A vehicle-cloud dual-link communication guarantee and strategy back-filling system for automatic driving key services

Abstract

The application provides a vehicle cloud double-link communication guarantee and strategy back-filling system for key services of automatic driving, and relates to the technical field of automatic driving vehicle networking communication. The vehicle cloud double-link communication guarantee and strategy back-filling system for key services of automatic driving comprises a vehicle end and a cloud end. The vehicle end is configured with at least two communication links, and is provided with a service level identification module, a link quality collection module, a local strategy execution module, a double-link communication agent and an instruction idempotent execution module. The cloud end comprises a multi-link access gateway, a deduplication and reordering module, an integrity check and optimization merging module, an instruction confirmation and retransmission module, a link scoring and strategy engine. Through level identification of service data and combined with real-time link state to implement differentiated transmission control, the key service obtains higher communication guarantee in the sending stage, and avoids different important degree data from competing for limited transmission resources to cause time delay accumulation or transmission congestion.

Description

Technical Field

[0001] This invention relates to the field of vehicle-to-everything (V2X) communication technology, specifically to a vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services. Background Technology

[0002] The field of autonomous driving vehicle-to-everything (V2X) communication technology is a comprehensive interdisciplinary field integrating autonomous driving technology, mobile communication technology, V2X technology, cloud computing technology, edge computing technology, and intelligent transportation technology. It primarily studies the information interaction and collaborative control mechanisms between vehicles, between vehicles and road infrastructure, between vehicles and network platforms, and between vehicles and cloud systems. Among these, a vehicle-cloud dual-link communication assurance and strategy feedback system for critical autonomous driving services refers to a vehicle-cloud collaborative communication assurance system built to meet the data transmission needs of critical business operations in autonomous driving scenarios. Through collaborative work between the vehicle and cloud sides, it reliably transmits, monitors, assesses, and optimizes the strategy for important business data generated during autonomous driving, ensuring stable, continuous, and highly reliable data interaction for critical business operations in complex network environments. Its purpose is to improve the reliability and availability of the vehicle-cloud communication link during the operation of autonomous vehicles, reducing the impact of network fluctuations, link interruptions, or communication quality degradation on autonomous driving services. By analyzing and optimizing historical communication data and operational results, the optimized communication strategy is fed back to the vehicle side, achieving dynamic adjustment and continuous optimization of communication resources, thereby improving the overall safety, stability, and intelligence level of the autonomous driving system.

[0003] While existing autonomous driving vehicle-to-everything (V2X) communication systems possess vehicle-cloud collaboration capabilities and communication strategy optimization capabilities, in actual operation they rely more on unified management of the overall communication status. They lack sufficiently granular transmission guarantee mechanisms to address the varying importance of different services. When perception data uploading, status feedback, remote control, and general information interaction occur simultaneously, critical services are easily affected by network load fluctuations, leading to queuing and waiting. In complex communication environments such as tunnels, overpasses, urban canyons, and network switching areas, data transmission results are significantly affected by the quality of individual transmissions. Even short-term network anomalies can cause missing critical information or delayed arrival of control information. For example, when a vehicle enters the edge of base station coverage, some important data may be received intermittently, affecting the accuracy of subsequent service judgments. For downlink control processes, if retransmissions, out-of-order delivery, or duplicate delivery occur, control content arriving at different times can increase the complexity of status management, leading to deviations between execution results and expected control timing. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services. This system solves the problem that while existing autonomous driving vehicle network communication systems possess vehicle-cloud collaboration capabilities and communication strategy optimization capabilities, they rely more on unified management of the overall communication status during actual operation. They lack sufficiently detailed transmission guarantee mechanisms for different service priorities. When perception data uploads, status feedback, remote control, and general information interaction occur simultaneously, key services are easily affected by network load fluctuations, resulting in queuing and waiting phenomena.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: a vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services, including the vehicle end and the cloud end; The vehicle terminal is configured with at least two communication links and includes a service level identification module, a link quality acquisition module, a local policy execution module, a dual-link communication proxy, and an instruction idempotent execution module. The cloud includes a multi-link access gateway, a deduplication and sorting module, an integrity verification and optimal merging module, a command confirmation and retransmission module, a link scoring and policy engine, a network signal quality map and early warning audit module; The system achieves parallel transmission of key uplink data via dual links, idempotent execution of downlink vehicle control commands via dual links, and policy feedback based on cloud-received results and link anomalies through vehicle-side collaboration.

[0006] Preferably, the service level identification module is used to classify the services to be transmitted into critical services, important services and ordinary services, and trigger different uplink and downlink communication protection strategies according to the service level.

[0007] Preferably, the dual-link communication agent is used to generate at least two sets of key business data carrying a unified business sequence number, timestamp, and data digest, and send them to the cloud in parallel through different communication links to support cross-link deduplication, sorting, and optimal merging in the cloud.

[0008] Preferably, the instruction idempotent execution module executes the downlink vehicle control instruction that arrives repeatedly only once, based on the instruction sequence number, idempotency token, and execution status table, and sends the repeated arrival information back to the cloud for auditing and network signal quality map updates.

[0009] Preferably, the network signal quality map is used to record the network quality and uplink reception results, downlink command confirmation rate, number of consecutive anomalies and duplicate arrival records for geographical areas, roads, time periods, operators, link types and service types, and supports the dynamic generation of end-side communication policies by the cloud policy engine.

[0010] Preferably, the link scoring and strategy engine dynamically generates communication strategies, including receive window length, link priority, retransmission period, degradation strategy and normal service rate limit threshold, based on service level, real-time link quality parameters, historical network signal quality map and uplink / downlink communication results, and feeds them back to the vehicle end.

[0011] Preferably, the early warning audit module is used to monitor link anomalies, data loss, retransmission status, selection and reception results, and instruction confirmation status, and triggers an early warning when an abnormal event occurs, while using the audit data to optimize subsequent strategy feedback.

[0012] Preferably, when the vehicle-side local policy execution module experiences a short-term disconnection or a decline in link quality, it independently executes critical business dual-link transmission, important business primary / backup hot-switching, and ordinary business speed limiting / pause based on cached policies to ensure the continuity of critical autonomous driving services.

[0013] Preferably, the system supports multiple link combinations, including but not limited to two physical SIM cards, a physical SIM and an eSIM, two eSIM profiles, a cellular network and a private network, a cellular network and a satellite network, or a cellular network and a fleet relay network.

[0014] Preferably, the system forms a closed-loop communication guarantee mechanism: dual-link transmission and idempotent execution at the vehicle end, deduplication sorting and optimal merging at the cloud, and network signal quality map updates and policy feedback are coupled to achieve the continuity, integrity, verifiability and auditability of uplink key data and downlink vehicle control commands in scenarios with weak coverage, link anomalies or multi-operator switching.

[0015] This invention provides a vehicle-cloud dual-link communication guarantee and policy feedback system for key autonomous driving operations. It has the following beneficial effects: This invention implements differentiated transmission control by classifying business data and combining it with real-time link status, ensuring higher communication guarantees for critical services during the transmission phase. This avoids data of varying importance competing for limited transmission resources, preventing latency accumulation or transmission congestion. By using two communication links to simultaneously carry the same critical data, a parallel transmission path is formed. Even if one link experiences attenuation, jitter, or momentary interruption, the other link can still maintain data delivery capability, thus improving the data reachability probability in complex road environments. The cloud side performs sequence sorting, integrity verification, and quality comparison based on the arrival results from different links, retaining only valid and higher-quality data content. This reduces the consumption of computing and storage resources by duplicate data and improves the accuracy of subsequent business processing results. For the downlink control process, a processing mechanism combining dual-path synchronous transmission and unique identifier verification is adopted to ensure that the same control... Even if the content arrives repeatedly via different links, it can still maintain a single effective execution state, avoiding control conflicts, action overlap, or state abnormalities caused by repeated triggering. By combining link operation results, data arrival status, and abnormal events, a continuous evaluation basis is formed, dynamically generating communication strategies that are more in line with the actual operating environment. The optimization results are fed back to the vehicle side to participate in subsequent transmission decisions, enabling communication behavior to be continuously adjusted according to changes in road environment, network coverage, and service load. At the same time, by accumulating network performance information in different regions and time periods over a long period of time, a network quality distribution recognition capability with spatial correlation characteristics is formed. Resource scheduling and link preparation can be implemented in advance before the vehicle approaches a communication risk area, reducing the impact of sudden communication anomalies on the continuity of critical autonomous driving services, and promoting a synergistic improvement effect in data transmission assurance, control execution reliability, and communication resource utilization efficiency. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall system architecture of the present invention; Figure 2 This is a structural block diagram of the vehicle-side functional modules of the present invention; Figure 3 This is a block diagram of the cloud-based functional module structure of the present invention; Figure 4 This is a schematic diagram of the dual-link communication guarantee mechanism of the present invention; Figure 5 This is a schematic diagram of the downlink control command idempotent execution mechanism of the present invention; Figure 6 This is a schematic diagram of the network signal quality map and strategy feedback closed loop of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example: like Figure 1-6 As shown, this embodiment of the invention provides a vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services, including a vehicle end and a cloud end; The vehicle is equipped with at least two communication links and includes a service level identification module, a link quality acquisition module, a local policy execution module, a dual-link communication proxy, and an instruction idempotent execution module. First, a multi-channel access configuration table is established, and the connection parameters corresponding to different operating networks, dedicated networks, or vehicle communication modules are written into the configuration table. Then, service information is processed by level identification, with the service level value denoted as G. The value range can be divided into basic service range, medium-level service range, and critical service range, with different transmission rules corresponding to different ranges. During vehicle operation, the status parameters of each communication channel are periodically collected, including latency T, packet loss rate P, signal strength S, and bandwidth utilization B, and then analyzed according to the scoring formula. The calculation is performed, where Q is the link quality score, and a, b, c, and d are weight coefficients. The weights are set based on historical operation records and service types, and the sum of each weight is a fixed value. When key service data arrives, a dual-channel transmission plan is generated based on the score result, and the service identifier code M is recorded. Subsequently, the received instructions are numbered and managed through local execution logic, and the number value is associated with the time tag for storage. When the same numbered instruction arrives again, the execution module verifies it according to the number consistency rule. The verification items include number matching degree, time window matching degree, and content hash matching degree. If all three matching results fall within the preset range, the single execution record is maintained. In the autonomous driving scenario, the path adjustment information and vehicle status information can be synchronously written into the corresponding buffers of the two communication paths, and the transmission scheduling is completed in combination with the above score calculation process, finally forming a unified communication record result that can be processed later.

[0019] The cloud-based system includes a multi-link access gateway, a deduplication and sorting module, an integrity verification and optimization merging module, a command confirmation and retransmission module, a link scoring and policy engine, a network signal quality map, and an early warning audit module. First, a multi-entry reception table is established to uniformly register data from different network entry points, and a reception sequence number R is generated for each data entry. Then, sorting is performed based on time stamps and service identifiers, with time differences calculated during the sorting process. ,in Indicates the first reception time. This indicates the second reception time. When the time difference falls within the preset synchronization interval, the process enters the same group processing flow, followed by integrity verification. The integrity index C can be represented as... ,in Indicates the number of valid fields. This indicates the total number of fields to be received. When the integrity index reaches a preset level range, the merging process begins. During merging, multiple source data are compared field by field, including location, status, and control fields. For fields with discrepancies, the best option is selected based on the source link score L. The link score can be calculated from historical success rate, average latency, and number of anomalies, and can be expressed as follows: ,in The configuration coefficient is determined based on historical statistical samples. After receiving the control information, the acknowledgment matching process is also required based on the confirmation number. When the acknowledgment does not enter the specified range, a resend record is generated. In the vehicle remote control scenario, multi-path data of the same control request can be merged and verified, and the records are integrated based on the scoring results, ultimately forming a unified processing result.

[0020] The system enables parallel transmission of key uplink data via dual links, idempotent execution of downlink vehicle control commands via dual links, and policy feedback based on cloud-received results and link anomalies through collaboration between the vehicle and the cloud.

[0021] First, a feedback association table is established between the vehicle side and the center side, mapping and associating service numbers, link numbers, and anomaly numbers. Then, the status data of each communication path is statistically analyzed within a continuous operating cycle. Status indicators include average latency (D), anomaly frequency (E), reception integrity (F), and signal fluctuation amplitude (V). The anomaly frequency can be obtained using E = n / T, where n represents the number of anomaly events and T represents the statistical period length. Finally, based on the comprehensive evaluation value... An assessment was conducted, in which For the normalized result, To evaluate the weights, the weights are determined based on historical operating samples and service level configurations. When the evaluation value enters the abnormal range, a feedback record is generated and written to the policy library. The policy library establishes corresponding rule sets according to different regions, time periods, and network types. In the autonomous driving operation scenario, when a certain path experiences continuous latency fluctuations and enters the abnormal level range, the corresponding region is marked and written to the network quality record table. Then, the sending order parameters, retransmission number parameters, and switching condition parameters are revised based on the historical statistical results of the same region. Each parameter is set through the historical sample mean, fluctuation range, and operating frequency. Subsequently, the revision results are sent to the terminal-side rule table to complete the update, ultimately forming a continuously updated feedback record result.

[0022] The service level identification module is used to classify the services to be transmitted into critical services, important services, and ordinary services, and trigger different uplink and downlink communication protection strategies according to the service level.

[0023] For vehicle-to-everything (V2X) service classification units, parameters such as service source identifier, message type identifier, latency tolerance level, data update frequency, and associated control objects are first extracted from the data stream to be processed. Then, each item is quantitatively processed according to preset evaluation items, where the latency index can be expressed as a latency score. The scope of business impact can be indicated by the impact score. This indicates that the degree of resource utilization can be expressed using a resource score. This means that each parameter is divided into low, medium, and high intervals based on historical operating records. A higher evaluation coefficient is assigned to a parameter in the high interval, and a lower evaluation coefficient is assigned to a parameter in the low interval. Then, a comprehensive evaluation value is used. Perform calculations, where This indicates the overall evaluation result. Indicates the weight of the corresponding evaluation item. This represents the score for the corresponding evaluation item, with the summation symbol. This indicates the cumulative calculation process for all evaluation items. Each weight is configured according to business management rules. For example, data related to vehicle operation status corresponds to a larger weight range, while periodic status reporting data corresponds to a smaller weight range. After the calculation is completed, the obtained evaluation results are matched with multiple level ranges. When the evaluation result falls into the highest level range, it is classified into the highest priority category. When the evaluation result falls into the middle level range, it is classified into the middle priority category. When the evaluation result falls into the basic level range, it is classified into the basic priority category. Then, the corresponding communication processing rules are associated based on the matching results, and a set of communication strategies corresponding to the business level is formed.

[0024] The dual-link communication agent is used to generate at least two copies of critical business data carrying a unified business sequence number, timestamp, and data summary, and send them to the cloud in parallel through different communication links to support cross-link deduplication, sorting, and optimal merging in the cloud.

[0025] For the dual-channel data forwarding unit, the service identifier information, generation time information, and content feature information are first extracted from the data entering the transmission queue. Uniform encoding is then performed on each item, and multiple identical data copies are generated. Each copy contains the same service identifier field, time association field, and digest verification field. The digest verification field can be calculated based on a digest formula. get, Indicates the summary results. Represents the original data content. The process involves generating a summary, followed by reading the status parameters of different communication channels, including link latency, packet loss rate, bandwidth utilization rate, and historical stability indicators. Each status parameter is divided into low, medium, and high ranges. When the link status parameter falls into the excellent range, it enters the priority transmission set; when the link status parameter falls into the normal range, it enters the backup transmission set. Subsequently, multiple data copies are written to different link transmission buffers. During vehicle operation, for example, when the same control message is synchronously forwarded via cellular and satellite communication links, each link independently records the transmission time and feedback status information, forming a link record table. Later, these records are associated and archived based on the same service identifier field, forming a unified data arrival record set.

[0026] The instruction idempotent execution module executes downlink vehicle control instructions that arrive repeatedly only once, based on the instruction sequence number, idempotency token, and execution status table, and sends the repeated arrival information back to the cloud for auditing and network signal quality map updates.

[0027] The instruction processing unit, upon receiving a control message, first parses the instruction identifier field, unique token field, and execution record field. The parsing results are then correlated with the local record database. Search criteria include instruction identifier matching degree, token consistency, and historical processing status. When a correlated record is found, the execution status information is further read and compared based on status markers. For example, status markers can be divided into unprocessed, processing, and processed intervals. If the corresponding record is in the processed interval, the execution process is not repeated; instead, a duplicate reception record is generated. If the corresponding record is in the unprocessed interval, it is written to the execution task queue, and then the status record table is updated. The status values ​​in the status record table can be... This indicates that the number of executions can be expressed as follows: This indicates that the record update time can be recorded using... This indicates that the state update process can be represented as follows: , This indicates the updated status result. This indicates the status update rule. In vehicle control scenarios, when the same control message arrives successively via different communication paths, the system reads its identifier field and token field to perform consistency verification, writes the number of repeated arrivals into the log entry, generates associated record information and writes it into the return queue, and finally forms a record result containing repeated reception flags, reception time information and processing status information.

[0028] The network signal quality map is used to record network quality and uplink reception results, downlink command confirmation rate, number of consecutive anomalies and duplicate arrival records for geographical areas, roads, time periods, operators, link types and service types. It supports the dynamic generation of end-side communication policies by the cloud policy engine.

[0029] The data recording base unit, constructed based on communication scenario information, divides the vehicle operating area into multiple grid units and establishes corresponding record items according to road type, time period type, access network type, and service message type. During the collection period, communication records uploaded by multiple vehicles are continuously aggregated. For each record, signal strength, latency, packet loss, and message arrival indicators are extracted. Signal strength is denoted as R, latency as T, and packet loss as L. The overall quality value Q is calculated, where α, β, and γ are weighting coefficients. Each weight is allocated according to historical statistical results in three intervals: large, medium, and small. When the overall quality value is in the preset high-quality interval, it is marked as a priority recording level; when the overall quality value is in the preset medium interval, it is marked as a normal recording level; and when the overall quality value is in the preset low-quality interval, it is marked as a restricted recording level. At the same time, the ratio of the number of received messages to the number of sent messages within a certain period is calculated to obtain the receiving ratio data. The ratio of the number of confirmed messages to the number of sent messages is calculated to obtain the confirmation ratio data. The number of consecutively abnormal record segments is calculated to form continuous abnormal statistics data. The number of times the same identifier message appears repeatedly is calculated to form repeated arrival statistics data. When a vehicle passes through a bridge section, tunnel section, or urban section, a corresponding area record set is formed and written into the corresponding storage table in chronological order. Finally, a data set result is formed for subsequent rule generation.

[0030] The link scoring and policy engine dynamically generates communication policies, including receive window length, link priority, retransmission period, degradation policy and normal service rate limit threshold, based on service level, real-time link quality parameters, historical network signal quality map and uplink / downlink communication results, and feeds them back to the vehicle end.

[0031] A communication decision-making basis table is established based on service priority information. Real-time link status data, historical area record data, and uplink / downlink communication record data are correlated. First, the signal quality value, latency value, packet loss rate, and acknowledgment ratio of the current link are extracted and normalized to form standard parameters, denoted as S1, S2, S3, and S4. Then, the following parameters are used... Calculate the link score P, where The scoring weight coefficients are set in intervals based on the impact of each indicator on the communication process in historical business records. A higher link level is assigned when the score value is in a higher interval, an intermediate link level when the score value is in a middle interval, and a lower link level when the score value is in a lower interval. When the vehicle enters a network fluctuation area, the corresponding parameter configuration table is read according to the scoring result. The reception duration parameter, transmission order parameter, retransmission interval parameter, and service flow control parameter are extracted from the parameter configuration table. A secondary correction is made based on the recent period's confirmation ratio change. For example, when the confirmation ratio is continuously at the lower edge of the preset fluctuation range, the retransmission interval parameter is adjusted to the corresponding correction range. When the confirmation ratio is continuously at the upper edge of the preset fluctuation range, the reception duration parameter is adjusted to the corresponding correction range. Then, the corresponding communication rule record is generated and sent to the vehicle execution unit to form the corresponding strategy configuration result.

[0032] The early warning audit module is used to monitor link anomalies, data loss, retransmission status, selection and reception results, and instruction confirmation status. It triggers early warnings when abnormal events occur and uses the audit data to optimize subsequent strategy feedback.

[0033] A full-process audit record table is established based on the basic unit of operation monitoring. This table continuously collects records of abnormal events, data transmission, and feedback during the communication process. First, it counts the number of sent messages (N1), received messages (N2), and acknowledgment messages (N3) according to a preset time window. The system acquires the loss ratio data and the confirmation ratio data using C = N3 / N1. The results are then written to the audit log table. Subsequently, records from multiple consecutive time windows are compared. An abnormal event record is generated when the anomaly count within a consecutive window exceeds the corresponding interval threshold. A retransmission event record is generated when the number of repeated transmissions enters a preset high-frequency interval. Selective reception event records are generated when the reception of the same service data differs across different links. The thresholds are set based on the average and fluctuation values ​​within historical statistical periods. For example, a threshold value is used... The calculation is performed, where M represents the historical average value, F represents the historical fluctuation value, and K represents the correction coefficient. Different value ranges are set according to different business levels. After the event identification is completed, the corresponding audit entries are generated. The audit entries formed in multiple periods are summarized and statistically analyzed to obtain abnormal frequency data, retransmission frequency data, and confirmation status change data. Then, the data input results required for subsequent rule adjustments are formed based on the statistical results.

[0034] When the vehicle-side local policy execution module experiences a short-term disconnection or a deterioration in link quality, it independently executes cached policies to send critical services via dual links, perform hot-switching of important services between primary and backup, and limit / suspend ordinary services, ensuring the continuity of critical autonomous driving services.

[0035] The local control unit, based on the pre-stored rule set on the vehicle side, first extracts the service identifier, service level identifier, link status record item, and historical transmission record item from the receive buffer. The service level identifier is divided into critical, important, and normal levels. The link status record item includes latency interval values, packet loss rate interval values, and signal strength interval values. Then, each record item is correlated, and a link quality evaluation parameter Q is constructed according to a preset evaluation formula. Q is obtained by weighting a latency weight item, a packet loss rate weight item, and a signal strength weight item, where the weight items are configured based on the proportion of historical statistical samples. When the link quality evaluation parameter falls within the normal range, the existing transmission mode is maintained. When the link quality evaluation parameter falls within the decreasing range for multiple consecutive sampling periods, the buffer rule invocation process is triggered. During execution, the critical-level service pairs are read first. The system encapsulates the same business data into multiple independent messages and writes them into different sending queues according to the corresponding sending rules. Then, it reads the switching rules corresponding to the important business and compares the evaluation parameters of the current bearer channel and the backup bearer channel item by item. When the evaluation result of the backup channel is continuously in the preferred range, the business mapping migration is performed. For ordinary business, the transmission frequency parameter is adjusted according to the transmission frequency parameter, which is obtained by the ratio of the available bandwidth range of the current link to the bandwidth range occupied by the business. For example, when an autonomous vehicle uploads environmental perception records, the critical data corresponds to multiple sending queues and is output synchronously. The important data is migrated to the backup channel according to the link evaluation result. The ordinary data enters the delayed transmission queue according to the adjusted frequency parameter and continues to execute the corresponding processing flow according to the caching rules until the corresponding business processing result is obtained.

[0036] The system supports various link combinations, including but not limited to two physical SIM cards, a physical SIM and an eSIM, two eSIM profiles, a cellular network and a private network, a cellular network and a satellite network, or a cellular network and a fleet relay network.

[0037] Based on a multi-communication medium collaborative configuration framework, the identity parameters, access parameters, and coverage parameters of different communication resources are first extracted and a unified resource description table is established. Each type of communication resource records its access stability range, transmission capacity range, and coverage continuity range. Then, a resource matching process is executed. First, the set of business requirement parameters is read, including data scale level, latency requirement level, and continuous connection requirement level. Then, resource association calculations are performed based on the matching coefficient M, which is formed by weighting the correspondence between requirement parameters and resource parameters. The weight configuration is determined based on the call frequency range in historical operation records. When a resource is matched… When a coefficient falls into the preferred range, it is included in the candidate resource set. Then, a combination analysis is performed on the candidate resource set. During the analysis, the degree of coverage overlap, handover continuity, and transmission capacity supplementation between different resources are compared item by item. For example, when a vehicle is in a cross-regional operation scenario, if the coverage continuity of the first communication resource is in the medium range and the coverage continuity of the second communication resource is in the better range, a mapping relationship between the main resource and the auxiliary resource is established according to the combination rules. When the resource status record changes, the matching coefficient is recalculated and the mapping relationship is updated. The resource screening, combination association, and mapping update processing are continuously completed in multiple sampling periods, and the corresponding resource configuration results are formed.

[0038] The system forms a closed-loop communication guarantee mechanism: dual-link transmission and idempotent execution at the vehicle end, deduplication and sorting and optimal merging at the cloud, and network signal quality map updates and policy feedback are coupled to ensure the continuity, integrity, verifiability and auditability of uplink key data and downlink vehicle control commands in scenarios with weak coverage, link anomalies or multi-operator switching.

[0039] Based on the collaborative processing flow among the vehicle-side processing unit, cloud-side processing unit, and network-side data unit, the uploaded data identifier, received time identifier, service sequence identifier, and link source identifier are first extracted and a unified record table is established. During execution, the vehicle-side processing unit first generates multiple source records for the same service. Then, the cloud-side processing unit reads the service sequence identifier from each source record and performs duplicate verification. The duplicate verification is completed by comparing the same service identifier with the sequence identifier. When multiple records meet the consistency condition, the preferred record is retained and the source information is registered. Subsequently, the order reconstruction process is performed according to the time sorting parameter, which is determined by the record. The generation order is associated with the receiving order, and the reconstruction result is written into the service processing queue. The network-side data unit synchronously summarizes the operation records of each area and extracts the signal quality parameter set, which includes coverage level parameters, latency level parameters, and stability level parameters. Regional quality records are formed according to the statistical period. When the level of a regional quality record changes, a new set of rule parameters is generated and sent to the vehicle-side cache area. For example, when a vehicle passes through an area with changed coverage, records from different sources first complete consistency verification, then complete sorting reconstruction and rule update processing, and then re-execute the service processing flow according to the updated rule parameters to form the corresponding data record results.

[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A vehicle-cloud dual-link communication assurance and strategy feedback system for key autonomous driving operations, characterized in that, Including both vehicle-side and cloud-side components; The vehicle terminal is configured with at least two communication links and includes a service level identification module, a link quality acquisition module, a local policy execution module, a dual-link communication proxy, and an instruction idempotent execution module. The cloud includes a multi-link access gateway, a deduplication and sorting module, an integrity verification and optimal merging module, a command confirmation and retransmission module, a link scoring and policy engine, a network signal quality map and early warning audit module; The system achieves parallel transmission of key uplink data via dual links, idempotent execution of downlink vehicle control commands via dual links, and policy feedback based on cloud-received results and link anomalies through vehicle-side collaboration.

2. The vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The service level identification module is used to classify the services to be transmitted into critical services, important services and ordinary services, and trigger different uplink and downlink communication protection strategies according to the service level.

3. The vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The dual-link communication proxy is used to generate at least two sets of key business data carrying a unified business sequence number, timestamp, and data digest, and send them to the cloud in parallel through different communication links to support cross-link deduplication, sorting, and optimal merging in the cloud.

4. The vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The instruction idempotent execution module executes downlink vehicle control instructions that arrive repeatedly only once, based on the instruction sequence number, idempotency token, and execution status table, and sends the repeated arrival information back to the cloud for auditing and network signal quality map updates.

5. A vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The network signal quality map is used to record network quality and uplink reception results, downlink command confirmation rate, number of consecutive anomalies and duplicate arrival records for geographical areas, roads, time periods, operators, link types and service types. It supports the dynamic generation of end-side communication policies by the cloud policy engine.

6. A vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The link scoring and strategy engine dynamically generates communication strategies, including receive window length, link priority, retransmission period, degradation strategy and normal service rate limit threshold, based on service level, real-time link quality parameters, historical network signal quality map and uplink / downlink communication results, and feeds them back to the vehicle end.

7. A vehicle-cloud dual-link communication guarantee and strategy feedback system for autonomous driving key business as described in claim 1, characterized in that: The early warning audit module is used to monitor link anomalies, data loss, retransmission status, selection results and instruction confirmation status, and triggers early warnings when abnormal events occur. At the same time, the audit data is used to optimize subsequent strategy feedback.

8. A vehicle-cloud dual-link communication guarantee and strategy feedback system for autonomous driving key business as described in claim 1, characterized in that: When the vehicle-side local policy execution module experiences a short-term disconnection or a deterioration in link quality, it independently executes critical business dual-link transmission, important business primary / backup hot-switching, and ordinary business speed limiting / pause based on cached policies to ensure the continuity of critical autonomous driving services.

9. A vehicle-cloud dual-link communication guarantee and strategy feedback system for key autonomous driving services as described in claim 1, characterized in that: The system supports multiple link combinations, including but not limited to two physical SIM cards, a physical SIM and an eSIM, two eSIM profiles, a cellular network and a private network, a cellular network and a satellite network, or a cellular network and a fleet relay network.

10. A vehicle-cloud dual-link communication guarantee and strategy feedback system for autonomous driving key business as described in claim 1, characterized in that: The system forms a closed-loop communication guarantee mechanism: dual-link transmission and idempotent execution at the vehicle end, deduplication and sorting and optimal merging at the cloud, network signal quality map updates and policy feedback are coupled to achieve continuity, integrity, verifiability and auditability of uplink key data and downlink vehicle control commands in scenarios with weak coverage, link anomalies or multi-operator switching.

Images