The signal power compensation value can be used to dynamically adjust the transmission power to overcome the signal attenuation caused by obstructions. Next, based on the target quality index and the signal power compensation value, the onboard unit can calculate the optimal transmission power for the target sub-channel. For example, if the target quality index of the target sub-channel is low, it indicates poor channel quality. In this case, in addition to increasing the transmission power using the compensation value, the onboard unit can further increase the transmission power to ensure that the signal strength can penetrate the obstruction and improve communication quality. Then, a suitable modulation scheme and coding rate can be selected based on the target quality index. Subchannels with a high target quality index indicate good signal-to-noise ratio and high communication quality. In this case, higher-order modulation and higher coding rates can be used to improve spectral efficiency and speed up data transmission. Subchannels with a low target quality index can choose lower-order modulation schemes and lower coding rates, sacrificing spectral efficiency for higher communication reliability and reducing data transmission errors.

[0104] In the above process, determining the signal power compensation value can effectively overcome the signal attenuation caused by obstructions, ensuring the strength and stability of the communication signal. Dynamic adjustment of the modulation scheme and coding rate can improve spectrum utilization efficiency and accelerate data transmission speed while ensuring communication quality, meeting the needs of different types of data transmission.

[0105] In the above embodiments of this disclosure, obtaining target occlusion information includes: acquiring image data collected by a visual sensor and point cloud data collected by radar; identifying occlusions around the vehicle based on the point cloud data to obtain first occlusion information; identifying occlusions around the vehicle based on the image data to obtain second occlusion information; and fusing the first occlusion information and the second occlusion information to obtain target occlusion information. Preferably, fusing the first occlusion information and the second occlusion information to obtain target occlusion information includes: determining the first occlusion based on position, vehicle speed, and heading angle. The system calculates a first initial weight value for object information and a second initial weight value for second occluded object information; performs semantic understanding on the first occluded object information to obtain a first semantic understanding result, and performs semantic understanding on the second occluded object information to obtain a second semantic understanding result; adjusts the first initial weight value based on the first semantic understanding result to obtain a first target weight value, and adjusts the second initial weight value based on the second semantic understanding result to obtain a second target weight value; and performs weighted processing on the first occluded object information and the second occluded object information based on the first target weight value and the second target weight value to obtain target occluded object information.

[0106] The aforementioned image data can refer to environmental image information collected through visual sensors, such as cameras. It can contain rich visual features, such as color, texture, and shape, which helps to identify object types, distances, and environmental changes. In communication, it can be used to assist in judging occluded and unoccluded objects.

[0107] The point cloud data mentioned above can refer to datasets generated by radar sensors, such as millimeter-wave radar, which measure the distance, azimuth, and radar cross-section of reflecting objects by transmitting and receiving electromagnetic waves. Each point can represent the location and attributes of an obstacle or an object in the environment, helping to accurately determine the location and shape of obstructions and assess the quality of communication links.

[0108] The aforementioned first obstruction information may refer to the information of obstructions around the vehicle identified based on radar point cloud data. This information may include the location, shape, size, and possible material characteristics of the obstruction, which helps to assess signal attenuation under non-line-of-sight communication conditions.

[0109] The aforementioned second occlusion information can refer to occlusion information identified through image analysis and semantic understanding techniques based on image data collected by a visual sensor. This second occlusion information supplements the types of occlusions identified by radar, improving the accuracy and completeness of occlusion identification.

[0110] The aforementioned first semantic understanding result can refer to the identification of the type and semantic information of obstructions after processing radar point cloud data through a deep learning model. For example, it can distinguish different types of obstacles such as vehicles, buildings, and trees, which helps to assess the impact of obstructions on communication links.

[0111] The aforementioned second semantic understanding result can refer to the occlusion type and environmental information obtained after semantic segmentation of image data. This second semantic understanding result can provide visual features and material information of the occlusion, helping to more accurately assess the impact of the occlusion on the signal.

[0112] The aforementioned first target weight value refers to the weight of the first obstruction information adjusted based on the vehicle's position, speed, and heading angle after evaluating the obstruction information. The first target weight value reflects the relative importance of radar obstruction information under current driving conditions.

[0113] The aforementioned second objective weight value can refer to the weight of the second occlusion information adjusted based on vehicle state data, reflecting the relative importance of image occlusion information under current driving conditions. This second objective weight value can be used to determine the contribution ratio of image information to radar information during data fusion.

[0114] As an optional implementation, visual sensors and radar sensors can acquire image data and point cloud data from the environment in real time. Visual sensors can provide rich and detailed environmental information, including features such as the color and shape of objects. Millimeter-wave radar can provide precise geometric information such as the distance and orientation of objects. Point cloud data can be processed by a deep learning model, using clustering algorithms to identify first occlusion information, including the location, shape, and size of the occlusion. Image data is processed by an image semantic segmentation model to identify second occlusion information, supplementing the type and material information of the occlusion. Then, the weight of the occlusion information can be adjusted according to the real-time status of the vehicle, such as position, speed, and heading angle. For example, in high-speed driving scenarios, point cloud data can provide more accurate occlusion location information, and the weight value of the first target can be higher. At complex intersections, image data can provide detailed type and material of the occlusion, and the weight value of the second target can be higher. Then, the first and second occlusion information can be fed into a semantic understanding module to obtain first and second semantic understanding results, respectively. Based on the results of the first and second semantic understandings, the weights are further adjusted to improve the fusion process of occlusion information. For example, if the first semantic understanding indicates that the occlusion is a metallic structure, the first target weight value can be further increased because metal attenuates the signal more severely. If the occlusion is vegetation, the second target weight value can be relatively high because image data is more advantageous in identifying vegetation occlusion. Finally, the first and second occlusion information can be weighted and fused according to the adjusted first and second target weight values ​​to generate target occlusion information, providing more accurate and comprehensive occlusion perception data for subsequent communication quality prediction and resource allocation.

[0115] In the aforementioned process, the complementary advantages of radar point cloud data and image data, along with weight adjustments based on state data, result in more accurate and comprehensive identification results. This effectively identifies various types of obstructions, improving the planning and improvement capabilities of communication links. Based on more accurate obstruction information, the communication system can make more rational resource allocation decisions, such as adjusting transmission power and selecting communication channels, effectively avoiding the communication quality degradation caused by obstructions.

[0116] The technical solution proposed in this disclosure is described below with reference to an optional embodiment. This disclosure proposes an adaptive vehicle-to-everything (V2X) communication resource allocation and interference coordination method for highly dynamic scenarios. This disclosure belongs to the field of vehicle-to-everything (V2X) communication technology in intelligent transportation systems, specifically involving an adaptive resource allocation and interference coordination method for high-speed vehicle movement scenarios. This disclosure addresses cellular V2X direct communication systems based on an interface (3GPP PC5). In complex scenarios such as high-speed vehicle movement (e.g., greater than 60 km / h), rapid topology changes (e.g., intersection merging), and dense occlusion non-line-of-sight (NLOS), this disclosure achieves dynamic resource allocation and interference suppression through multi-source perception fusion and distributed collaborative decision-making. This disclosure is applicable to the improvement of direct communication links in V2X systems.

[0117] This disclosure aims to effectively address the following key issues: Eliminating channel state prediction inaccuracies caused by high-speed vehicle movement; accurately controlling position errors within 3 meters at vehicle speeds up to 100 km / h, thus improving the reliability and accuracy of the communication system. In environments with sudden interference, controlling the packet error rate to within 10% with a confidence level exceeding 95%, ensuring accurate data transmission and providing a solid guarantee for safety-critical applications such as autonomous driving. Successfully overcoming communication interruption challenges caused by environmental occlusion; reducing packet loss to below 35% in non-line-of-sight scenarios, improving communication continuity and stability, and maintaining good communication quality even in complex urban environments. Achieving dynamic adaptation to network load fluctuations, significantly increasing resource utilization to over 80%, improving resource allocation strategies, reducing resource waste, and enhancing the overall performance and capacity of the communication system to meet communication needs in different traffic scenarios.

[0118] To address the aforementioned issues, this disclosure proposes a multimodal data fusion system and method for vehicle-to-everything (V2X) based on edge computing. The specific technical solutions are as follows: A lightweight edge computing node design, employing modular design, supports on-demand expansion, reducing hardware costs and resource consumption. Field-programmable gate array (FPGA) acceleration technology is introduced to improve data processing speed and reduce latency. A multimodal data fusion method with dynamic weight adjustment can dynamically adjust fusion weights based on vehicle dynamic characteristics, such as speed, acceleration, and environmental complexity, improving fusion accuracy. Natural language processing (NLP) technology is introduced to semantically understand the vehicle's surrounding environment, further improving fusion weights. Adaptive communication protocols and data transmission paths are improved, employing frequency band adaptive technology to dynamically adjust data transmission paths and reduce communication latency. A software-defined network (SDN)-based communication protocol improvement method dynamically adjusts the network topology and improves data transmission paths. A localized decision-priority mechanism is implemented, with edge nodes possessing a localized decision-priority mechanism to prioritize tasks with high real-time requirements, ensuring rapid response to critical decisions. For tasks with high real-time requirements, decision results can be generated directly at the edge nodes, reducing reliance on the cloud. A reinforcement learning-based task offloading strategy dynamically determines whether to offload parts of the task to the cloud based on its complexity and real-time requirements. An improved task offloading threshold is introduced, dynamically adjusted to adapt to different traffic scenarios, thereby enhancing the overall performance of the communication system.

[0119] Figure 2 is a schematic diagram of a communication system architecture according to an embodiment of the present invention. As shown in Figure 2, it includes: a multi-source sensing fusion module, capable of radar point cloud clustering and visual semantic segmentation; a communication quality prediction module, capable of spatiotemporal joint prediction using a multilayer perceptron-long short-term memory model; a distributed decision-making module, capable of potential game improvement and generating a dynamic resource pool; a dynamic optimization module, capable of achieving closed-loop feedback; a power control table, capable of achieving adaptive signal-to-noise ratio; and a modulation and coding switching threshold, capable of driving the packet error rate.

[0120] To ensure the efficient and stable operation of this disclosure, the various components of the communication system can be deployed reasonably, and the specific deployment specifications are as follows:

[0121] The vehicle-mounted unit can undertake edge computing decision-making tasks, process and analyze the collected vehicle data in real time, and make decisions such as communication resource allocation quickly to ensure low latency and high efficiency of in-vehicle communication.

[0122] Millimeter-wave radar can collect point cloud data of the vehicle's surrounding environment in real time, including information such as the location and distance of obstacles, providing basic data support for occlusion perception modeling, helping to predict occlusion situations in advance, and improving communication links.

[0123] Visual sensors can acquire images of the vehicle's surrounding environment and identify obstacle types, such as metal trucks, concrete buildings, and vegetation, using visual semantic segmentation technology. The high frame rate and high dynamic range of visual sensors ensure clear and accurate images even under complex lighting conditions, providing rich environmental information for communication quality prediction.

[0124] Roadside units can be responsible for local traffic flow prediction, collecting and analyzing traffic flow data within a certain area, predicting changes in traffic conditions, distributing global load information to vehicles, helping vehicles understand the surrounding communication environment in advance, improving resource allocation strategies, and achieving effective management and coordination of large-scale traffic scenarios.

[0125] The real-time data acquisition in this disclosure, such as based on a 50ms period, is fundamental to the entire process in vehicle-to-everything (V2X) communication systems. Accurate and timely acquisition of vehicle status data is crucial for subsequent communication resource allocation and interference coordination. A VehicleState class is defined to encapsulate various key vehicle status information, specifically including: ID, position (pos), heading angle, point cloud data (radar_points), and image data (camera_frame).

[0126] The number can be a vehicle identifier, which can be used to distinguish different vehicles in the communication network and ensure that the data is accurately associated with the corresponding vehicle.

[0127] By combining navigation satellite systems with real-time dynamic differential positioning technology, centimeter-level vehicle positioning accuracy can be achieved, accurately determining the vehicle's position coordinates (x, y, z) in space, providing reliable location information for communication distance calculation, occlusion judgment, and other purposes.

[0128] Doppler radar can be used to calibrate and measure vehicle speed, obtaining real-time speed values, which helps predict vehicle trajectories and plan communication resources in advance. In high-speed moving scenarios, the accuracy of speed information can affect the rationality of resource allocation.

[0129] Heading angle information can reflect the vehicle's direction of travel. Combined with speed and position information, it can more comprehensively depict the vehicle's motion state, help predict the relative position changes of the vehicle and surrounding vehicles, and improve communication link establishment and maintenance strategies.

[0130] The point cloud coordinate data (x, y, RCS) acquired by millimeter-wave radar includes radar cross-section (RCS), which reflects the scattering characteristics of radar waves by obstacles. By analyzing the point cloud coordinates and RCS, the location, shape, and reflection characteristics of obstacles can be identified, providing crucial data support for occlusion perception modeling, and consequently influencing decisions such as communication path planning and power control.

[0131] The high dynamic range image data acquired by the visual sensor, with its high frame rate and high dynamic range characteristics, ensures the acquisition of clear and detailed images in complex lighting environments. This provides high-quality input data for visual semantic segmentation algorithms, accurately identifies obstacle types, and assists in communication quality prediction and resource allocation decisions.

[0132] During data acquisition, the onboard unit can acquire the aforementioned vehicle status information in real time through various sensors at 50ms intervals, and encapsulate the vehicle status information into a VehicleState object, providing a continuous, real-time data stream for subsequent processing modules. For example, in response to a vehicle traveling on urban roads, the onboard unit can update the vehicle's location information through the positioning system and measure the vehicle speed using Doppler radar every 50ms. Simultaneously, millimeter-wave radar and visual sensors can also collect point cloud data and image data of the surrounding environment in real time, jointly constructing a comprehensive and dynamic vehicle surrounding environment perception model, laying a solid foundation for subsequent communication improvement decisions.

[0133] The occlusion perception modeling disclosed herein is a crucial step in addressing communication challenges in non-line-of-sight scenarios. By fusing data from millimeter-wave radar and visual sensors, it accurately perceives occlusion around the vehicle, providing important data for communication quality prediction and resource allocation. A clustering algorithm (DBSCAN) can be used to process the point cloud data acquired by the millimeter-wave radar. The clustering algorithm parameters can be set to eps = 1.5m and min_samples = 10. Through cluster analysis, the point cloud data can be divided into different clusters, each representing a potential obstacle. Based on the clustering results, a two-dimensional occlusion heatmap can be generated with a grid accuracy of 0.5m × 0.5m, visually displaying the degree of occlusion in different areas around the vehicle. This provides a visual reference for communication path planning and power adjustment, helping the communication system to identify occlusion areas in advance and improve the communication link. Semantic segmentation processing can be performed on images acquired by the visual sensor, accurately identifying different obstacle types in the images, such as metal trucks, concrete buildings, and vegetation. Based on the different attenuation effects of different types of obstructions on communication signals, corresponding compensation values ​​can be assigned to different types of obstructions. The specific compensation values ​​are shown in Table 1 below:

[0134] Table 1

[0135] Based on the occlusion heatmap generated by visual semantic segmentation and radar point cloud clustering, path loss is dynamically compensated. For example, when a vehicle is traveling on an urban road and a metal truck appears ahead, the visual semantic segmentation algorithm can quickly identify the metal truck as an occlusion. Combining this with the distance to the vehicle (assuming it's within 30 meters), the algorithm can compensate for communication path loss according to pre-set rules, with a specific compensation value of 15dB. This dynamic compensation mechanism effectively addresses signal attenuation issues in occluded environments, improving the strength and quality of communication signals and ensuring a reliable communication link for vehicles in non-line-of-sight scenarios, thus guaranteeing accurate information transmission.

[0136] The communication quality prediction in this disclosure can be based on a multilayer perceptron-long short-term memory (LSSM) model. The performance of the communication quality prediction module can affect the accuracy and effectiveness of resource allocation decisions. Accurate spatiotemporal joint prediction of communication channel quality can be achieved by combining the LSM model with an attention mechanism. The training dataset can be 100km of measured urban road data, covering various typical scenarios, including complex environments such as tunnels, overpasses, and intersections. Data labels can include key communication quality indicators such as reference signal received power, packet error rate, and signal-to-noise ratio, with a sampling rate of 20ms. A rich training dataset ensures that the LSM model has good generalization ability in different types of scenarios and accurately predicts the trend of communication channel quality changes. Online inference can proceed as follows: a fully connected layer extracts spatial features, the LSM captures temporal dependencies, attention focuses on key nodes, and the sub-channel quality index is output. The input tensor can contain 12-dimensional feature data, which is then processed by a fully connected layer to extract spatial features and initially capture the correlation between the various feature dimensions. Subsequently, the Long Short-Term Memory (LSTM) layer processes the feature sequences, uncovers temporal dependencies, and extracts the dynamic changes in communication channel quality. The attention mechanism focuses on key time points and feature dimensions, enhancing attention to important information and improving the accuracy and reliability of predictions. Finally, through Q-score learning, the Quality Indices (CQIs) of each sub-channel are output, providing a quantitative basis for resource allocation decisions. The Multilayer Perceptron-LLSTM model meets the stringent low-latency requirements of real-time communication systems, ensuring timely and accurate resource allocation decisions in rapidly changing communication environments. For example, in a typical non-line-of-sight scenario where a vehicle is about to enter a tunnel, the communication quality prediction model can accurately predict the quality indices of each sub-channel within the next 300ms based on multi-dimensional feature data such as the vehicle's current position, speed, heading angle, and surrounding obstructions, combined with historical communication quality data, such as reference signal received power and signal-to-noise ratio. Based on the prediction results, the resource allocation decision module can allocate communication resources to higher-quality sub-channels in advance, while adjusting the transmission power and modulation and coding scheme, and improving the communication link to ensure that vehicles can maintain stable communication after entering the tunnel, thus ensuring the accurate transmission of driving safety-related information, such as vehicle status updates and surrounding traffic warnings.

[0137] The distributed resource allocation module in this disclosure can efficiently allocate and coordinate communication resources based on communication quality prediction results and the current load status of the communication system, combined with a dynamic resource pool partitioning algorithm and a game-theoretic decision-making mechanism. The resource pool size can adaptively adjust according to the network load rate (ρ). When the network load rate (ρ) is less than 0.3, the resource pool size is 5; when 0.3 is less than or equal to the network load rate (ρ) less than 0.7, the resource pool size is 8; and when the network load rate (ρ) is greater than or equal to 0.7, the resource pool size is 12. This dynamic adjustment mechanism ensures that the resource pool size matches the network load status, avoiding resource waste under low load and providing sufficient resources to meet communication needs under high load. Furthermore, security messages, such as basic security messages, are prioritized for allocation to the high-frequency resource pool, ensuring priority transmission and reliable reception of critical security information, guaranteeing communication priority for vehicles in emergency situations, and improving the overall security and reliability of the communication system.

[0138] This disclosure presents a three-stage conflict resolution strategy. During the monitoring phase, vehicles continuously monitor the received power of the reference signal. If the received power falls below a set threshold (e.g., -95dBm), the vehicle determines the current sub-channel is idle and can attempt to occupy resources. This monitoring mechanism helps vehicles understand channel occupancy in advance, avoiding resource collisions and conflicts caused by blind competition. If a busy channel or a conflict in resource allocation requests is detected, an exponential backoff algorithm is initiated, with a backoff window size of W = [2Npool, 1] seconds. By randomly selecting a backoff time, the vehicle re-attempts its resource allocation request after the backoff time expires. This exponential backoff strategy can, to some extent, disperse the resource allocation requests of vehicles, reduce the probability of resource collisions, alleviate channel congestion, and ensure the stable operation of the communication system. Based on an improved potential game model, each vehicle calculates its own utility value under different resource allocation strategies according to a defined utility function, and iteratively updates its strategy until a Nash equilibrium is reached. In the Nash equilibrium state, the resource allocation strategies of each vehicle reach a relatively stable and optimal state, achieving efficient allocation and competitive coordination of communication resources. For example, in a scenario involving 120 vehicles at an intersection, a game-theoretic decision-making process allows vehicles to quickly converge to a Nash equilibrium, achieving an orderly allocation of communication resources. This ensures accurate interaction and timely transmission of traffic information, guaranteeing the safety and efficiency of traffic flow at the intersection. In practical applications, when multiple vehicles simultaneously compete for communication resources at an intersection, each vehicle can calculate its utility value under its current strategy based on its position, speed, heading angle, and the resource occupancy of surrounding vehicles. Subsequently, a function can be called to solve for the Nash equilibrium, determine the optimal resource allocation strategy, and update the resource occupancy map. After multiple iterations, a Nash equilibrium is reached when the larger difference between the payoffs of two adjacent iterations is less than 0.01, completing the resource allocation process. This process can be completed within 20ms, ensuring the real-time performance and efficiency of the communication system.

[0139] To further improve communication performance, this disclosure achieves precise control and improvement of the communication link by dynamically adjusting the transmit power and modulation / coding scheme. The transmit power can be dynamically adjusted based on the predicted signal-to-noise ratio (SNR). Specific adjustment rules are shown in Table 2 below:

[0140] Table 2

[0141] For example, in a strong non-line-of-sight scenario where a vehicle is driving on a city street and surrounding buildings severely obstruct the communication signal, the predicted signal-to-noise ratio is 3dB. Based on a power control strategy, the communication system can increase the transmit power by 6dB to enhance signal penetration, ensuring effective transmission to the receiver, reducing communication interruptions and data loss due to obstruction, and improving communication continuity and stability.

[0142] The adaptive modulation and coding scheme mechanism in this disclosure can adaptively adjust the modulation and coding scheme according to the predicted packet error rate to achieve a better balance between spectral efficiency and communication reliability. The specific adjustment scheme is shown in Table 3 below:

[0143] Table 3

[0144] When the predicted packet error rate (PER) is low, such as less than 1%, it indicates good communication channel quality. In this case, a high-order modulation scheme, such as 64-quadrature amplitude modulation (64QAM), and a higher coding rate, such as 3 / 4, can be used to achieve high spectral efficiency, such as 4.5 bps / Hz. This can improve the data transmission rate of the communication system and meet the needs of large-volume data transmission, such as high-definition video streaming and real-time traffic map updates. However, under poor channel conditions with a predicted PER of greater than 5%, to ensure communication reliability and reduce the data transmission error rate, a low-order modulation scheme, such as quadrature phase shift keying (QPSK), and a lower coding rate, such as 1 / 3, can be used. Although the spectral efficiency decreases to 0.67 bps / Hz, it can significantly improve the anti-interference capability of communication and ensure the accurate transmission of critical information, such as safety warning messages. For example, on a highway, vehicles travel in a wide line-of-sight (LOS) environment, where the communication channel quality is good, and the predicted PER is 0.5%. At this point, the communication system can select a 64 quadrature amplitude modulation (QAM) modulation scheme with a coding rate of 3 / 4, achieving a spectral efficiency of 4.5 bps / Hz. This enables rapid transmission of various information, such as vehicle driving data and traffic updates, enhancing the user's communication experience. When the vehicle enters a tunnel and is in a strong non-line-of-sight scenario, and the predicted packet error rate rises to 8%, the communication system can quickly switch to a quadrature phase shift keying (QPSK) modulation scheme with a coding rate of 1 / 3. This ensures that even in adverse communication conditions, the vehicle can still stably receive safety-related warning information, such as road construction warnings and emergency braking notifications, guaranteeing driving safety.

[0145] Through the aforementioned implementation steps, including real-time data acquisition, occlusion perception modeling, communication quality prediction, distributed resource allocation, and dynamic parameter optimization, this disclosure can achieve precise allocation and effective coordination of vehicle-to-everything (V2X) communication resources in complex and highly dynamic scenarios, significantly improving the performance indicators of the communication system, such as reducing resource collision rate, reducing message latency, reducing packet error rate in non-line-of-sight scenarios, and improving channel utilization, thus providing strong support for the efficient, stable, and safe operation of intelligent transportation systems.

[0146] The radar-vision fusion non-line-of-sight compensation model and accurate communication quality prediction employed in this disclosure can detect occlusion and channel changes in advance, improve resource allocation decisions, and effectively avoid resource collisions and conflicts. Regarding message latency, this disclosure can reduce message latency, which is crucial for real-time communication applications in high-speed mobile scenarios, such as collaborative decision-making and control between autonomous vehicles. It ensures that vehicles receive and respond to messages sent by surrounding vehicles and traffic infrastructure in a timely manner, avoiding potential safety risks caused by excessive latency.

[0147] When addressing challenges in non-line-of-sight scenarios, this disclosure can reduce packet error rates. Due to the adoption of a dynamic parameter optimization mechanism, including adaptive adjustment of power control and modulation / coding schemes, communication parameters can be dynamically improved based on obstruction and channel quality, overcoming the attenuation effect of obstructions and enhancing communication reliability and stability. It can also improve channel utilization; the dynamic resource pool partitioning algorithm and distributed game theory decision-making mechanism can efficiently allocate and utilize limited communication resources, reducing resource waste and improving the overall capacity and efficiency of the communication system. Furthermore, this disclosure can quickly achieve Nash equilibrium, enabling rapid convergence of resource allocation decisions, ensuring the real-time performance and efficiency of the communication system, and meeting the demands of rapidly changing communication environments.

[0148] In high-speed merging zone early warning scenarios, the packet loss rate of basic safety messages exceeds 40% due to metal obstruction by guardrails in merging zones on highways. This significantly impacts vehicles' accurate perception and timely response to traffic conditions in merging areas, posing a substantial threat to driving safety. To address this issue, the solution disclosed herein is as follows: Millimeter-wave radar monitors the vehicle's surrounding environment in real time, identifying continuous metal reflection points. Combined with visual semantic segmentation results, these continuous metal reflection points are marked as dynamically obstructed areas, accurately locating the source and extent of the obstruction. A power compensation mechanism is triggered, increasing the transmission power by 9dB, enhancing the penetration capability of the communication signal, effectively overcoming the signal attenuation caused by metal guardrails, improving signal strength and quality, and ensuring reliable transmission of basic safety messages to the target vehicle. A dynamic resource pool management mechanism reserves a dedicated resource pool to provide priority and dedicated communication resources for the transmission of basic safety messages in merging zones, reducing resource contention and collisions, and ensuring stable transmission of critical safety information. The communication quality prediction model can switch to merging mode. In merging mode, the weight of historical communication data in the Long Short-Term Memory (LSTM) layer can be increased, making full use of historical communication patterns and current state information to more accurately predict the communication channel quality in the merging zone. This further improves resource allocation decisions and ensures that vehicles can receive and process basic safety messages in a timely manner in complex merging scenarios, allowing for proactive safety measures and preventing traffic accidents. Through these targeted measures, this disclosure effectively solves the communication interruption and packet loss problems caused by metal obstructions in high-speed merging zones, improving the communication reliability and stability in merging zones and providing strong protection for safe driving in high-speed merging scenarios.

[0149] In densely populated urban intersections, a large number of vehicles simultaneously compete for limited communication resources, causing the signal-to-noise ratio to drop sharply to -5dB. This severely degrades communication quality, impacting safe information exchange between vehicles and traffic flow. In this intersection scenario, multiple vehicles simultaneously engage in a game-theoretic decision-making process. Each vehicle first calculates its utility value under its current resource allocation strategy based on its position, speed, heading angle, and the resource occupancy of surrounding vehicles. This utility function comprehensively considers interference costs and communication quality gains, ensuring that vehicles minimize interference with other vehicles while meeting their own communication needs. Subsequently, a function is called to solve for the Nash equilibrium, determining the optimal resource allocation strategy and updating the resource occupancy map. Each vehicle adjusts its resource allocation decision based on the solution results, iteratively updating until a Nash equilibrium is reached. During this process, resource allocation conflicts between vehicles gradually decrease, and communication resources are rationally allocated and utilized. The entire game decision-making process can quickly converge to Nash equilibrium in a short period of time, achieving efficient allocation of communication resources. This ensures that in dense urban intersection scenarios, vehicles can transmit and receive safety-related information in a timely and accurate manner, such as vehicle turning intentions and emergency braking warnings. This effectively improves the safety and efficiency of intersection traffic, reduces the probability of traffic accidents, and improves the operation of urban traffic flow.

[0150] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0151] According to another aspect of the present invention, a communication resource allocation apparatus is also provided. This apparatus can execute the communication resource allocation method of the above embodiments. The specific implementation method and preferred application scenarios are the same as those of the above embodiments, and will not be described in detail here.

[0152] Figure 3 is a schematic diagram of a communication resource allocation device according to an embodiment of the present disclosure, applied to an on-board unit in a vehicle. As shown in Figure 3, the device includes the following: an acquisition module 302, a determination module 304, and an allocation module 306.

[0153] The acquisition module 302 is configured to acquire vehicle status data and communication load data of the area where the vehicle is located in response to receiving a resource allocation request. The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information refers to the information of objects around the vehicle that obstruct the transmission of communication signals. The communication load data is used to indicate the occupancy status of multiple sub-channels. The determination module 304 is configured to determine the target quality index of multiple sub-channels within a preset time period based on the status data. The target quality index is used to characterize the carrying capacity of multiple sub-channels. The allocation module 306 is configured to allocate sub-communication resources corresponding to the target sub-channels to the vehicle based on the target quality index and the communication load data.

[0154] The determination module is also configured to acquire the historical quality indices of multiple sub-channels; fuse the status data and historical quality value data using a communication quality prediction model to obtain spatiotemporal characteristics; and use the communication quality prediction model and spatiotemporal characteristics to predict the target quality indices of multiple sub-channels within a preset time period in the future.

[0155] The communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; a determination module is further configured to use the encoding module to encode state data and historical quality indices to obtain input features; use the spatial feature extraction module to extract the spatial positional relationships between input features to obtain spatial features; and use the temporal feature extraction module to extract the temporal dependencies between spatial features to obtain spatiotemporal features; preferably, the communication quality prediction model further includes: an attention module and an output module; using the communication quality prediction model and spatiotemporal features, the target quality index is predicted, including: using the attention module to perform attention processing on the spatiotemporal features to obtain attention features; and using the output module to perform Q-value learning on the attention features to obtain the target quality index.

[0156] The allocation module is further configured to determine the number of candidate resources in the target resource pool based on communication load data, wherein the target resource pool contains sub-communication resources corresponding to the target sub-channel; adjust the sub-communication resources corresponding to multiple sub-channels contained in the original resource pool based on the number of candidate resources to obtain the target resource pool; and allocate sub-channel resources corresponding to the target sub-channel to the vehicle from the target resource pool based on the target quality index and status data; preferably, the original resource pool is a resource pool that matches the target communication priority of the vehicle among multiple resource pools, wherein different resource pools correspond to different communication priorities.

[0157] The allocation module is further configured to determine the initial resource allocation strategy for vehicles from the target resource pool based on the target quality index; update the initial resource allocation strategy based on the status data to obtain the target resource allocation strategy; and allocate sub-channel resources corresponding to the target sub-channel to the vehicles based on the target resource allocation strategy.

[0158] The allocation module is further configured to: determine a first utility value corresponding to the initial resource allocation strategy based on state data, wherein the first utility value characterizes the benefit obtained by allocating resources to vehicles based on the initial resource allocation strategy; update the initial resource allocation strategy to obtain an updated resource allocation strategy; determine a second utility value corresponding to the updated resource allocation strategy based on state data; determine the utility difference between the first utility value and the second utility value; in response to the utility difference being greater than or equal to a preset value, use the updated resource allocation strategy as the initial resource allocation strategy, and repeatedly execute the steps of determining the first utility value corresponding to the initial resource allocation strategy based on state data, updating the initial resource allocation strategy to obtain an updated resource allocation strategy, and determining the second utility value corresponding to the updated resource allocation strategy based on state data, until the utility difference is less than the preset value; in response to the utility difference being less than the preset value, determine the updated resource allocation strategy as the target resource allocation strategy.

[0159] The acquisition module is further configured to: acquire the occupancy status of the target sub-channel; determine a backoff time in response to the occupancy status indicating that the target sub-channel is occupied, wherein the backoff time is the waiting time for re-allocating communication resources to the vehicle; and regenerate a resource allocation request in response to the arrival of the backoff time. Preferably, acquiring the occupancy status of the target sub-channel includes: acquiring the reference signal received power of the target sub-channel; determining that the occupancy status indicates that the target sub-channel is idle in response to the reference signal received power being less than a preset power; and determining that the occupancy status indicates that the target sub-channel is occupied in response to the reference signal received power being greater than or equal to the preset power.

[0160] The acquisition module is further configured to determine the signal power compensation value based on the target obstruction information; determine the transmit power of the target sub-channel based on the target quality index and the signal power compensation value; and determine the modulation scheme and coding rate of the target sub-channel based on the target quality index.

[0161] The acquisition module is further configured to acquire image data collected by a visual sensor and point cloud data collected by radar; identify occlusions around the vehicle based on the point cloud data to obtain first occlusion information; identify occlusions around the vehicle based on the image data to obtain second occlusion information; fuse the first occlusion information and the second occlusion information to obtain target occlusion information; preferably, fusing the first occlusion information and the second occlusion information to obtain target occlusion information includes: determining a first [missing information] based on position, vehicle speed, and heading angle. An initial weight value and a second initial weight value for the second occlusion information; semantic understanding is performed on the first occlusion information to obtain a first semantic understanding result, and semantic understanding is performed on the second occlusion information to obtain a second semantic understanding result; the first initial weight value is adjusted based on the first semantic understanding result to obtain a first target weight value, and the second initial weight value is adjusted based on the second semantic understanding result to obtain a second target weight value; the first occlusion information and the second occlusion information are weighted based on the first target weight value and the second target weight value to obtain the target occlusion information.

[0162] Embodiments of this disclosure also provide an on-board unit, including: a memory storing an executable program; and a processor configured to run the program, wherein the program executes the methods in the various embodiments of this disclosure when it runs.

[0163] Embodiments of this disclosure also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium resides to perform the methods of the various embodiments of this disclosure.

[0164] Embodiments of this disclosure also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of this disclosure.

[0165] Embodiments of this disclosure also provide a computer program product including a non-volatile computer-readable storage medium configured to store a computer program that, when executed by a processor, implements the methods described in various embodiments of this disclosure.

[0166] Embodiments of this disclosure also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of this disclosure.

[0167] In the above embodiments of this disclosure, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0168] In the several embodiments provided in this disclosure, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0169] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0171] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0172] The above description is only a preferred embodiment of this disclosure. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles of this disclosure, and these improvements and modifications should also be considered within the scope of protection of this disclosure.

Claims

1. A communication resource allocation method, applied to an on-board unit in a vehicle, the method comprising: In response to receiving a resource allocation request, the status data of the vehicle and the communication load data of the area where the vehicle is located are obtained. The status data includes at least one of the following: position, vehicle speed, heading angle and target obstruction information around the vehicle. The target obstruction information is information about objects around the vehicle that obstruct the transmission of communication signals. The communication load data is configured to indicate the occupancy status of multiple sub-channels. Based on the status data, a target quality index for the multiple sub-channels is determined within a future preset time period, wherein the target quality index is configured to characterize the carrying capacity of the multiple sub-channels; Based on the target quality index and the communication load data, sub-communication resources corresponding to the target sub-channel are allocated to the vehicle.

2. The method according to claim 1, wherein, The step of determining the target quality index of the multiple sub-channels within a preset time period based on the state data includes: Obtain the historical quality index of the multiple sub-channels; The spatiotemporal characteristics are obtained by fusing the state data and the historical quality index using a communication quality prediction model. Using the communication quality prediction model and the spatiotemporal characteristics, the target quality index of the multiple sub-channels within the future preset time period is predicted.

3. The method according to claim 2, wherein, The communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; the process of fusing the state data and the historical quality index using the communication quality prediction model to obtain spatiotemporal features includes: The state data and the historical quality index are encoded using the encoding module to obtain input features; The spatial feature extraction module is used to extract the spatial positional relationship between the input features to obtain spatial features; The temporal dependency relationship between the spatial features is extracted using the temporal feature extraction module to obtain the spatiotemporal features.

4. The method according to claim 2, wherein, The communication quality prediction model further includes an attention module and an output module; the step of using the communication quality prediction model and the spatiotemporal features to predict the target quality index of the multiple sub-channels within the future preset time period includes: The attention module is used to perform attention processing on the spatiotemporal features to obtain attention features; The output module is used to learn the Q-value of the attention features to obtain the target quality index.

5. The method according to claim 1, wherein, The process of allocating sub-communication resources corresponding to a target sub-channel to the vehicle based on the target quality index and the communication load data includes: Based on the communication load data, the number of candidate resources in the target resource pool is determined, wherein the target resource pool includes the sub-communication resources corresponding to the target sub-channel; Based on the number of candidate resources, the sub-communication resources corresponding to multiple sub-channels contained in the original resource pool are adjusted to obtain the target resource pool; Based on the target quality index and the status data, sub-channel resources corresponding to the target sub-channel are allocated to the vehicle from the target resource pool.

6. The method according to claim 5, wherein, The original resource pool is a resource pool among multiple resource pools that matches the target communication priority of the vehicle, wherein different resource pools correspond to different communication priorities.

7. The method according to claim 5, wherein, The step of allocating sub-channel resources corresponding to the target sub-channel from the target resource pool to the vehicle based on the target quality index and the status data includes: Based on the target quality index, the initial resource allocation strategy for the vehicle is determined from the target resource pool; Based on the state data, the initial resource allocation strategy is updated to obtain the target resource allocation strategy; Based on the target resource allocation strategy, the vehicle is allocated the sub-channel resources corresponding to the target sub-channel.

8. The method according to claim 7, wherein, The step of updating the initial resource allocation strategy based on the state data to obtain the target resource allocation strategy includes: Based on the state data, a first utility value corresponding to the initial resource allocation strategy is determined, wherein the first utility value is configured to characterize the benefit obtained by allocating resources to the vehicle based on the initial resource allocation strategy. The initial resource allocation strategy is updated to obtain the updated resource allocation strategy; Based on the status data, determine the second utility value corresponding to the updated resource allocation strategy; Determine the utility difference between the first utility value and the second utility value; In response to the utility difference being greater than or equal to a preset value, the updated resource allocation strategy is used as the initial resource allocation strategy, and the steps of determining the first utility value corresponding to the initial resource allocation strategy based on the state data, updating the initial resource allocation strategy to obtain the updated resource allocation strategy, and determining the second utility value corresponding to the updated resource allocation strategy based on the state data are repeated until the utility difference is less than the preset value. In response to the utility difference being less than the preset value, the updated resource allocation strategy is determined to be the target resource allocation strategy.

9. The method according to any one of claims 1 to 8, wherein, The method further includes: Obtain the occupancy status of the target sub-channel; In response to the occupancy status indicating that the target subchannel is occupied, a backoff time is determined, wherein the backoff time is the waiting time for the vehicle to be allocated communication resources again; In response to the arrival of the backoff time, the resource allocation request is regenerated.

10. The method according to claim 9, wherein, The step of obtaining the occupancy status of the target sub-channel includes: Obtain the reference signal received power of the target sub-channel; In response to the reference signal received power being less than a preset power, the occupancy status is determined to indicate that the target sub-channel is idle; In response to the reference signal received power being greater than or equal to the preset power, the occupancy status is determined to indicate that the target sub-channel is occupied.

11. The method according to any one of claims 1 to 8, wherein, The method further includes: Based on the target obstruction information, a signal power compensation value is determined; Based on the target quality index and the signal power compensation value, the transmit power of the target sub-channel is determined; Based on the target quality index, the modulation scheme and coding rate of the target sub-channel are determined.

12. The method according to any one of claims 1 to 8, wherein, The step of obtaining the target occlusion information includes: Acquire image data from visual sensors and point cloud data from radar; Based on the point cloud data, the obstructions around the vehicle are identified to obtain the first obstruction information; Based on the image data, the obstructions around the vehicle are identified to obtain the second obstruction information; The first occlusion information and the second occlusion information are fused to obtain the target occlusion information.

13. The method according to claim 12, wherein, The process of fusing the first occlusion information and the second occlusion information to obtain the target occlusion information includes: Based on the location, the vehicle speed, and the heading angle, a first initial weight value for the first obstruction information and a second initial weight value for the second obstruction information are determined. Semantic understanding is performed on the first occlusion information to obtain a first semantic understanding result, and semantic understanding is performed on the second occlusion information to obtain a second semantic understanding result; The first initial weight value is adjusted based on the first semantic understanding result to obtain the first target weight value, and the second initial weight value is adjusted based on the second semantic understanding result to obtain the second target weight value; Based on the first target weight value and the second target weight value, the first occlusion information and the second occlusion information are weighted to obtain the target occlusion information.

14. A vehicle-mounted unit, comprising: Memory, which stores executable programs; A processor is configured to run the program, wherein the program performs the following steps when it runs: In response to receiving a resource allocation request, the system acquires vehicle status data and communication load data of the area where the vehicle is located. The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information is information about objects around the vehicle that obstruct the transmission of communication signals. The communication load data is configured to indicate the occupancy status of multiple sub-channels. Based on the status data, a target quality index for the multiple sub-channels is determined within a future preset time period, wherein the target quality index is configured to characterize the carrying capacity of the multiple sub-channels; Based on the target quality index and the communication load data, sub-communication resources corresponding to the target sub-channel are allocated to the vehicle.

15. The vehicle-mounted unit according to claim 14, wherein, The program also performs the following steps when it runs: Obtain the historical quality index of the multiple sub-channels; The spatiotemporal characteristics are obtained by fusing the state data and the historical quality index using a communication quality prediction model. Using the communication quality prediction model and the spatiotemporal characteristics, the target quality index of the multiple sub-channels within the future preset time period is predicted.

16. The vehicle-mounted unit according to claim 15, wherein, The communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; the program also performs the following steps during runtime: The state data and the historical quality index are encoded using the encoding module to obtain input features; The spatial feature extraction module is used to extract the spatial positional relationship between the input features to obtain spatial features; The temporal dependency relationship between the spatial features is extracted using the temporal feature extraction module to obtain the spatiotemporal features.

17. The vehicle-mounted unit according to claim 15, wherein, The communication quality prediction model further includes an attention module and an output module; the program also performs the following steps during runtime: The attention module is used to perform attention processing on the spatiotemporal features to obtain attention features; The output module is used to learn the Q-value of the attention features to obtain the target quality index.

18. A computer-readable storage medium comprising a stored executable program, wherein, When the executable program is executed, it controls the device containing the computer-readable storage medium to perform the following steps: In response to receiving a resource allocation request, the system acquires vehicle status data and communication load data of the area where the vehicle is located. The status data includes at least one of the following: position, vehicle speed, heading angle, and target obstruction information around the vehicle. The target obstruction information is information about objects around the vehicle that obstruct the transmission of communication signals. The communication load data is configured to indicate the occupancy status of multiple sub-channels. Based on the status data, a target quality index for the multiple sub-channels is determined within a future preset time period, wherein the target quality index is configured to characterize the carrying capacity of the multiple sub-channels; Based on the target quality index and the communication load data, sub-communication resources corresponding to the target sub-channel are allocated to the vehicle.

19. The computer-readable storage medium according to claim 18, wherein, When the executable program is executed, it also controls the device containing the computer-readable storage medium to perform the following steps: Obtain the historical quality index of the multiple sub-channels; The spatiotemporal characteristics are obtained by fusing the state data and the historical quality index using a communication quality prediction model. Using the communication quality prediction model and the spatiotemporal characteristics, the target quality index of the multiple sub-channels within the future preset time period is predicted.

20. The computer-readable storage medium according to claim 19, wherein, The communication quality prediction model includes: an encoding module, a spatial feature extraction module, and a temporal feature extraction module; when the executable program runs, it also controls the device containing the computer-readable storage medium to perform the following steps: The state data and the historical quality index are encoded using the encoding module to obtain input features; The spatial feature extraction module is used to extract the spatial positional relationship between the input features to obtain spatial features; The temporal dependency relationship between the spatial features is extracted using the temporal feature extraction module to obtain the spatiotemporal features.