Large-scale motorcade formation control method and system

Abstract

The invention provides a large-scale motorcade formation control method and system, and belongs to the technical field of intelligent network connection automobile formation, the method is applied to a global management node G node, and the method comprises the following steps: establishing star flash wireless communication connection with a plurality of terminal execution nodes T nodes; obtaining resource state information of each T node, and maintaining a distributed system resource state database; decomposing a cooperative driving calculation task into a plurality of sub-tasks, and dynamically allocating the sub-tasks to corresponding T nodes based on resource state information; receiving a local intelligent driving conclusion uploaded by each T node; and carrying out fusion processing on all local intelligent driving conclusions, generating a global control instruction, and issuing the global control instruction to each T node for execution. According to the method, the formation is constructed into a distributed computing platform, parallel processing of computing tasks and dynamic resource scheduling are achieved, the upper limit of single-node computing power is broken through, communication bandwidth occupation is reduced, the single-point fault risk is eliminated, and the cooperative control efficiency of the large-scale vehicle formation is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent connected vehicle platooning technology, and in particular to a method and system for controlling large-scale vehicle platooning. Background Technology

[0002] Intelligent connected vehicle platooning technology can effectively improve road traffic efficiency and reduce fuel consumption through collaborative perception and decision-making among vehicles. Existing platooning control schemes typically adopt a centralized architecture, with the lead vehicle acting as a single master node responsible for global path planning and issuing decision commands, while following vehicles only execute control commands and upload perception data.

[0003] However, when dealing with large-scale vehicle platoons (such as 100-500 or more vehicles), existing technologies have the following shortcomings: the computing power of the centralized master node has a physical limit, and the computing latency increases exponentially as the platoon size expands, which cannot meet the requirements of real-time control; all raw sensing data must be uploaded to the master node, which consumes a large amount of communication bandwidth and is prone to network congestion and packet loss; the system performance is overly dependent on the health status of the master node, and a single point of failure will cause the entire platoon to fail.

[0004] Therefore, how to achieve low-latency, high-reliability, and scalable collaborative control in large-scale formation scenarios has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art, and proposes a method and system for large-scale vehicle formation control.

[0006] In a first aspect, embodiments of the present invention provide a large-scale vehicle formation control method, applied to a system comprising a global management node G and multiple terminal execution nodes T, wherein the method is executed by the G node and includes:

[0007] Establish a star-flash wireless communication connection with multiple T nodes, obtain the resource status information of each T node, and maintain a distributed system resource status database; wherein, the distributed system resource status database is composed of multiple T nodes and is used to record various resource status information of each T node in real time.

[0008] The computational task of cooperative driving is decomposed into multiple subtasks that can be processed in parallel, and the subtasks are dynamically allocated to the corresponding T nodes based on the resource status information.

[0009] It receives local intelligent driving conclusions uploaded by each T node, performs fusion processing on all received local intelligent driving conclusions, generates global control commands, and sends them to each T node for execution.

[0010] In some embodiments, obtaining the resource status information of each T node includes:

[0011] The G node receives heartbeat data packets sent by each T node at a high frequency. The heartbeat data packets contain update information on node identifier, computing resource dimension, and task status dimension.

[0012] The G node updates the information of the communication resource dimension according to the reception time and data quality of the heartbeat data packet;

[0013] The G node detects the online status of the T node based on the reception of the heartbeat data packet. If it does not receive a heartbeat data packet from a certain T node within a set time, it marks the T node as offline and stops assigning tasks to it.

[0014] In some embodiments, the fusion process includes:

[0015] The G node integrates the local intelligent driving information uploaded by each T node into a global high-precision dynamic map;

[0016] The G node performs conflict detection on the conclusions of different T nodes regarding the same target, and confirms the authenticity and risk level of the target through a weighted voting mechanism;

[0017] The G node generates global control commands based on the fusion results.

[0018] In some embodiments, the method further includes: when the G node detects a conflict between the local intelligent driving conclusion of the T node and the conclusion calculated by the G node itself, or receives an emergency instruction from the T node, the G node activates the high-speed computing module to re-analyze the data.

[0019] The G node updates the conclusion weight and credibility of the corresponding T node based on the results of the reanalysis.

[0020] In some embodiments, the method further includes:

[0021] The G node assigns some T nodes as child G nodes. The child G nodes are responsible for the allocation of sub-tasks and the fusion of local conclusions within their management scope, and then upload the fusion results to the main G node.

[0022] Secondly, embodiments of the present invention provide a large-scale vehicle formation control method, applied to a system comprising a global management node G and multiple terminal execution nodes T, wherein the method is executed by any T node and includes:

[0023] Establish a StarScan wireless communication connection with the global management node G, and send a heartbeat data packet to the G node. The heartbeat data packet contains the resource status information of the T node.

[0024] The system receives subtasks assigned by the G node, acquires surrounding environmental information through local sensors, processes the assigned subtasks locally, generates local intelligent driving conclusions, and uploads the local intelligent driving conclusions to the G node; wherein, the local intelligent driving conclusions include perception conclusions, prediction conclusions, and local decision suggestions.

[0025] Receive and execute global control commands issued by the G node.

[0026] In some embodiments, generating local intelligent driving conclusions includes:

[0027] The T-node identifies the target and generates a perception conclusion that includes the target type, location, and confidence level.

[0028] The T node generates a prediction conclusion that includes a collision time prediction;

[0029] The T-node generates local decision-making suggestions for the vehicle and its formation.

[0030] Thirdly, embodiments of the present invention provide a large-scale vehicle platooning control system, comprising:

[0031] A global management node G is deployed on any vehicle in the platoon to execute the large-scale vehicle platooning control method provided above for application to the G node;

[0032] Multiple terminal execution nodes T nodes, each T node corresponding to the remaining vehicles in the formation, and each T node is used to execute the large-scale vehicle formation control method provided above for application to T nodes;

[0033] The G node and T node exchange information via a star-flash wireless communication network.

[0034] In some embodiments, the distributed system resource status database includes the following dimensions of information:

[0035] The node identifier dimension includes the unique identifier and geographical location of node T;

[0036] The computing resource dimension includes processor model, number of cores, CPU utilization, and remaining memory;

[0037] The communication resource dimension includes the signal-to-noise ratio and network latency of the T-node and G-node links;

[0038] Sensor status dimension, including the model and data quality of lidar and millimeter-wave radar;

[0039] The task status dimension includes the number of subtasks currently being executed, the average completion time of historical tasks, and the average completion quality.

[0040] In some embodiments, the system also supports a layered architecture:

[0041] The G node will allocate some T nodes as child G nodes. Each child G node is responsible for managing the allocation of subtasks and the fusion of local conclusions within its jurisdiction, and will upload the fusion results to the main G node.

[0042] This invention provides a large-scale convoy control method where the G node decomposes the computational task of cooperative driving into multiple sub-tasks and dynamically allocates them to the corresponding T nodes based on resource status information. This design distributes the entire computational load, originally borne by a single master node, across the computing units of all vehicles in the convoy for parallel processing. The T nodes are no longer merely terminals executing instructions, but participate in computation as part of a distributed computing resource pool. The resulting technical effects are: breaking through the physical limit of single-node computing power, eliminating the exponential growth in computational latency as the convoy size increases; and effectively utilizing the previously idle computing resources of following vehicles, significantly improving the overall energy efficiency of the system.

[0043] After processing its assigned subtasks locally, node T only uploads its local intelligent driving conclusions to node G, rather than uploading the raw perception data. Since the amount of local intelligent driving conclusions (such as target type, location, confidence level, and decision recommendations) is much smaller than the raw point cloud or image data, communication bandwidth consumption is significantly reduced. The resulting technical benefits are: alleviating network congestion, reducing transmission latency and packet loss risk, and enabling real-time collaborative control of large-scale formations. Node G plays the role of task scheduling and result fusion, rather than being the sole brain responsible for all computations as in traditional solutions.

[0044] By maintaining a distributed system resource status database, the G node can monitor the resource status and online status of each T node in real time, enabling dynamic task allocation and fault detection. When a T node goes offline, the G node can mark it and stop assigning new tasks, while the system continues to operate normally. The resulting technical benefits are: eliminating absolute dependence on a single node, giving the system fault immunity; and allowing for dynamic expansion of the platoon size, with newly added vehicles automatically becoming part of the resource pool, significantly enhancing system flexibility. Attached Figure Description

[0045] Figure 1 A flowchart illustrating a large-scale vehicle formation control method applied to a G-node, provided as an embodiment of the present invention;

[0046] Figure 2 A flowchart illustrating a large-scale vehicle formation control method applied to a T-node, provided by an embodiment of the present invention;

[0047] Figure 3 This is a structural block diagram of an application scenario of the method provided in the embodiments of the present invention;

[0048] Figure 4 This is a schematic diagram of the distributed task scheduling process between G nodes and T nodes in the method provided in the embodiments of the present invention;

[0049] Figure 5 This is a schematic diagram of the information fusion processing of the G node in the method provided in the embodiment of the present invention. Detailed Implementation

[0050] To enable those skilled in the art to better understand the technical solutions of the present invention, exemplary embodiments of the present invention are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0051] Where there is no conflict, the various embodiments of the present invention and the features thereof may be combined with each other.

[0052] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0053] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0054] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having the meaning consistent with their meaning in the context of the relevant art and the invention, and will not be interpreted as having an idealized or overly formal meaning unless expressly so defined herein.

[0055] In the technical solution of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information all comply with relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution follows relevant national laws and regulations (e.g., the "Information Security Technology - Personal Information Security Specification"). For example: appropriate measures are taken for personal information access control; restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; and explicit identity targeting is eliminated when using personal information to avoid precisely locating a specific individual.

[0056] Existing intelligent connected vehicle cooperative driving technologies face the following insurmountable technical bottlenecks when dealing with large-scale vehicle platoons (such as 100-500 vehicles or even larger):

[0057] 1. Centralized computing bottleneck: The computing model that uses a single master node (such as the lead vehicle) for global decision-making has a physical limit to its computing power. As the formation size increases, the computing latency grows exponentially, which cannot meet the requirements of real-time control.

[0058] 2. Data communication bottleneck: All raw sensing data is uploaded to the master node, which consumes a lot of communication bandwidth, causing network congestion and increasing transmission delay and packet loss risk.

[0059] 3. Poor system flexibility and single point of failure: The system performance is overly dependent on the health of a single master node. Once the master node fails, the entire queuing system faces the risk of failure. Furthermore, the system's computing power is fixed and cannot be dynamically expanded.

[0060] 4. Low resource utilization efficiency: The computing resources of a large number of following vehicles (T nodes) in the formation are idle, failing to form a synergistic effect, resulting in low overall system energy efficiency.

[0061] The core idea of ​​this invention is to initiate a paradigm shift from "centralized computing" to "collaborative computing." We construct the entire vehicle platoon as a dynamic and scalable "distributed computing platform," enabling the computing capabilities of all vehicles to work collaboratively through a complete mechanism.

[0062] To address at least one of the technical problems existing in the aforementioned related technologies, this invention provides a large-scale vehicle formation control method, applied to a system comprising a global management node G and multiple terminal execution nodes T, wherein the method is executed by the G node, referencing... Figure 1 ,include:

[0063] Establish a star-flash wireless communication connection with multiple T nodes, obtain the resource status information of each T node, and maintain a distributed system resource status database; wherein, the distributed system resource status database is composed of multiple T nodes and is used to record various resource status information of each T node in real time.

[0064] The computational task of cooperative driving is decomposed into multiple subtasks that can be processed in parallel, and the subtasks are dynamically allocated to the corresponding T nodes based on the resource status information.

[0065] It receives local intelligent driving conclusions uploaded by each T node, performs fusion processing on all received local intelligent driving conclusions, generates global control commands, and sends them to each T node for execution.

[0066] A system architecture for establishing a logically centralized, physically distributed vehicle-mounted computing cluster is proposed. This architecture can be divided into two logical node types: G nodes and T nodes. Information exchange between different nodes will utilize high-speed, low-latency, and highly reliable wireless signal transmission technologies, such as StarFlash 3.0. Figure 3 As shown.

[0067] G-node (Global Management Node): As the "brain" and "scheduler" of the system, it is responsible for three core functions: task decomposition, resource scheduling, and result fusion, and global decision-making. Specifically, it breaks down complex problems, then acts as a central hub to schedule problem-solving, relies on global computing resources to solve the problems, and finally integrates the computational results in the G-node to issue the final instructions.

[0068] T-node cluster (terminal execution node cluster): Represents all vehicles in the formation. Each T-node is an independent intelligent computing unit, and all managed vehicles together form a scalable distributed computing resource pool. Each T-node is not only an execution terminal but also a powerful computing unit.

[0069] refer to Figure 4 The first step is spatial modeling of the environment surrounding the vehicle platoon. This involves establishing connections between the G node and all T nodes to determine the size of the vehicle platoon. Then, the entire surrounding environment is decomposed into multiple parallel subtasks. This decomposition strategy is strongly coupled with the spatial location of the vehicles and the sensor field of view. Each environment is divided into computational networks of varying sizes based on the T node's computing resources, communication resources, and sensor status, and then allocated to the corresponding T node.

[0070] Node G is responsible for maintaining an internal dynamic, real-time, and multi-dimensional distributed system resource status database. Its core function is to provide accurate and quantifiable decision-making basis for distributed task scheduling algorithms, ensuring that each computational subtask is assigned to the most suitable node T for execution, thereby optimizing the overall system performance.

[0071] In some embodiments, obtaining the resource status information of each T node includes:

[0072] The G node receives heartbeat data packets sent by each T node at a high frequency. The heartbeat data packets contain update information on node identifier, computing resource dimension, and task status dimension.

[0073] The G node updates the information of the communication resource dimension according to the reception time and data quality of the heartbeat data packet;

[0074] The G node detects the online status of the T node based on the reception of the heartbeat data packet. If it does not receive a heartbeat data packet from a certain T node within a set time, it marks the T node as offline and stops assigning tasks to it.

[0075] The distributed system resource status database contains various resource and status dimensions for different nodes T. The database includes information such as node identifier, computing resources, communication resources, sensor status, and task status.

[0076] The node identification dimension consists of the T node ID as the unique identifier of the node and the actual geographical location of the T node itself obtained from the positioning.

[0077] The computing resource dimension consists of resources such as processor model, number of cores, CPU utilization, and remaining memory size.

[0078] The communication resource dimension consists of resources such as the signal-to-noise ratio and network latency between the T node and the G node link.

[0079] The sensor state dimension consists of different models of lidar, millimeter-wave radar and their corresponding data quality.

[0080] The task status dimension consists of the number of subtasks currently being executed, the average completion time of historical tasks, and the average completion quality.

[0081] Each T node sends lightweight heartbeat data packets to the G node at a high-frequency communication rate, updating the resource database of the G node in dimensions such as node identifier, computing resources, and task status.

[0082] After receiving the heartbeat packet data, the G node updates the resource library of the communication resource dimension based on the reception time and data quality.

[0083] The G node allocates distributed tasks to different nodes according to its own algorithm and updates the task status dimension in the resource table. At the same time, the G node also monitors the heartbeat frequency of each T node. If a T node does not transmit heartbeat data packets after a period of time, it will be marked as offline and will stop being assigned new tasks.

[0084] In some embodiments, the fusion process includes:

[0085] The G node integrates the local intelligent driving information uploaded by each T node into a global high-precision dynamic map;

[0086] The G node performs conflict detection on the conclusions of different T nodes regarding the same target, and confirms the authenticity and risk level of the target through a weighted voting mechanism;

[0087] The G node generates global control commands based on the fusion results.

[0088] refer to Figure 5 Information fusion processing of G nodes:

[0089] The T-node's own sensors acquire surrounding road information, quickly identifying vehicles, pedestrians, lane lines, and other information, generating corresponding pixel coordinates and confidence levels. Then, the intelligent driving information status and conclusions are fused with the T-node's own status, and when sending a heartbeat, the final intelligent driving data and conclusions are transmitted to the G-node.

[0090] The G node receives signal transmissions from all T nodes and stitches together all local intelligent driving information. Figure 1 This is integrated into a global, unified, high-precision dynamic map. After comparing and fusing the intelligent driving conclusions transmitted from different T-nodes with its own calculated intelligent driving conclusions, the final intelligent driving conclusion command is issued to each T-node.

[0091] In some embodiments, the method further includes: when the G node detects a conflict between the local intelligent driving conclusion of the T node and the conclusion calculated by the G node itself, or receives an emergency instruction from the T node, the G node activates the high-speed computing module to re-analyze the data.

[0092] The G node updates the conclusion weight and credibility of the corresponding T node based on the results of the reanalysis.

[0093] When the intelligent driving conclusion of node T conflicts with the intelligent driving conclusion calculated by node G itself, or when node T issues certain emergency commands, node G will slow down and activate its own high-speed computing module to perform a more detailed and accurate intelligent driving situation analysis. This updates the weight ratio of the conclusions of different nodes T and the reliability of the node data.

[0094] In some embodiments, the method further includes: the G node assigns some T nodes as child G nodes, the child G nodes are responsible for the sub-task allocation and local conclusion fusion within their management scope, and upload the fusion results to the main G node.

[0095] In this embodiment, the specific interaction details can be formulated based on the logical judgment of the G node and the signal quality of the T node. The T node can directly upload vehicle and road information, use its own computing resources to judge intelligent driving information, and calculate intelligent driving conclusions.

[0096] The G node is broken down into smaller formations. In addition to the main G node, some T stages are also assigned as sub-G nodes. Intelligent driving tasks are assigned layer by layer, and intelligent driving conclusions are also uploaded layer by layer.

[0097] In some embodiments, a large-scale vehicle formation control method is provided, applied to a system comprising a global management node G and multiple terminal execution nodes T, wherein the method is executed by any T node, referencing Figure 2 ,include:

[0098] Establish a StarScan wireless communication connection with the global management node G, and send a heartbeat data packet to the G node. The heartbeat data packet contains the resource status information of the T node.

[0099] The system receives subtasks assigned by the G node, acquires surrounding environmental information through local sensors, processes the assigned subtasks locally, generates local intelligent driving conclusions, and uploads the local intelligent driving conclusions to the G node; wherein, the local intelligent driving conclusions include perception conclusions, prediction conclusions, and local decision suggestions.

[0100] Receive and execute global control commands issued by the G node.

[0101] In some embodiments, generating local intelligent driving conclusions includes:

[0102] The T-node identifies the target and generates a perception conclusion that includes the target type, location, and confidence level.

[0103] The T node generates a prediction conclusion that includes a collision time prediction;

[0104] The T-node generates local decision-making suggestions for the vehicle and its formation.

[0105] In some embodiments, a large-scale vehicle formation control method is provided, applied to a system comprising a global management node (G node) and multiple terminal execution nodes (T nodes), wherein the method is executed by any T node and includes the following steps:

[0106] This embodiment describes the execution process of a T-node in distributed collaborative platoon control from the perspective of the T-node. As a following vehicle in the platoon, the T-node is not only the terminal for executing control commands, but also an active participant in the distributed computing resource pool.

[0107] First, node T establishes a star-flash wireless communication connection with node G and sends heartbeat data packets to node G at high frequency. These heartbeat data packets carry node T's resource status information, such as computing resources (CPU utilization, remaining memory), communication resources (signal-to-noise ratio, network latency), sensor status (radar model, data quality), and current task status. Node G uses this information to maintain a global resource database and provides a basis for decision-making regarding subsequent task allocation.

[0108] Subsequently, node T receives subtasks dynamically assigned by node G. These subtasks are typically strongly correlated with the spatial location of node T and its sensor field of view, such as obstacle detection in a specific area, lane line recognition, or local trajectory prediction. Node T utilizes local sensors (such as LiDAR, millimeter-wave radar, and cameras) to acquire information about the surrounding environment, performs computational processing locally, and generates local intelligent driving conclusions. These conclusions include:

[0109] Perceptual conclusions: Identify information such as the type, location, and confidence level of the target;

[0110] Prediction conclusion: Predict the trajectory of the target or the time of collision with this vehicle;

[0111] Local decision recommendations: Preliminary control recommendations (such as braking, steering) or coordination recommendations for the formation.

[0112] Instead of uploading raw perception data, the T-node uploads the aforementioned local intelligent driving conclusions to the G-node, thus significantly reducing communication bandwidth consumption. After receiving the conclusions from all T-nodes, the G-node performs fusion processing, generates global control commands, and issues them.

[0113] Finally, node T receives and executes global control commands from node G, such as adjusting vehicle speed, steering angle, or maintaining platoon spacing. These commands are generated based on globally optimal decisions, ensuring the safety and coordination of the entire platoon.

[0114] This embodiment achieves distributed processing of computing tasks through local computing and collaborative uploading of T nodes, effectively alleviating the computing and communication bottlenecks of centralized architecture, while making full use of the computing resources of each vehicle in the formation, thus improving the overall efficiency and reliability of the system.

[0115] In some embodiments, a large-scale vehicle platooning control system is provided, comprising:

[0116] A global management node G is deployed on any vehicle in the platoon to execute the large-scale vehicle platooning control method applied to the G node provided in the above embodiments;

[0117] Multiple terminal execution nodes T nodes, each T node corresponding to the remaining vehicles in the formation, and each T node is used for the large-scale vehicle formation control method applied to T nodes provided in the above embodiments;

[0118] The G node and T node exchange information via a star-flash wireless communication network.

[0119] In some embodiments, the distributed system resource status database includes the following dimensions of information:

[0120] The node identifier dimension includes the unique identifier and geographical location of node T;

[0121] The computing resource dimension includes processor model, number of cores, CPU utilization, and remaining memory;

[0122] The communication resource dimension includes the signal-to-noise ratio and network latency of the T-node and G-node links;

[0123] Sensor status dimension, including the model and data quality of lidar and millimeter-wave radar;

[0124] The task status dimension includes the number of subtasks currently being executed, the average completion time of historical tasks, and the average completion quality.

[0125] In some embodiments, the system also supports a hierarchical architecture: the G node allocates some T nodes as sub-G nodes, each sub-G node is responsible for managing the sub-task allocation and local conclusion fusion within its jurisdiction, and uploads the fusion results to the main G node.

[0126] In some embodiments, a large-scale vehicle platooning control system is provided, comprising:

[0127] This embodiment describes the overall implementation scheme of distributed cooperative formation control from the perspective of system architecture. The system consists of two types of logical nodes: a global management node (G node) and multiple terminal execution nodes (T nodes). The nodes communicate with each other through a StarBlink wireless communication network for high-speed, low-latency, and highly reliable information exchange.

[0128] The G node is deployed on any vehicle in the platoon, typically a vehicle with strong computing power, complete sensor configuration, and stable communication links. The G node executes the large-scale vehicle platoon control method applied to the G node in the aforementioned embodiments. Its core responsibilities include: establishing connections with all T nodes and maintaining a distributed system resource status database containing resource status information of each node; decomposing the computational tasks of cooperative driving into multiple sub-tasks that can be processed in parallel, and dynamically allocating them to appropriate T nodes based on resource status information; receiving local intelligent driving conclusions uploaded by each T node, performing fusion processing to generate a global high-precision dynamic map; confirming the authenticity and risk level of targets through conflict detection and weighted voting mechanisms, and finally generating global control commands and issuing them to each T node.

[0129] T-nodes correspond to the remaining vehicles in the platoon, with each vehicle acting as an independent intelligent computing unit. Each T-node executes the large-scale vehicle platooning control method applied to T-nodes in the aforementioned embodiments. Its core responsibilities include: sending heartbeat data packets to G-nodes at high frequency to update its own resource status information; receiving sub-tasks assigned by G-nodes, acquiring surrounding environmental information using local sensors, completing computational processing locally, generating local intelligent driving conclusions (including perception conclusions, prediction conclusions, and local decision suggestions), and uploading them to G-nodes; and receiving and executing global control commands issued by G-nodes, such as adjusting vehicle speed, steering angle, or maintaining platoon spacing.

[0130] During system operation, G nodes and T nodes form a tight collaborative closed loop: T nodes report resource status in real time via heartbeat packets, providing decision-making basis for G nodes' task scheduling; G nodes distribute computing tasks to various T nodes for parallel processing, and generate globally optimal decisions by aggregating local conclusions; T nodes, as components of the distributed computing resource pool, both execute local computing tasks and receive and execute global instructions. This architectural design realizes a paradigm shift from "centralized computing" to "collaborative computing."

[0131] This embodiment utilizes the division of labor and cooperation between G nodes and T nodes to construct the entire vehicle platoon into a dynamic and scalable distributed computing platform. The system features the following characteristics: the computational load is distributed across all vehicles, breaking through the single-node computing power limit; only intelligent driving conclusions are uploaded, rather than raw data, significantly reducing communication bandwidth consumption; dynamic task scheduling and offline node detection mechanisms eliminate the risk of single-point failures; and full utilization of each vehicle's computing resources improves the overall system efficiency.

[0132] In some embodiments, the solutions of this application are described below in conjunction with specific scenarios.

[0133] I. Scenario and System Initialization.

[0134] On a highway, a logistics convoy consisting of 1,000 heavy trucks is traveling at night.

[0135] Formation objectives: Maintain minimum safe distance, achieve maximum fuel efficiency, and respond to unexpected traffic incidents.

[0136] System initialization and node allocation:

[0137] Trigger: The formation begins to be formed, and all vehicles broadcast their own status (number of computing cores, sensor configuration, and location) through the StarFlash 3.0 network.

[0138] Election: A vehicle located at the front of the formation, equipped with a multi-core heterogeneous processor and complete sensors, is elected as the G node due to its strong computing power, good location, and stable communication link.

[0139] Network topology: The G node broadcasts its leadership status through the StarNet network and allocates communication time slot resources to all T nodes (e.g., 0.5ms for the emergency channel, 2ms for the control channel, and 5ms for the data channel). A global resource table is established within the G node and begins to be dynamically updated.

[0140] II. Standardized Collaborative Driving Process.

[0141] The sensors of T-node-500 (located in the middle of the formation) detected an abandoned tire (target #Tire-01) 120 meters ahead of it.

[0142] Its local model completes recognition within 15μs and generates a local intelligent driving conclusion package:

[0143] Perception conclusion: Target #Tire-01, type: static obstacle, confidence level: 99.8%, located in the center of this lane.

[0144] Prediction: This vehicle will collide in 4.2 seconds.

[0145] Local decision-making recommendations: It is recommended that this vehicle initiate emergency braking, and that the entire formation make a slight adjustment to the left to avoid the obstacle.

[0146] This conclusion packet, along with the vehicle status, is uploaded to node G via a heartbeat packet.

[0147] Conclusion: Node G received reports about #Tire-01 from multiple nodes, including T-500, T-499, and T-501, almost simultaneously.

[0148] Conflict Detection and Fusion:

[0149] The conclusions of node G regarding all nodes are highly consistent ("static obstacle, high confidence").

[0150] Through weighted voting (taking into account the distance of nodes from the target and sensor accuracy), #Tire-01 was quickly identified as a real, high-risk global threat.

[0151] Global decision-making:

[0152] Scenario judgment: This is a "sudden obstacle" scenario, triggering the collaborative obstacle avoidance decision-making process.

[0153] Authoritative calculation: The trajectory prediction of node G calculates two alternative global paths within 20μs: a) overall braking; b) overall smooth detour to the left.

[0154] Multi-objective optimization: Command generation > Based on the principle of safety > comfort > efficiency, select scheme b within 10μs, because the overall detour has less impact on formation stability and fuel economy.

[0155] Generate instructions: Node G generates the final intelligent driving conclusion instructions and embeds them with a timestamp accurate to the corresponding synchronization time.

[0156] The technical solution of this invention brings at least the following beneficial effects:

[0157] Safety: From perception to the generation of global collaborative obstacle avoidance commands, the entire process has a latency of <50μs, which is hundreds of times faster than human drivers or traditional autonomous driving systems.

[0158] High efficiency: When 1,000 vehicles move as a whole, the traffic efficiency at the intersection is increased by more than 50%, and energy savings of 15-20% are achieved due to reduced wind resistance.

[0159] Reliability: Through distributed sensing and decision-making, the system is immune to failures of single or multiple nodes, and its MTBF (Mean Time Between Failures) far exceeds 10,000 hours.

[0160] Economy: There is no need to equip each vehicle with a top-of-the-line computing chip. By pooling resources, ultra-high performance can be achieved using a combination of medium-power computing hardware, which greatly reduces the cost of upgrading a single vehicle to be intelligent.

[0161] In the embodiments of this application, the transmission of vehicle-related information needs to be precise and confidential. The signal must be transmitted directionally without affecting distant vehicles, preventing interference with the normal communication of other vehicles. Data from different T nodes is rapidly transmitted to the G node, which requires significant computing power to receive the rapidly transmitted data from the T nodes. In other words, in the embodiments of this application, the vehicle platooning system has high requirements for the directionality and anti-interference capabilities of communication. Specifically, data transmission between the G node and the T node is conducted via star-flash wireless communication technology. This technology supports beamforming and directional transmission mechanisms, concentrating signal energy in the direction of the target receiving node, thereby reducing signal leakage to non-target vehicles, avoiding interference with the normal communication of vehicles outside the platoon or at greater distances, and improving the spatial reuse efficiency of spectrum resources.

[0162] Meanwhile, due to the large convoy size (reaching hundreds or even thousands of vehicles), each T-node sends heartbeat data packets and intelligent driving conclusions to the G-node at high frequency. The G-node needs to receive and process a large amount of concurrent data in a very short time. To this end, the G-node is equipped with a high-performance multi-core heterogeneous processor and adopts a data parallel processing and priority scheduling mechanism, which can simultaneously parse, verify, and update the data streams of multiple T-nodes, ensuring that real-time requirements are still met in high-concurrency scenarios.

[0163] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0164] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0165] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0166] The computer program instructions used to perform the operations of this invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions. This electronic circuitry can execute the computer-readable program instructions to implement various aspects of the invention.

[0167] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0168] Various aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0169] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0170] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0171] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0172] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for illustrative purposes only and should be construed as such, and is not intended to be limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of the invention as set forth in the appended claims.

Claims

1. A method for large-scale vehicle formation control, characterized in that, Applied to a system comprising a global management node G and multiple terminal execution nodes T, the method is executed by the G node and includes: Establish a star-flash wireless communication connection with multiple T nodes, obtain the resource status information of each T node, and maintain a distributed system resource status database; wherein, the distributed system resource status database is composed of multiple T nodes and is used to record various resource status information of each T node in real time. The computational task of cooperative driving is decomposed into multiple subtasks that can be processed in parallel, and the subtasks are dynamically allocated to the corresponding T nodes based on the resource status information. It receives local intelligent driving conclusions uploaded by each T node, performs fusion processing on all received local intelligent driving conclusions, generates global control commands, and sends them to each T node for execution.

2. The method according to claim 1, characterized in that, The process of obtaining the resource status information of each T node includes: The G node receives heartbeat data packets sent by each T node at a high frequency. The heartbeat data packets contain update information on node identifier, computing resource dimension, and task status dimension. The G node updates the information of the communication resource dimension according to the reception time and data quality of the heartbeat data packet; The G node detects the online status of the T node based on the reception of the heartbeat data packet. If it does not receive a heartbeat data packet from a certain T node within a set time, it marks the T node as offline and stops assigning tasks to it.

3. The method according to claim 1 or 2, characterized in that, The fusion process includes: The G node integrates the local intelligent driving information uploaded by each T node into a global high-precision dynamic map; The G node performs conflict detection on the conclusions of different T nodes regarding the same target, and confirms the authenticity and risk level of the target through a weighted voting mechanism; The G node generates global control commands based on the fusion results.

4. The method according to claim 3, characterized in that, Also includes: When the G node detects a conflict between the local intelligent driving conclusion of the T node and the conclusion calculated by the G node itself, or receives an emergency command from the T node, the G node starts the high-speed computing module to re-analyze. The G node updates the conclusion weight and credibility of the corresponding T node based on the results of the reanalysis.

5. The method according to claim 1, characterized in that, The method further includes: The G node assigns some T nodes as child G nodes. The child G nodes are responsible for the allocation of sub-tasks and the fusion of local conclusions within their management scope, and then upload the fusion results to the main G node.

6. A method for large-scale vehicle formation control, characterized in that, Applied to a system comprising a global management node G and multiple terminal execution nodes T, the method is executed by any of the T nodes and includes: Establish a StarScan wireless communication connection with the global management node G, and send a heartbeat data packet to the G node. The heartbeat data packet contains the resource status information of the T node. The system receives subtasks assigned by the G node, acquires surrounding environmental information through local sensors, processes the assigned subtasks locally, generates local intelligent driving conclusions, and uploads the local intelligent driving conclusions to the G node; wherein, the local intelligent driving conclusions include perception conclusions, prediction conclusions, and local decision suggestions. Receive and execute global control commands issued by the G node.

7. The method according to claim 6, characterized in that, The generated local intelligent driving conclusions include: The T-node identifies the target and generates a perception conclusion that includes the target type, location, and confidence level. The T node generates a prediction conclusion that includes a collision time prediction; The T-node generates local decision-making suggestions for the vehicle and its formation.

8. A large-scale vehicle platooning control system, characterized in that, include: A global management node G, deployed on any vehicle in the platoon, is used to execute the method of any one of claims 1 to 5; Multiple terminal execution nodes T nodes, each T node corresponding to the remaining vehicles in the formation, each T node being used to execute the method of any one of claims 6 to 7; The G node and T node exchange information via a star-flash wireless communication network.

9. The system according to claim 8, characterized in that, The distributed system resource status database includes the following dimensions of information: The node identifier dimension includes the unique identifier and geographical location of node T; The computing resource dimension includes processor model, number of cores, CPU utilization, and remaining memory; The communication resource dimension includes the signal-to-noise ratio and network latency of the T-node and G-node links; Sensor status dimension, including the model and data quality of lidar and millimeter-wave radar; The task status dimension includes the number of subtasks currently being executed, the average completion time of historical tasks, and the average completion quality.

10. The system according to claim 8, characterized in that, The system also supports a layered architecture: The G node will allocate some T nodes as child G nodes. Each child G node is responsible for managing the allocation of subtasks and the fusion of local conclusions within its jurisdiction, and will upload the fusion results to the main G node.

Images