Intelligent Arbitration Method and System Based on Multi-Protocol Vehicle Network

Abstract

This invention discloses an intelligent arbitration method and system based on a multi-protocol vehicular network. The method includes: parsing messages from different communication protocols and converting them into a three-dimensional feature vector containing data type, real-time characteristics, and data size; collecting load characteristic parameters of the vehicular network's operating status to form a load feature vector; predicting network load trends for a preset time period based on the three-dimensional feature vector and the load feature vector using a GRU model; triggering a decision based on scheduling triggering parameters and a preset dynamic threshold; and adjusting arbitration when the triggering conditions are met. This invention enables predictive scheduling optimization based on protocol type, message attributes, and real-time network load status, thereby improving communication resource utilization efficiency and ensuring the real-time transmission quality of critical messages.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of intelligent driving technology, specifically to an intelligent arbitration method and system, electronic device, and storage medium based on a multi-protocol vehicle network. Background Technology

[0002] With the development of autonomous driving, the in-vehicle electronic and electrical architecture is gradually evolving from a domain-centralized to a vehicle-centric computing architecture. The vehicle's internal communication network has expanded from the traditional CAN bus to a multi-protocol converged network, such as CAN FD, CAN XL, and high-speed Ethernet. However, the existing CAN XL (20Mbit / s, 2048-byte data field), CAN FD (5Mbit / s in high-speed mode), and Ethernet (hundreds of megabits / gigabits of bandwidth) protocols have significant differences in characteristics. Traditional arbitration mechanisms use fixed rules (such as CAN's bit-by-bit comparison requiring 11-29 bits of checksum), and when the load rate reaches 50%, the arbitration time can be as high as 7.5 × 10⁻⁻⁶. 4 The conflict intensifies when multiple protocols run concurrently, making it unsuitable for the high real-time requirements of autonomous driving.

[0003] Moreover, in autonomous driving, the priority requirements for data such as LiDAR point clouds (Ethernet transmission), braking signals (CAN FD transmission), and diagnostic data (CAN XL transmission) change dynamically, but existing systems use preset priorities, which can easily lead to delays in safety-related data under high load (such as braking signals being blocked by low-priority diagnostic data). Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the purpose of this invention is to provide an intelligent arbitration method, system, and electronic device based on a multi-protocol vehicle network, which realizes predictive allocation of communication resources, avoids the delay expansion problem caused by traditional passive scheduling, and ensures deterministic communication delay guarantee for safety-related data such as critical control messages under complex operating conditions.

[0005] To address the aforementioned problems, the first aspect of this invention discloses an intelligent arbitration method based on a multi-protocol vehicular network, comprising the following steps:

[0006] The system parses messages from different communication protocols and converts the parsing results into a three-dimensional feature vector containing data type, real-time characteristics, and data size.

[0007] The load characteristic parameters of the vehicle network operation status are collected to form a load characteristic vector. The load characteristic parameters include bandwidth utilization, packet collision frequency, node response time and vehicle operation status parameters.

[0008] Based on the three-dimensional feature vector and the load feature vector, the network load trend for a preset time period is predicted using the GRU model; the load prediction result is the output of the GRU model, including predicted bandwidth utilization, packet collision probability and communication latency trend.

[0009] Based on the prediction results, the scheduling trigger determination parameters are calculated, and the trigger determination is made according to the scheduling trigger determination parameters and the preset dynamic threshold.

[0010] When the triggering conditions are met, arbitration adjustment is performed; the arbitration adjustment includes: generating a three-dimensional evaluation matrix containing security weight, real-time coefficient and load redundancy based on the security level, latency sensitivity and communication resource occupation requirements of the message to be transmitted; obtaining an arbitration priority score based on the three-dimensional evaluation matrix; and executing a predictive time slot reservation mechanism based on the arbitration priority score, and dynamically rearranging the priority of the message to be transmitted and scheduling communication resources.

[0011] Optionally, the parsing of messages from different communication protocols includes:

[0012] Acquire messages from different communication protocols and parse the identifier of CAN XL protocol data, the variable data field length of CAN FD protocol data, and the tag parameters of Ethernet protocol data respectively.

[0013] Optionally, the step of determining the trigger based on the scheduling trigger determination parameters and a preset dynamic threshold includes:

[0014] Based on the dynamic threshold setting, three levels of load threshold conditions are set. When the scheduling trigger determination parameters meet the first level load threshold condition, regular arbitration is performed; when the scheduling trigger determination parameters meet the second level load threshold condition and the third level load threshold condition, the trigger condition is determined to be met.

[0015] Optionally, the dynamic prioritization and communication resource scheduling of the messages to be transmitted includes dynamic channel switching scheduling and protocol bandwidth occupancy adjustment: when the scheduling trigger determination parameter meets the second-level load threshold condition, the CAN XL protocol message arbitration priority strategy is adopted; when the scheduling trigger determination parameter meets the third-level load threshold condition, Ethernet bandwidth reservation is initiated, and non-real-time data is transmitted through the CAN FD protocol.

[0016] Optionally, the conflict scheduling strategy of the communication resource scheduling adopts a time slot division mechanism to allocate independent time resource segments to messages of different protocols or different levels in order to reduce arbitration conflicts.

[0017] The implementation of a predictive time slot reservation mechanism, and the dynamic prioritization and communication resource scheduling of the packets to be transmitted include: reserving independent communication time slots for critical packets before the scheduling trigger judgment parameters meet the second-level load threshold conditions;

[0018] When an automatic emergency braking trigger signal is detected, the communication resource scheduling prioritizes the execution of a predictive time slot reservation strategy to map the braking control message to the Ethernet TSN channel.

[0019] Optionally, the method employs arbitration priority score inheritance to maintain the original arbitration priority score when the message is migrated across protocols; and when a V2X environment event warning message is received, the cross-protocol priority inheritance mechanism is executed to migrate the message to a higher bandwidth communication protocol channel.

[0020] Optionally, the preset time period is 100ms, and the communication resource scheduling ensures that the transmission delay of critical safety messages does not exceed 20ms; when the vehicle is in high-speed automatic lane changing mode, its scheduling level is raised by at least one level according to the real-time coefficient of the trajectory planning message, the raised scheduling level is less than or equal to the highest level, and the protocol migration strategy is triggered to ensure that the transmission delay of the trajectory planning message is not greater than 10ms.

[0021] A second aspect of this invention discloses an intelligent arbitration system based on a multi-protocol vehicle network, comprising:

[0022] The message parsing module is used to parse messages from different communication protocols and convert the parsing results into a three-dimensional feature vector containing data type, real-time characteristics and data size.

[0023] The load acquisition module is used to collect load characteristic parameters of the vehicle network operation status and form a load characteristic vector. The load characteristic parameters include bandwidth utilization, packet collision frequency, node response time and vehicle operation status parameters.

[0024] The prediction calculation module is used to predict the network load trend for a preset time period based on the three-dimensional feature vector and the load feature vector using the GRU model; the load prediction result is the output of the GRU model, including the predicted bandwidth utilization, packet collision probability and communication latency trend.

[0025] The trigger judgment module is used to calculate the scheduling trigger judgment parameters based on the prediction results, and to make a trigger judgment based on the scheduling trigger judgment parameters and a preset dynamic threshold.

[0026] The arbitration adjustment module is used to perform arbitration adjustment when triggering conditions are met. The arbitration adjustment includes: generating a three-dimensional evaluation matrix containing security weight, real-time coefficient and load redundancy based on the security level, latency sensitivity and communication resource occupation requirements of the message to be transmitted; obtaining an arbitration priority score based on the three-dimensional evaluation matrix; and executing a predictive time slot reservation mechanism based on the arbitration priority score, and dynamically rearranging the priority of the message to be transmitted and scheduling communication resources.

[0027] A third aspect of the present invention discloses an electronic device, comprising: a memory storing executable program code; a processor coupled to the memory; the processor calling the executable program code stored in the memory to execute the intelligent arbitration method based on a multi-protocol vehicle network disclosed in the first aspect of the present invention.

[0028] A fourth aspect of the present invention discloses a computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the intelligent arbitration method based on a multi-protocol vehicle network disclosed in the first aspect of the present invention.

[0029] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows:

[0030] This invention achieves consistency in message classification and control strategies across protocols by uniformly parsing multi-protocol data formats and generating quantifiable three-dimensional feature vectors. It employs the GRU model to perform trend analysis on the time-varying load of the vehicular network and outputs a dynamic trigger scheduling strategy based on the prediction results, realizing predictive allocation of communication resources and avoiding the latency expansion problem caused by traditional passive scheduling. By constructing a three-dimensional evaluation matrix containing security weights, real-time coefficients, and load redundancy, it calculates arbitration priority scores. Based on the arbitration priority scores, it executes a predictive time slot reservation mechanism and dynamically rearranges the priorities of messages to be transmitted and schedules communication resources, ensuring deterministic communication latency guarantees for safety-related data such as critical control messages even under complex operating conditions.

[0031] Furthermore, this invention enables millisecond-level response scheduling in scenarios such as high-speed lane changes, emergency braking, or receiving V2X emergency warnings, thereby significantly improving the reliability of the vehicle network, real-time transmission capabilities, and the safety of autonomous driving operation. It avoids the problem of existing static scheduling schemes being unable to dynamically adjust according to real-time autonomous driving scenarios, which may lead to increased delays and response lags in critical control messages. Attached Figure Description

[0032] Figure 1 This is a flowchart illustrating an intelligent arbitration method based on a multi-protocol vehicle network according to an embodiment of the present invention.

[0033] Figure 2 This is a schematic diagram of the structure of an intelligent arbitration system based on a multi-protocol vehicle network provided in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. Detailed Implementation

[0035] This specific embodiment is merely an explanation of the embodiments of the present invention and is not intended to limit the embodiments of the present invention. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but as long as they are within the scope of the claims of the embodiments of the present invention, they are protected by patent law.

[0036] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 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 protection scope of the embodiments of the present invention.

[0037] The term "comprising" and any variations thereof in the specification and claims of this application are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product or device.

[0038] In embodiments of the present invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0039] Example 1

[0040] Please refer to Figure 1-3 As shown, an intelligent arbitration method based on a multi-protocol vehicular network is proposed. Figure 1 As shown, it includes the following steps:

[0041] Step S110: Parse messages from different communication protocols and convert the parsing results into a three-dimensional feature vector containing data type, real-time characteristics and data size;

[0042] Specifically, including:

[0043] Acquire messages from different communication protocols and parse the identifier of CAN XL protocol data, the variable data field length of CAN FD protocol data, and the tag parameters of Ethernet protocol data respectively.

[0044] In this embodiment, the core differences between different protocols, such as the 29-bit ID of CAN XL and the VLAN tag of Ethernet, are standardized through "three-dimensional feature vectors," namely data type, real-time performance, and data volume. The arbitration decision engine no longer needs to adapt to the underlying format of each protocol. As can be seen from data processing theory, a unified format can reduce the intermediate links of "protocol conversion-format adaptation," which can theoretically compress the decision response time to within 10ms. Combined with the low collision latency of TDM (Time Division Multiplexing), the total latency can be ≤20ms, which meets the theoretical requirements of ASIL-D level for "real-time performance."

[0045] For example, in an Automatic Emergency Braking (AEB) triggering scenario, the parsing of braking control messages includes structured parsing and semantic mapping: Structured parsing obtains the message identifier, data length, and time sensitivity parameters; semantic mapping rules identify the message type and map braking control messages to preset safety messages; the system determines the target transmission delay to be no greater than 10ms based on the real-time control requirements of safety messages. That is, it is identified as "safety type," with a target transmission delay ≤10ms and a message length of 64 bytes, obtaining parameters such as message type, time sensitivity, and data size. The original braking signal message attributes are "safety type - 10ms level - 64 bytes," where type: safety braking signal, time sensitivity: 10ms control cycle, data length: 64 bytes. Then, it is converted into a three-dimensional feature vector [1.0, 0.95, 0.031] containing data category, real-time characteristics, and data size through a preset mapping table.

[0046] In this step, the transformation of the three-dimensional feature vector can be generated based on a preset mapping table, which includes: message category encoding value, real-time level value, and data size normalization value. Specifically, it can be generated through a segmented interval mapping function. Specifically, the real-time level value is calculated using a segmented and hierarchical time sensitivity model.

[0047] In this step, heterogeneous messages from different protocols are transformed into basic input parameters that can be used for GRU model prediction analysis by structured parsing and semantic mapping of multi-protocol vehicle communication frames and by using a unified feature vector encoding method.

[0048] In this example, the unified parsing layer achieves "protocol adaptation-hardware decoupling." The unified parsing layer parses the core parameters of each protocol (such as the 29-bit ID of CAN XL → standardized feature vector) through a "preset mapping relationship." Essentially, this decouples the protocol adaptation logic from the hardware interface. Hardware compatibility theory dictates that existing vehicle controller hardware interfaces already support the physical layer protocols of CAN XL / CAN FD / Ethernet; adaptation is only required at the software layer through a "mapping table," without modifying the hardware circuitry.

[0049] For example, the "variable data field" output by the CAN FD node can be interpreted as a "data volume" characteristic by software without replacing the hardware chip. Theoretically, it can be compatible with more than 80% of mainstream controllers (because mainstream controllers all comply with ISO11898-1 / CAN XL and IEEE 802.3 Ethernet standards).

[0050] Step S120: Collect load characteristic parameters of the vehicle network operation status to form a load characteristic vector;

[0051] The load characteristic parameters may include, but are not limited to, bandwidth utilization, message collision frequency, node response time, and vehicle operating status parameters.

[0052] In practical implementation, the bandwidth utilization rate can be calculated by using the gateway bus monitoring module to statistically determine the proportion of total data traffic to available bandwidth within a unit time window. The bandwidth utilization rate is calculated by dividing the amount of data transmitted per unit time by the available communication bandwidth. The message collision frequency is calculated based on the ratio of message retransmission count to the total number of messages. The node response time is obtained by statistically analyzing the difference between message sending and feedback time. The vehicle operating status parameters are determined by the vehicle behavior state machine output by the autonomous driving decision module, using state machine mapping. Specifically, the vehicle operating status parameters are output by the behavior state machine in the autonomous driving decision module, and the behavior state machine converts the driving state into a quantifiable vehicle operating status code through a preset state machine mapping rule.

[0053] For example, the bandwidth utilization rate is obtained through real-time link statistics and is recorded as 0.63; the message collision rate is obtained based on the collision statistics within the preset sampling window and is recorded as 0.047; the average node response time is obtained through ECU request-response cycle measurement and is recorded as 6.5ms; the vehicle behavior state is output by the autonomous driving decision module and "AEB_ACTIVE" is converted into state code 3 through state machine mapping, thereby forming the load feature vector [0.63, 0.047, 6.5, 3].

[0054] Step S130: Based on the three-dimensional feature vector and the load feature vector, perform load prediction on the network load trend for a preset time period using the GRU model; the load prediction result is the output of the GRU model, including predicted bandwidth utilization, packet collision probability and communication latency trend.

[0055] Among them, the predicted bandwidth utilization rate is the future communication bus bandwidth utilization trend, the predicted message collision probability is the probability of retransmission and arbitration failure, and the predicted communication delay trend is the predicted end-to-end delay value of the message.

[0056] In this embodiment, the input to the GRU model is a time-series feature matrix concatenated from message-level feature vectors and system load feature vectors. The input window uses the past 10-50 scheduling cycles (corresponding to approximately 500ms of historical sequence data) to predict network load trends within the next 100-300ms. The GRU model, with its inherently suited "short-sequence load prediction" scenario, effectively captures the short-term correlation of load fluctuations in vehicular networks. For example, the pattern of increased steering signals and vehicle control messages shortly after braking is triggered. According to time-series prediction theory, when the input feature dimension (12 types of load parameters) matches the prediction window (100ms), the GRU model error rate can be stably controlled within 5%, sufficient to trigger arbitration adjustments in advance and avoid increased time consumption due to "passive congestion response."

[0057] In this embodiment, the three-dimensional feature vector and the load feature vector are concatenated to construct the model input sequence input to the GRU model. For example, the input vector in the above step is: [1.00,0.98,0.03125,0.63,0.047,6.5,3].

[0058] For example, in an automatic emergency braking triggering scenario, the system samples the input sequence of the GRU model in the last 500ms. For example, the input feature vector [1.00,0.98,0.03125,0.63,0.047,6.5,3] changes over time to form a training window.

[0059] Step S140: Calculate the scheduling trigger determination parameters based on the prediction results, and make a trigger determination based on the scheduling trigger determination parameters and the preset dynamic threshold;

[0060] In practice, scheduling triggering judgment parameters can be generated by weighted fusion based on the importance of different indicators in the prediction results.

[0061] Specifically, the scheduling trigger determination parameters are generated by weighted fusion prediction of bandwidth occupancy, prediction of message collision probability and prediction of communication delay, and compared with dynamic thresholds to determine whether to perform arbitration adjustment.

[0062] For example, the GRU model predicts an output of [0.87, 0.091, 14.2ms], and the weighted fusion calculation yields a trigger decision parameter of 0.6043.

[0063] In practice, the dynamic threshold can be calculated based on vehicle operating conditions, protocol environment and safety status, or it can be preset. For example, the dynamic threshold can be preset to 0.3 or 0.5.

[0064] The step of triggering based on the scheduling triggering determination parameters and a preset dynamic threshold includes:

[0065] Based on the dynamic threshold setting, three levels of load threshold conditions are set. When the scheduling trigger determination parameters meet the first level load threshold condition, regular arbitration is performed; when the scheduling trigger determination parameters meet the second level load threshold condition and the third level load threshold condition, the trigger condition is determined to be met.

[0066] In practice, for example, the dynamic thresholds are preset to 0.3 and 0.5. Based on the dynamic thresholds, three levels of load threshold conditions are set as follows:

[0067] Level 1 load threshold condition: If the scheduling trigger judgment parameter is less than 0.3, perform regular arbitration.

[0068] Level 2 load threshold condition: scheduling trigger judgment parameter 0.30–0.5;

[0069] Level 3 load threshold condition: The scheduling trigger judgment parameter is greater than 0.5.

[0070] Step S150: When the triggering condition is met, perform arbitration adjustment;

[0071] In this step, when the triggering condition is met, arbitration adjustment is performed, which can reduce the risk of communication blockage within a preset time period (e.g., 100ms).

[0072] The arbitration adjustment includes:

[0073] Step S1501: Based on the security level, latency sensitivity, and communication resource consumption requirements of the message to be transmitted, generate a three-dimensional evaluation matrix including security weight, real-time coefficient, and load redundancy, and obtain the arbitration priority score based on the three-dimensional evaluation matrix.

[0074] In this step, the security weight is used to reflect the importance of the message type in the functional safety system, the real-time coefficient is used to characterize the sensitivity of the message to communication latency, and the load redundancy reflects the degree of matching between the currently available network resources and the communication resources required by the message.

[0075] In this step, the "security weight × real-time coefficient × load redundancy" matrix is ​​essentially a "precise match between demand and capability". As we know from the system resource scheduling theory, when data demand (such as entertainment data "low security - low real-time") matches protocol capability (such as CAN FD "medium bandwidth - medium real-time"), the waste of "high-capacity protocol carrying low-demand data" can be avoided.

[0076] For example, non-real-time diagnostic data (security weight 0.3, real-time performance 0.2) theoretically does not need to occupy Ethernet gigabit bandwidth. After being downgraded to CAN FD transmission, the idle Ethernet bandwidth can be released to high-frequency data such as LiDAR point clouds, theoretically increasing the utilization rate by more than 50%.

[0077] In this embodiment, the design of the "safety weight" in the three-dimensional evaluation matrix follows the "functional safety priority principle": According to ISO 26262 theory, in autonomous driving, "braking signals and LiDAR data" belong to "safety-critical data," and their priority must be higher than non-critical data such as "entertainment and diagnostics." By setting the safety weight to 1.0 (non-critical data ≤ 0.3), it can theoretically be ensured that: regardless of load changes, safety data always has the highest priority in arbitration decisions and will not be blocked by low-priority data; the transmission success rate theoretically approaches 100% (99.99% is a theoretically conservative value).

[0078] In this embodiment, the three-dimensional evaluation matrix is ​​generated based on the packet's own attributes and the current communication resource status, while the GRU prediction results are used to identify network congestion trends in advance and determine whether to initiate the arbitration optimization process.

[0079] In practice, the security level can be set according to the following data:

[0080] For example: when the message type is braking / steering, the safety level ASIL is ASIL-D and the safety weight is 1.00;

[0081] The message type is power control, the security level is risk level ASIL-C, and the security weight is 0.85;

[0082] The message type is perception fusion, the security level is risk level ASIL-B, and the security weight is .65;

[0083] The message type is a V2X event, the security level is risk level ASIL-A, and the security weight is 0.50;

[0084] The message type is OTA / Entertainment, the security level is QM, and the security weight is 0.10.

[0085] The real-time coefficient is based on the message period and the maximum allowable transmission delay. For example, the real-time coefficient = min(1, message period / maximum allowable transmission delay). For example, the trajectory planning message period requires 10ms and the maximum allowable delay is 20ms: the real-time coefficient = min(1, 10 / 20) = 0.5.

[0086] Among them, load redundancy is related to the current actual operating status of the network, and is equal to the current remaining bandwidth / required bandwidth. If the load redundancy is low, it means that the current resources are tight and packets should be given priority.

[0087] In practice, the system converts the three-dimensional evaluation matrix into comparable scores for ranking arbitration.

[0088]

[0089] For example, safety weight = 1.00 (risk level ASIL-D);

[0090] Real-time performance factor = 0.95 (requirement ≤ 10ms);

[0091] Load redundancy = 0.42;

[0092] ;

[0093] Step S1502: Based on the arbitration priority score, execute the predictive time slot reservation mechanism, and dynamically rearrange the priorities of the packets to be transmitted and schedule communication resources.

[0094] In practice, the scheduling level can be determined based on the arbitration priority score, and then the priority can be dynamically rearranged and communication resources can be scheduled based on the assigned scheduling level.

[0095] For example, the arbitration priority score is ≥0.80, and the scheduling level is Class A (preemptive): immediate scheduling, other messages can be interrupted;

[0096] The arbitration priority score is between 0.60 and 0.79, and the scheduling level is Class B (high-speed scheduling level): reserved time slots, priority transmission;

[0097] The arbitration priority score is between 0.30 and 0.59, and the scheduling level is Class C (controlled scheduling level): queued sending, resources can be shared;

[0098] The arbitration priority score is ≤0.29, and the scheduling level is Class D (background level): it can be delayed, paused, or downgraded.

[0099] Optionally, the conflict scheduling strategy of the communication resource scheduling adopts a time slot division mechanism to allocate independent time resource segments to messages of different protocols or different levels in order to reduce arbitration conflicts.

[0100] Implementing a predictive time slot reservation mechanism and dynamically prioritizing and scheduling communication resources for packets to be transmitted includes:

[0101] Before the scheduling trigger determination parameters meet the second-level load threshold conditions, independent communication time slots are reserved for critical messages; when an automatic emergency braking trigger signal is detected, the communication resource scheduling prioritizes the execution of the predictive time slot reservation strategy, mapping the braking control message to the Ethernet TSN channel.

[0102] For example, in an automatic emergency braking triggering scenario, the arbitration priority score of the braking control message is 0.92, classifying it as Class A. The system executes a predictive time slot reservation strategy and inserts the message at the front of the scheduling queue, while simultaneously suspending OTA update messages and in-vehicle entertainment traffic. Subsequently, the message is migrated to the Ethernet TSN channel and sent using a preemptive scheduling method, ensuring that the end-to-end transmission latency remains within 10ms.

[0103] In this step, the core problem of the traditional "passive response to congestion" approach is that "resource allocation lags behind load changes." However, by using the GRU model to predict in advance, it can determine whether to make arbitration adjustments. For example, before the scheduling trigger judgment parameters meet the second-level load threshold conditions, resource scheduling theory shows that "advance allocation" can avoid the contradiction of "competing for resources when overloaded" and "resources being idle when under low load."

[0104] For example, the scheduling trigger determination parameters are calculated based on the prediction results of the GRU model. Before the scheduling trigger determination parameters meet the second-level load threshold conditions, 20% of the bandwidth is reserved for safe data in advance, and non-real-time data is transferred to CANFD. In theory, this can allow each protocol bus to always run in the high-efficiency load range, and the overall utilization rate can be increased from 40% to 85%.

[0105] In this embodiment, the traditional CAN protocol's "bit-by-bit comparison arbitration" has an inherent risk of conflict (multiple nodes need to compete bit by bit when sending messages simultaneously). The TDM time slot allocation mechanism theoretically eliminates this conflict through "preset time slot allocation": independent time slots are allocated for CANXL, CAN FD, and Ethernet protocol messages. For example, CAN XL protocol messages occupy 20% of the time slots to transmit large data packets, while CAN FD protocol messages occupy 30% of the time slots to transmit real-time small messages. Each protocol message exclusively occupies the bus within its dedicated time slot, eliminating the need for arbitration bit competition. Communication theory shows that the conflict handling time for time slot allocation depends only on the "time slot switching delay" (theoretically, it can be as low as...). (seconds), far lower than traditional bitwise comparisons. This reduces arbitration time significantly.

[0106] In practice, bandwidth limits are imposed on OTA software upgrade traffic. When the scheduling trigger judgment parameters meet the second-level and third-level load threshold conditions, OTA message transmission is suspended to ensure the real-time performance of autonomous driving control messages.

[0107] This invention parses and converts messages from different communication protocols (CAN FD, CAN XL, Ethernet) into a unified three-dimensional feature vector, enabling comparison and arbitration of multi-protocol messages within the same decision model. This improves the consistency and compatibility of cross-protocol communication and achieves unified scheduling and control capabilities across multiple protocols. By utilizing the GRU model to predict future load trends (bandwidth utilization, message collision probability, and communication latency trends) of the communication network in advance, arbitration strategy adjustments can be triggered before load deterioration, avoiding the lag problem of "arbitration after congestion" in traditional scheduling methods. It achieves predictive network load perception and dynamic scheduling. Furthermore, by constructing a three-dimensional evaluation matrix including security weights, real-time coefficients, and load redundancy, an arbitration priority score is generated. Combined with a predictive time slot reservation mechanism, critical control messages (such as braking, steering, and trajectory planning) maintain low-latency and high-reliability communication capabilities under high load conditions, ensuring deterministic communication performance for critical safety messages.

[0108] Meanwhile, the present invention adopts a mechanism of predicting in advance and then driving dynamic priority rearrangement and communication resource scheduling, so that low-priority messages and non-real-time data give way reasonably, and has adaptive scheduling capability, thereby reducing message collision rate, alleviating network bottlenecks, improving overall communication efficiency, and enhancing network resource utilization efficiency and scheduling stability.

[0109] Optionally, the priority dynamic rearrangement and communication resource scheduling include dynamic channel switching scheduling and protocol bandwidth occupancy adjustment: when the scheduling trigger determination parameter meets the second-level load threshold condition, the CAN XL protocol message arbitration priority strategy is adopted; when the scheduling trigger determination parameter meets the third-level load threshold condition, Ethernet bandwidth reservation is initiated, and non-real-time data is transmitted through the CAN FD protocol.

[0110] Among them, the CAN XL protocol message arbitration priority strategy prioritizes the transmission opportunity of CAN XL messages in conflict arbitration when the network load reaches the second level threshold (i.e., moderate congestion), giving them a higher arbitration priority when competing with CAN FD or other protocols.

[0111] During the above implementation process, the overall utilization rate of multi-protocol bus bandwidth increased from 40% to over 85%, the idle bandwidth utilization rate of Ethernet decreased by 70%, and the congestion rate of CAN FD bus dropped to below 5%.

[0112] Optionally, the method of the present invention adopts arbitration priority score inheritance to maintain the original arbitration priority score when the message is migrated across protocols; and when a V2X environment event warning message is received, the cross-protocol priority inheritance mechanism is executed to migrate the message to a higher bandwidth communication protocol channel.

[0113] In this step, an arbitration priority score inheritance mechanism is executed when a message is migrated across protocols, so that the migrated message retains the arbitration priority score before the migration; and when a V2X emergency environmental event warning message is received, a cross-protocol scheduling upgrade is performed on the message, so that it is migrated to a communication protocol channel with higher bandwidth capabilities for transmission.

[0114] In this step, priority scores are retained during cross-protocol migration to ensure scheduling consistency (otherwise, packets might be downgraded and delayed or blocked after migration). When a packet migrates from one protocol to another, the score is not recalculated; instead, an inheritance mechanism is executed.

[0115] For example, when a vehicle receives a "multi-vehicle emergency braking" warning message from the V2X infrastructure while driving on a highway, the cross-protocol priority inheritance mechanism is triggered during message migration. The score of 0.92 remains unchanged without rollback or recalculation, and the message is migrated from the CAN XL channel to the Ethernet TSN preemption queue. During the migration, the arbitration priority score remains unchanged, and the scheduling is not reordered, reducing the end-to-end transmission delay of the warning message from approximately 22ms to 9ms, thus allowing it to participate in trajectory planning and emergency braking decisions earlier.

[0116] Specifically, the preset time period is 100ms, and the communication resource scheduling ensures that the transmission delay of critical safety messages does not exceed 20ms; when the vehicle is in high-speed automatic lane changing mode, its scheduling level is increased by at least one level according to the real-time coefficient of the trajectory planning message, and the protocol migration strategy is triggered to ensure that the transmission delay of the trajectory planning message is not greater than 10ms.

[0117] In this step, when the detected vehicle is in high-speed automatic lane-changing mode, its scheduling level is raised by at least one level based on the real-time coefficient of the trajectory planning message. The highest scheduling level is Class A. A cross-protocol migration strategy is executed so that the message is sent in the Ethernet time-sensitive network channel with an independent reserved time slot to ensure that the end-to-end transmission delay of the trajectory planning message does not exceed 10ms.

[0118] In this embodiment, when the vehicle performs automatic lane change decision-making in high-speed cruising mode, the trajectory planning module continuously outputs predicted path nodes, and the real-time coefficient of this message is increased to the highest level. After detecting that the vehicle is in high-speed automatic lane change mode, the scheduling level of the trajectory planning message is upgraded from Class B to Class A, and the protocol migration is triggered to transfer the message from the CAN XL communication channel to the Ethernet TSN reserved time slot. Finally, the end-to-end transmission delay of the message is reduced from about 18ms to 8ms, ensuring that the planning execution is synchronized with the vehicle's lateral control and avoiding the risk of vehicle deviation caused by delay.

[0119] In this step, the transmission success rate of autonomous driving safety-related data (such as LiDAR and braking signals) reaches 99.99%, and the adjustment response time is ≤10ms, avoiding the loss of safety signals under high load.

[0120] As a specific implementation, when a change in vehicle operating status is detected or an emergency-category V2X message is received, a dynamic promotion mechanism is executed to adjust the arbitration priority score; when a message is migrated across protocols, a priority inheritance mechanism is executed to keep the arbitration priority score unchanged after the migration.

[0121] Specifically, when an external environmental event or change in vehicle operating status leads to an increase in message security level, the system recalculates the arbitration priority. If a protocol migration is subsequently triggered, a priority inheritance mechanism is executed, ensuring that the migrated message retains the upgraded arbitration priority score, thereby avoiding scheduling level resets or delay fluctuations caused by protocol switching. For example, when a vehicle receives a "chain braking warning" from a roadside unit: the original priority score is 0.72, which is then increased to 0.92. The priority inheritance mechanism is activated, the message is immediately migrated to a new channel, OTA data is paused, entertainment stream scheduling communication is delayed, and the "chain braking warning" message time is reduced from 23ms to 9ms.

[0122] In practice, the protocol migration can be performed by the vehicle communication gateway, and the message data format remains unchanged during the migration process, with only the protocol encapsulation format and transmission channel being updated.

[0123] In this embodiment, the dynamic priority reordering allows the interruption of low-priority data transmission when a high-priority message enters the queue.

[0124] In this embodiment, the OTA pause policy only applies to non-security software upgrades and complies with ISO26262 and ISO21434 communication security standards.

[0125] After verification, the method of the present invention reduces arbitration time by more than 60% under high load scenarios (load rate 60%-80%), and the data transmission delay between CAN XL and Ethernet protocols is ≤20ms, meeting the ISO26262 functional safety ASIL-D level requirements. It is expected that the arbitration time will be reduced by 60%+ under high load, with a delay of ≤20ms, meeting the ASIL-D level safety requirements.

[0126] In this embodiment, edge AI sensing nodes are deployed on the vehicle gateway, enabling real-time collection of 12 types of load characteristics (such as packet collision frequency and node response time), avoiding the transmission delays of "centralized sensing." Edge computing theory dictates that the response time of local sensing + local decision-making depends solely on "hardware computing power" (the mainstream CPU computing power of vehicle gateways has reached 1GHz, and the time to process 12 types of characteristics is ≤5ms). Combined with a dynamic threshold triggering mechanism (activating adjustment upon exceeding a threshold), the priority adjustment response time can theoretically be controlled within 10ms, ensuring that security data is not lost due to adjustment delays.

[0127] In this embodiment, the system of the present invention has compatibility extension support for seamless access to CAN XL / CANFD / Ethernet protocols without the need to modify existing vehicle node hardware, and is compatible with more than 80% of mainstream vehicle controllers.

[0128] In this embodiment, the computing power of the vehicle gateway is typically limited (e.g., ARM Cortex-A55 quad-core 1.5GHz), while the GRU model achieves lightweighting by "simplifying the number of hidden layers (≤3 layers) and reducing feature dimensions (12 types of core parameters)". According to embedded AI theory, the lightweight GRU consumes ≤100M FLOPS of computing power per prediction, far below the computing power limit of the vehicle gateway (≥1GFLOPS), and can run without additional hardware upgrades. Simultaneously, the arbitration decision engine adopts a modular design (independent parsing, prediction, and decision layers), which can be adapted to the interface protocols of different controllers through software upgrades, further reducing the need for hardware modifications.

[0129] Example 2

[0130] This invention discloses an intelligent arbitration system based on a multi-protocol vehicle network, such as... Figure 2 As shown, Figure 2 It is an intelligent arbitration system based on a multi-protocol vehicle network, including:

[0131] The message parsing module 210 is used to parse messages from different communication protocols and convert the parsing results into a three-dimensional feature vector containing data type, real-time characteristics and data size.

[0132] The load acquisition module 220 is used to collect load characteristic parameters of the vehicle network operation status and form a load characteristic vector. The load characteristic parameters include bandwidth utilization, packet collision frequency, node response time and vehicle operation status parameters.

[0133] The prediction calculation module 230 is used to predict the network load trend for a preset time period based on the three-dimensional feature vector and the load feature vector using a GRU model; the load prediction result is the output of the GRU model, including the predicted bandwidth utilization, packet collision probability and communication latency trend.

[0134] The trigger judgment module 240 is used to calculate the scheduling trigger judgment parameters based on the prediction results, and to make a trigger judgment based on the scheduling trigger judgment parameters and a preset dynamic threshold.

[0135] The arbitration adjustment module 250 is used to perform arbitration adjustment when the triggering conditions are met. The arbitration adjustment includes: generating a three-dimensional evaluation matrix containing security weight, real-time coefficient and load redundancy based on the security level, latency sensitivity and communication resource occupation requirements of the message to be transmitted; obtaining an arbitration priority score based on the three-dimensional evaluation matrix; and executing a predictive time slot reservation mechanism based on the arbitration priority score, and dynamically rearranging the priority of the message to be transmitted and scheduling communication resources.

[0136] Example 3

[0137] Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. For example... Figure 3 As shown, the electronic device may include:

[0138] Memory 310 storing executable program code;

[0139] Processor 320 coupled to memory 310;

[0140] The processor 320 calls the executable program code stored in the memory 310 to execute some or all of the steps in the intelligent arbitration method based on a multi-protocol vehicle network in Embodiment 1.

[0141] This invention discloses a computer-readable storage medium storing a computer program that causes a computer to perform some or all of the steps in a smart arbitration method based on a multi-protocol vehicle network as described in Embodiment 1.

[0142] This invention also discloses a computer program product, wherein when the computer program product is run on a computer, the computer performs some or all of the steps in the intelligent arbitration method based on a multi-protocol vehicle network in Embodiment 1.

[0143] This invention also discloses an application publishing platform, which is used to publish computer program products. When the computer program products are run on a computer, the computer performs some or all of the steps in the intelligent arbitration method based on a multi-protocol vehicle network in Embodiment 1.

[0144] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0145] 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; they can be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0146] Furthermore, the functional units in the various embodiments of the present invention 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.

[0147] 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-accessible memory. Based on this understanding, the technical solution of the present invention, 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 memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of the present invention.

[0148] In the embodiments provided by this invention, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0149] Those skilled in the art will understand that some or all of the steps in the various methods of the embodiments described can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0150] The foregoing has provided a detailed description of an intelligent arbitration method, apparatus, electronic device, and storage medium based on a multi-protocol vehicle network disclosed in the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A smart arbitration method based on a multi-protocol vehicular network, characterized in that, It includes the following: The system parses messages from different communication protocols and converts the parsing results into a three-dimensional feature vector containing data type, real-time characteristics, and data size. Collect load characteristic parameters of the vehicle network operation status to form a load feature vector; Based on the three-dimensional feature vector and the load feature vector, the network load trend for a preset time period is predicted using the GRU model; the load prediction result is the output of the GRU model, including predicted bandwidth utilization, packet collision probability and communication latency trend. Based on the load prediction results, the scheduling trigger determination parameters are calculated, and the trigger determination is made according to the scheduling trigger determination parameters and the preset dynamic threshold. When the triggering conditions are met, arbitration adjustment is performed; the arbitration adjustment includes: generating a three-dimensional evaluation matrix containing security weight, real-time coefficient and load redundancy based on the security level, latency sensitivity and communication resource occupation requirements of the message to be transmitted; obtaining an arbitration priority score based on the three-dimensional evaluation matrix; and executing a predictive time slot reservation mechanism based on the arbitration priority score, and dynamically rearranging the priority of the message to be transmitted and scheduling communication resources.

2. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The parsing of messages from different communication protocols includes: Acquire messages from different communication protocols and parse the identifier of CAN XL protocol data, the variable data field length of CAN FD protocol data, and the tag parameters of Ethernet protocol data respectively.

3. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The step of triggering based on the scheduling triggering determination parameters and a preset dynamic threshold includes: Based on the dynamic threshold setting, three levels of load threshold conditions are set. When the scheduling trigger determination parameters meet the first level load threshold condition, regular arbitration is performed; when the scheduling trigger determination parameters meet the second level load threshold condition and the third level load threshold condition, the trigger condition is determined to be met.

4. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The dynamic prioritization and communication resource scheduling of the messages to be transmitted includes dynamic channel switching scheduling and protocol bandwidth occupancy adjustment, specifically including: when the scheduling trigger determination parameter meets the second-level load threshold condition, adopting the CAN XL protocol message arbitration priority strategy; when the scheduling trigger determination parameter meets the third-level load threshold condition, initiating Ethernet bandwidth reservation and transmitting non-real-time data through the CAN FD protocol.

5. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The conflict scheduling strategy for communication resource scheduling adopts a time slot division mechanism to allocate independent time resource segments to messages of different protocols or different levels in order to reduce arbitration conflicts. The implementation of a predictive time slot reservation mechanism, and the dynamic prioritization and communication resource scheduling of the packets to be transmitted include: reserving independent communication time slots for critical packets before the scheduling trigger judgment parameters meet the second-level load threshold conditions; When an automatic emergency braking trigger signal is detected, the communication resource scheduling prioritizes the execution of a predictive time slot reservation strategy to map the braking control message to the Ethernet TSN channel.

6. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The method adopts arbitration priority score inheritance to maintain the original arbitration priority score when the message is migrated across protocols; When a V2X environment event warning message is received, a cross-protocol priority inheritance mechanism is executed to migrate the message to a higher bandwidth communication protocol channel.

7. The intelligent arbitration method based on a multi-protocol vehicular network according to claim 1, characterized in that, The preset time period is 100ms, and the communication resource scheduling ensures that the transmission delay of critical safety messages does not exceed 20ms. When the vehicle is in high-speed automatic lane changing mode, its scheduling level is raised by at least one level according to the real-time coefficient of the trajectory planning message. The raised scheduling level is less than or equal to the highest level, and the protocol migration strategy is triggered to ensure that the transmission delay of the trajectory planning message is not greater than 10ms.

8. An intelligent arbitration system based on a multi-protocol vehicle network, characterized in that, It includes: The message parsing module is used to parse messages from different communication protocols and convert the parsing results into a three-dimensional feature vector containing data type, real-time characteristics and data size. The load acquisition module is used to collect load characteristic parameters of the vehicle network operation status and form a load feature vector. The prediction calculation module is used to predict the network load trend for a preset time period based on the three-dimensional feature vector and the load feature vector using the GRU model; the load prediction result is the output of the GRU model, including the predicted bandwidth utilization, packet collision probability and communication latency trend. The triggering judgment module is used to calculate the scheduling triggering judgment parameters based on the load prediction results, and to make a triggering judgment based on the scheduling triggering judgment parameters and a preset dynamic threshold. The arbitration adjustment module is used to perform arbitration adjustment when the triggering conditions are met; The arbitration adjustment includes: generating a three-dimensional evaluation matrix containing security weights, real-time coefficients, and load redundancy based on the security level, latency sensitivity, and communication resource usage requirements of the messages to be transmitted; obtaining an arbitration priority score based on the three-dimensional evaluation matrix; and executing a predictive time slot reservation mechanism based on the arbitration priority score, and dynamically rearranging the priorities and scheduling communication resources for the messages to be transmitted.

9. An electronic device, characterized in that, It includes: a memory storing executable program code; a processor coupled to the memory; the processor calling the executable program code stored in the memory to execute the intelligent arbitration method based on a multi-protocol vehicle network as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, wherein the computer program causes the computer to execute the intelligent arbitration method based on a multi-protocol vehicle network as described in any one of claims 1-7.

Images