Multi-pilot vehicle formation control method based on formation graph connectivity judgment

Abstract

The invention provides a multi-pilot vehicle formation control method based on formation graph connectivity judgment, and the method comprises the steps: determining the motion state of each target vehicle in a vehicle formation when the formation operation mode of the vehicle formation is a single-pilot mode; if it is determined that a target vehicle in the vehicle formation is abnormal according to the motion state, splitting the vehicle formation into a plurality of sub-formations in a multi-navigation mode; and when each sub-formation satisfies a preset formation fusion condition, recombining the sub-formations into a vehicle formation in the single navigation mode. The method is used for overcoming the defect that in the prior art, the whole formation loses efficacy when the vehicles cannot continue to follow stably in the vehicle formation in a single-pilot mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent transportation systems and automatic driving technology, and particularly relates to a multi-leader vehicle platoon control method based on platoon graph connectivity determination. BACKGROUND

[0002] Under the background of rapid development of intelligent networking and automatic driving technology, vehicle platoon is widely used in typical scenarios such as highway driving, automatic logistics vehicle scheduling, emergency escort, etc. Vehicle platoon keeps multiple vehicles in orderly, stable and efficient cooperative driving through unified control strategy, thereby improving road capacity and reducing energy consumption.

[0003] In the prior art, most platoon control methods adopt single-leader mode, that is, one master vehicle guides the whole platoon, and other vehicles follow the trajectory of the master vehicle. This centralized control architecture performs well in ideal traffic environment, but faces various challenges in actual road scenarios. Especially in complex traffic conditions (such as multi-lane lane changing, temporary queue insertion, road congestion) or communication and perception interruption, some vehicles may not be able to continue to receive leader information and maintain stable following, resulting in platoon structure rupture, control instability, and even platoon failure. SUMMARY

[0004] The present application provides a multi-leader vehicle platoon control method based on platoon graph connectivity determination to overcome the defect that the vehicle platoon under single-leader mode in the prior art fails as a whole when some vehicles cannot continue to follow stably.

[0005] The present application provides a multi-leader vehicle platoon control method based on platoon graph connectivity determination, comprising the following steps: When the platoon operation mode of the vehicle platoon is single-leader mode, the motion state of each target vehicle in the vehicle platoon is determined; If it is determined according to the motion state that there is an abnormal target vehicle in the vehicle platoon, the vehicle platoon is split into multiple sub-platoons in multi-leader mode; When each sub-platoon meets the preset platoon fusion condition, the sub-platoons are reorganized into the vehicle platoon in single-leader mode.

[0006] In some embodiments, the determination of the motion state of each target vehicle in the vehicle platoon comprises: each target vehicle in the vehicle platoon as a node, and a following relationship between the target vehicles as a directed edge, wherein a node attribute of the node includes a vehicle state and a platoon state, the vehicle state includes a vehicle position, a motion speed, a heading angle, a control input value, and a lane in which the vehicle is located, and the platoon state includes a vehicle role, a vehicle control mode, and a vehicle driver state, and an edge attribute of the directed edge includes an inter-vehicle communication quality, an inter-vehicle perception reachability, and an inter-vehicle relative distance; construct a directed dynamic graph corresponding to the vehicle platoon based on the nodes and the directed edges; filter a control subgraph from the directed dynamic graph according to the edge attribute; obtain a motion state of each target vehicle in the vehicle platoon according to the nodes and the directed edges in the control subgraph when the vehicle platoon is moving.

[0007] In some embodiments, the determining, according to the motion state, that there is an abnormal target vehicle in the vehicle platoon includes: determining, according to the motion state, a communication state of each target vehicle with other vehicles, an inter-vehicle distance, and a driver state of each target vehicle; determining that there is an abnormal target vehicle in the vehicle platoon when any of the following conditions is met: the communication state is a communication link interruption; the inter-vehicle distance is greater than a preset vehicle distance threshold; the driver state indicates that a driving duration exceeds a safety threshold or a driver requests to give up driving.

[0008] In some embodiments, when the vehicle role of the abnormal target vehicle is a lead vehicle in a single-lead mode, and the type of the abnormality is a passive type, the splitting the vehicle platoon into multiple sub-platoons in a multiple-lead mode includes: dividing the vehicle platoon into multiple sub-platoons; for each target vehicle in the sub-platoon, counting a number of following vehicles corresponding to the target vehicle, and determining a target vehicle with a maximum number of following vehicles as a candidate vehicle; when the number of candidate vehicles is 1, determining the candidate vehicle as a lead vehicle of the sub-platoon, and broadcasting the lead vehicle to the sub-platoon; when the number of candidate vehicles is greater than 1, selecting a candidate vehicle with a highest state index from candidate vehicles allowed to lead as the lead vehicle of the sub-platoon, and broadcasting the lead vehicle to the sub-platoon, wherein the state index is positively correlated with a communication quality index and a control stability index.

[0009] In some embodiments, when the vehicle role of the abnormal target vehicle is a leading vehicle in a single-leader mode, and the abnormality type is an active type, the splitting the vehicle platoon into a plurality of sub-platoons in a multiple-leader mode comprises: dividing the vehicle platoon into a plurality of sub-platoons, and broadcasting a de-weight signal of the abnormal target vehicle in the sub-platoons, the de-weight signal being used to represent abandoning the vehicle role of the leading vehicle; after each of the sub-platoons receives the de-weight signal, determining a communication quality indicator, a control stability indicator, and a leading feasibility indicator of each target vehicle in the sub-platoon; respectively weighting the communication quality indicator, the control stability indicator, and the leading feasibility indicator, and summing the weighted results to obtain a comprehensive indicator; taking the target vehicle with the highest comprehensive indicator as the leading vehicle in the sub-platoon, and broadcasting the leading vehicle to the sub-platoon.

[0010] In some embodiments, the preset platoon fusion condition comprises: a real-time relative distance between the target vehicles in each sub-platoon is less than a preset distance threshold; the driving directions of the target vehicles in each sub-platoon are consistent; the abnormal target vehicle has re-established a communication link.

[0011] In some embodiments, the recombining the sub-platoon into the vehicle platoon in the single-leader mode comprises: determining a state stability duration and a vehicle formation error of the recombined vehicle platoon; when the state stability duration exceeds a preset time threshold, the vehicle formation error is less than a preset error range, and the leading vehicle in the single-leader mode has been determined, determining that the recombined vehicle platoon has been converted from a multiple-leader mode into a vehicle platoon in a single-leader mode; updating the nodes and directed edges in the control sub-graph according to the recombined vehicle platoon.

[0012] In some embodiments, when the platoon operation mode of the vehicle platoon is a single-leader mode, the method further comprises: determining a following reference vehicle for each target vehicle in the vehicle platoon; for each target vehicle, obtaining an expected offset between the target vehicle and the corresponding following reference vehicle, and constructing an expected trajectory of the target vehicle based on the expected offset; The preset state prediction model is called to optimize the expected trajectory, generate a motion control instruction of the target vehicle, and the motion control instruction is sent to the corresponding target vehicle to control the motion of the target vehicle.

[0013] In some embodiments, the calling of the preset state prediction model to optimize the expected trajectory, generate the motion control instruction of the target vehicle, comprises: determining a vehicle state of each target vehicle and a following vehicle state of a corresponding following reference vehicle; determining a difference between a target value of the vehicle state and a reference value of the following vehicle state, and constructing a target function according to a target control input value included in the target value and the difference; determining a constraint condition of the target function, the constraint condition comprising: the target value is within a preset threshold range, the target control input value is within a preset control threshold range, an actual relative distance between the target vehicle and a preceding vehicle is greater than a preset safety distance, and a fluctuation value of the target control input value is less than a preset fluctuation threshold value; under the constraint condition, the preset state prediction model is called to optimize the target function to obtain the motion control instruction of the target vehicle.

[0014] The application also provides a multi-pilot vehicle platoon control system based on platoon graph connectivity judgment, and the control system comprises the following modules: A determination module is configured to determine a motion state of each target vehicle in a vehicle platoon when a platoon operation mode of the vehicle platoon is a single-pilot mode; A switching module is configured to split the vehicle platoon into a plurality of sub-platoons in a multi-pilot mode if it is determined according to the motion state that there is a target vehicle in the vehicle platoon that is abnormal; A recovery module is configured to recombine the sub-platoons into the vehicle platoon in the single-pilot mode when each sub-platoon satisfies a preset platoon fusion condition.

[0015] The application also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the multi-pilot vehicle platoon control method based on platoon graph connectivity judgment according to any one of the above embodiments when executing the computer program.

[0016] The application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and the computer program is executable on a processor to implement the multi-pilot vehicle platoon control method based on platoon graph connectivity judgment according to any one of the above embodiments.

[0017] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the multi-leader vehicle formation control method based on formation graph connectivity judgment as described above.

[0018] The multi-leader vehicle platooning control method based on platooning graph connectivity judgment provided by this invention, when a vehicle platoon is in single-leader mode and a target vehicle malfunctions, changes the platooning operation mode from single-leader to multi-leader mode, thereby splitting the vehicle platoon into multiple sub-platoons for movement, thus avoiding platooning structure breakage due to malfunctions. When the sub-platoons meet preset platooning merging conditions, the platoons are reorganized to restore the vehicle platooning in single-leader mode. This ensures the control stability of the vehicle platooning under complex traffic conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced one by one below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the multi-leader vehicle formation control method based on formation graph connectivity judgment provided by the present invention.

[0021] Figure 2 This is a schematic diagram illustrating the principle of the multi-leader vehicle formation control method based on formation graph connectivity judgment provided by the present invention.

[0022] Figure 3 This is one of the schematic diagrams of an abnormal scenario of vehicle formation structure fracture provided by the present invention.

[0023] Figure 4 This is the second schematic diagram of an abnormal scenario of vehicle formation structure fracture provided by the present invention.

[0024] Figure 5 This is a schematic diagram of the layered structure of the directed dynamic graph provided by the present invention.

[0025] Figure 6 This is a schematic diagram illustrating the principle of the formation operation mode switching provided by the present invention.

[0026] Figure 7 This is a schematic diagram of the structure of the multi-leader vehicle formation control system based on formation diagram connectivity judgment provided by the present invention.

[0027] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0029] The multi-leader vehicle formation control method based on formation graph connectivity judgment of the present invention is described below with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating the multi-leader vehicle formation control method based on formation graph connectivity judgment provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps 101 to 103, which are described in detail below.

[0030] Step 101: When the vehicle formation operation mode is single navigator mode, determine the motion state of each target vehicle in the vehicle formation.

[0031] During normal vehicle platooning operations, a single-leader mode is used. This means that one target vehicle in the platoon acts as the lead vehicle, and the remaining target vehicles follow its trajectory. Therefore, platooning control should not only monitor the lead vehicle's movement status but also determine the movement status of each target vehicle within the platoon. Vehicle movement status is divided into vehicle status and platoon status. Vehicle status includes vehicle position, speed, heading angle, control input values, and lane location. Placing status includes vehicle role, vehicle control mode, driver status, inter-vehicle communication quality, inter-vehicle perception accessibility, and relative distance between vehicles. Specifically, vehicle role indicates whether it is the lead vehicle; vehicle control mode indicates whether it is in manual or automatic driving mode; driver status indicates the driver's fatigue level, concentration level, and physiological load; inter-vehicle communication quality indicates whether communication with other vehicles is good or bad; inter-vehicle perception accessibility indicates whether the status information of other vehicles, such as position and obstruction, can be obtained via radar and cameras; and relative distance between vehicles indicates the distance between them.

[0032] Since there are many parameters related to the motion state of each target vehicle in a vehicle formation, it is impossible to fully control and obtain them through conventional technical means. Therefore, this embodiment of the invention uses a directed dynamic graph to model the system structure of the moving vehicle formation in order to depict the information interaction structure of the vehicle formation as it changes over time.

[0033] In some embodiments, the motion state of each target vehicle in a vehicle platoon can be determined in the following ways, which are described in detail below.

[0034] When constructing a directed dynamic graph of vehicle formation, each target vehicle in the vehicle formation is treated as a node, and the following relationship between the target vehicles is treated as a directed edge.

[0035] Here is as Figure 5 As shown, a directed dynamic graph is denoted as , Each target vehicle is denoted as i, and the set of vehicles is a set of nodes, i.e. Edges are used to represent the following relationship between target vehicles, i.e., "who follows whom". Therefore, directed edges are used here, and the direction of the edge describes the direction of following. The set of directed edges is the [blank]. Specifically, if there exists a directed edge (j, i) ∈ This indicates that target vehicle i follows target vehicle j in the current vehicle formation, using target vehicle j as the direct reference formation, receiving state information and adjusting its own control behavior accordingly. Furthermore, for directed edges, this embodiment of the invention uses an adjacency matrix... Set of opposite edges A formal description is provided. Among them... This indicates that target vehicle i follows target vehicle j at time t; otherwise... This indicates that there is no following relationship between the two. Due to the directionality of information flow, the adjacency matrix is ​​typically an asymmetric matrix. The adjacency matrix not only reflects the following topology between vehicles but also directly determines the information-dependent paths in the control algorithm. Its sparsity reflects the local interaction characteristics of the system, while its dynamic changes reflect the structural update process of the topology when events such as edge breakage, fusion, and role switching occur. Furthermore, this adjacency structure is used in conjunction with the vehicle role of target vehicle i, and can be used to identify the local connectivity states of the lead vehicle (whose corresponding incoming edge is 0), the last vehicle (whose corresponding outgoing edge is 0), and candidate vehicles.

[0036] See also Figure 5 In a directed dynamic graph, each node also has node attributes, including vehicle status. and formation status Among them, vehicle status Including vehicle location Speed ​​of movement Heading angle , control input value and the lane Wherein, vehicle position represents the two-dimensional spatial coordinates of the target vehicle, motion speed represents the longitudinal speed of the target vehicle, and heading angle is taken as... This represents the angle between the target vehicle and the reference direction, used to characterize the current direction of travel of the target vehicle. Control input value This includes acceleration, steering wheel angle, or a combination of both, used to propel the target vehicle. The lane in which it is located. This indicates the lane number currently occupied by the target vehicle. Therefore, the vehicle state of target vehicle i is determined. , means as follows (1) The formation status of the target vehicles Including vehicle roles Vehicle control mode and the vehicle driver's status Vehicle Role Vehicle Role This indicates the functional division of vehicles within the current vehicle formation. The lead vehicle is responsible for path planning and control target setting, while other following vehicles perform follow control according to a distributed algorithm. Lead vehicle candidates represent target vehicles currently in the candidate lead vehicle state, and can participate in role switching or election processes as needed. Vehicle roles can dynamically evolve based on factors such as communication status, driver status, and control status. Vehicle Control Mode The value is {0, 1}, used to distinguish the current control mode of the target vehicle. Indicates manual driving. This indicates the autonomous driving mode. This parameter plays a crucial role in state management and control strategy execution. For example, in vehicle platooning reorganization, autonomous vehicles prioritize control coordination, while manually driven vehicles are treated as passive units. (Vehicle driver state) This parameter is used to quantify the physiological and cognitive state of the target vehicle's driver, such as fatigue level, concentration, and physiological load. It can be estimated using driving behavior data, physiological sensors, or manual input. In this embodiment of the invention, the driver's state is not only used to determine whether the current vehicle is suitable for continued manual driving, but also serves as an important constraint when selecting a potential lead vehicle, avoiding the selection of a vehicle driven by a driver in poor condition.

[0037] In addition, directed dynamic graphs are denoted as Each directed edge also has edge attributes, and the set of edge attributes is represented by the adjacency matrix between nodes. To represent this. The edge properties of directed edges include the quality of communication between vehicles. Vehicle-to-vehicle accessibility Relative distance between vehicles Communication quality This parameter represents the communication link status between target vehicle i and target vehicle j at time t, including indicators such as link strength, communication latency, and packet loss rate. This parameter plays a crucial role in determining whether a directed edge can be used for real-time control information transmission. (Inter-vehicle perception reachability) This refers to whether target vehicle i can directly obtain the status information of target vehicle j through local sensors (such as radar or cameras), and is usually related to environmental occlusion, relative position, and sensor field of view. Relative distance between vehicles. Used to represent Euclidean distance or relative position information between vehicles. This attribute can be used for edge activation determination, control weight allocation, or triggering specific events (such as fusion or collision avoidance).

[0038] The definition of edge attributes provides crucial decision-making criteria for vehicle platoon control. For example, in edge selection, directed edges can be filtered based on communication quality or perceived reachability, retaining only reliable links for state prediction and optimization. Furthermore, vehicle role credibility assessment can be performed; determining whether following vehicles remain reliable can be triggered by mechanisms such as perceived link interruption or distance exceeding limits. Additionally, platoon restructuring can be triggered: when critical edge attributes remain abnormal (e.g., prolonged communication interruption), operations such as switching platoon operation modes, platoon restructuring, or determining the lead vehicle can be initiated.

[0039] Therefore, based on nodes and directed edges, a directed dynamic graph corresponding to vehicle platooning is constructed. Based on the following relationships between each target vehicle, nodes are connected by directed edges. Then, node attributes and edge attributes are loaded into each node, thus constructing a directed dynamic graph. It can vividly and concretely depict the motion state of the vehicle formation at each moment t during the movement process.

[0040] like Figure 2 As shown, after constructing the directed dynamic graph, when controlling vehicle platooning, it is necessary to filter out the control subgraph from the directed dynamic graph based on edge attributes. The purpose of this filtering is to check whether there are any breaks in the platooning structure, and then select reliable edges from the platoon for control and optimization. Examples of broken platooning structures include... Figure 3 As shown, target vehicle 1 was the lead vehicle, but due to congestion in some lanes, it was overtaken by vehicle 4, which was acting as a follower vehicle, causing confusion in the following relationship between target vehicle 1 and target vehicle 4. For example... Figure 4 As shown, the intrusion of a large number of other vehicles caused the communication quality and perception reachability of target vehicle No. 1 and target vehicle No. 2 in the vehicle formation to be interrupted.

[0041] When it is determined that there is no structural break in the current vehicle formation, a control subgraph is selected from the directed dynamic graph. During vehicle formation movement, the motion state of each target vehicle in the formation is obtained based on the nodes and directed edges in the control subgraph. Here, nodes and directed edges are directly extracted from the control subgraph, and the corresponding node and edge attributes are obtained, including the vehicle state and formation state of each target vehicle. In other words, the motion state of each target vehicle in the formation is obtained for subsequent formation control.

[0042] This invention proposes a graph modeling approach, which models the real-time vehicle formation as a directed dynamic graph. By filtering and extracting attributes from the directed dynamic graph, the motion state of each target vehicle can be obtained in real time. Thus, through the dynamic topology of the directed dynamic graph, not only can the motion state information of the target vehicles be determined, but the topological evolution of the vehicle formation caused by communication interruption or environmental changes can also be vividly depicted, providing necessary information support and basis for the subsequent control of the vehicle formation.

[0043] Step 102: If it is determined from the motion status that there is an abnormality in the target vehicle in the vehicle formation, then the vehicle formation is split into multiple sub-formations in the multi-navigation mode.

[0044] Furthermore, during the real-time operation of the vehicle platoon, the directed dynamic graph can be monitored in real time, monitoring each node and directed edge, and updating node and edge attributes in real time. This allows for the identification of abnormal nodes, that is, determining whether a target vehicle in the vehicle platoon is malfunctioning. Here, determining whether a target vehicle in the vehicle platoon is malfunctioning is achieved by identifying whether there are any breaks in the platoon formation.

[0045] In actual vehicle platooning operations, the vehicle organization structure is not static but dynamically adjusted based on various factors such as communication status, driver status, and spatial relationships. Therefore, an operation mode management mechanism is needed to support efficient collaboration and safe operation of target vehicles under different operating states. To this end, this invention proposes an event-driven operation mode switching framework that supports three basic platooning operation modes: single-leader mode, multi-leader mode, and fusion mode. Switching between each platooning operation mode requires a corresponding trigger event; that is, the platooning operation mode is switched based on the trigger event.

[0046] Specifically, such as Figure 6 As shown, the triggering events include formation breakup event, fusion completion event, and fusion start event. In other words, the switching between single navigation mode, multi-navigation mode, and fusion mode is triggered by formation breakup event, fusion completion event, and fusion start event.

[0047] In some embodiments, determining that a target vehicle in a vehicle formation is abnormal based on its motion state can be achieved in the following ways.

[0048] First, based on the motion state, the communication status of each target vehicle with other vehicles, the distance between vehicles, and the driver status of each target vehicle are determined. For each target vehicle, its communication status with other vehicles can be determined through inter-vehicle perceived reachability. Vehicle spacing can be determined by the relative distance between vehicles. This can be determined by the driver status of each target vehicle, which can be determined through the vehicle driver status. To determine.

[0049] Further judgment is made based on the vehicle's movement status. An anomaly is determined when any one of the following conditions is met: communication link interruption, vehicle spacing exceeding a preset distance threshold, driver status indicating driving duration exceeding a safety threshold, or driver requesting relinquishment. These three conditions serve as the reference criteria for determining whether a platoon breakup event has been triggered. The running status of the target vehicle within the platoon is used to determine if any one of these three conditions is met, thus determining whether a platoon breakup event has been triggered. Specifically, a communication link interruption indicates impaired inter-vehicle perceptual reachability. This indicates that the target vehicles cannot reliably exchange state information, making coordinated control difficult. Furthermore, a vehicle spacing greater than a preset distance threshold indicates a significant distance between the vehicles. The following indicators suggest that the following distance is too large, exceeding the perception range or control delay tolerance. The driver status indicator indicates that the driving duration has exceeded a safety threshold or the driver has requested to relinquish control, thus reflecting the vehicle's driver status. This indicates a decline in driver function, such as when a driver's continuous driving control time exceeds a limit, or when the driver is subjectively uncomfortable or fatigued. The driver has applied for a demotion or initiated a demotion based on a fatigue assessment model.

[0050] Therefore, if the current vehicle formation meets any of the above three conditions, it can be determined that the current vehicle formation has triggered a formation breakage event, and at this time it is determined that there is an anomaly in the target vehicle within the vehicle formation. If it is determined from the motion state that there is an anomaly in the target vehicle within the vehicle formation, the vehicle formation is split into multiple sub-formations in multi-navigation mode. That is to say, when it is determined that the vehicle formation has triggered a formation breakage event, the formation operation mode needs to be switched from single-navigation mode to multi-navigation mode. In specific implementation, the vehicle formation is split into multiple sub-formations in multi-navigation mode.

[0051] When the target vehicle exhibiting the anomaly is a lead vehicle in single-leader mode, it indicates that the vehicle formation has been disrupted due to the lead vehicle's malfunction. In this case, it's necessary to determine whether the anomaly is passive or active. Active anomalies indicate that the lead vehicle's operational status is insufficient to continue its lead role, specifically including a continuous decline in communication quality, persistent fluctuations or expansion of control errors, or the driver voluntarily relinquishing control and exiting the formation for safety or personal reasons. Passive anomalies, on the other hand, indicate that the lead vehicle loses control due to unforeseen events, such as a communication link interruption, the distance between the lead vehicle and following vehicles exceeding the limit, loss of physical reachability, or detection of driver fatigue or vehicle malfunction.

[0052] Furthermore, the lead vehicle is re-elected based on the corresponding election method according to the different types of anomalies, thereby splitting the vehicle formation into multiple sub-formations in the multi-leader mode and switching the single-leader mode to the multi-leader mode.

[0053] In some embodiments, when the target vehicle that causes the anomaly is a lead vehicle in single lead mode and the anomaly type is passive, the vehicle formation can be split into multiple sub-formations in multi-lead mode. This can be achieved in the following ways, which are explained in detail below.

[0054] First, the vehicle convoy is divided into multiple sub-convoys. The division can be based on equal distribution, with any target vehicles that are not evenly distributed forming a separate sub-convoy. After the division is completed, a new lead vehicle is elected for each sub-convoy. To ensure the smooth operation of the convoy, the vehicle convoy transitions from a single lead vehicle to multiple lead vehicles during this process.

[0055] In passive mode, this will automatically trigger a re-election of the lead vehicle for each sub-formation. Specifically, for each target vehicle in the sub-formation, the number of following vehicles corresponding to the target vehicle is counted, and the target vehicle with the largest number of following vehicles is determined as the candidate vehicle.

[0056] In practical implementation, it can be derived from a directed dynamic graph. In the control subgraph of the selection process, all nodes with available states are identified for each sub-group, forming a candidate node set corresponding to the sub-group. Then, the node with the most incoming edges is selected from the candidate node set, which means the node followed by the most other vehicles. This indicates that it is in a local core position and can be used as a candidate vehicle.

[0057] There may be multiple candidate vehicles. When there is only one candidate vehicle, it is designated as the lead vehicle of the sub-formation, and the lead vehicle is broadcast to the sub-formation. Alternatively, if only one node with the most incoming edges is selected from the candidate node set, the target vehicle represented by this node can be designated as the lead vehicle of the sub-formation.

[0058] When the number of candidate vehicles is greater than 1, the candidate vehicle with the highest status index is selected from the candidate vehicles that are allowed to lead the way as the lead vehicle of the sub-formation and the lead vehicle is broadcast to the sub-formation. The status index is positively correlated with the communication quality index and the control stability index.

[0059] Here, if multiple nodes with the highest number of incoming edges are selected from the candidate node set, then it is necessary to first obtain the node attributes of each node, such as the driver status of the target vehicle. It can determine whether the target vehicle meets the standards for allowing navigation, such as whether the fatigue level exceeds the fatigue threshold, thereby filtering out nodes that are not allowed to navigate.

[0060] Then, for the selected nodes, the edge attributes of the directed edges containing those nodes are further obtained to calculate the node's state index. The state index is positively correlated with both the communication quality index and the control stability index; the higher the communication quality index or control stability index, the higher the state index. Edge attributes can be selected as inter-vehicle communication quality... and inter-vehicle perceived accessibility inter-vehicle communication quality Communication quality metrics can be used to measure this; for example, a good inter-vehicle communication quality score is 1, otherwise it's 0. Control stability metrics can be measured through inter-vehicle perceived reachability. To measure this, better inter-vehicle perception accessibility corresponds to a higher control stability index (valued at 1), otherwise 0. Then, weights are assigned based on control and communication importance, and the communication quality and control stability indices are weighted separately. The weighted results are then summed to obtain the state index. Finally, the node with the highest state index is selected, and the candidate vehicle corresponding to this node is chosen as the leader vehicle of the sub-formation.

[0061] Therefore, by repeating the lead vehicle election process described above, a lead vehicle can be selected for each sub-formation. This selection is then broadcast within each sub-formation to inform other vehicles that a new lead vehicle has been elected to take over control. Furthermore, after each sub-formation completes its lead vehicle election, the attributes of nodes and directed edges in the control subgraph are updated to reflect the motion state of each target vehicle. This allows the vehicle formation to transition from a single-leader mode to a multi-leader mode, ensuring smooth operation even in the event of anomalies.

[0062] In this embodiment of the invention, when the lead vehicle causes an anomaly in the vehicle formation due to objective passive reasons, an anomaly response mechanism is implemented. When the lead vehicle is abnormal, the formation is actively adjusted or the leading strategy is changed, the vehicle formation is immediately divided into multiple sub-formations, and a new lead vehicle is re-elected for each sub-formation. This allows the vehicle formation to complete the transition from a single leading mode to a multi-leading mode, achieving stable operation of the entire vehicle formation during anomalies and improving the flexibility of vehicle formation operation.

[0063] In some embodiments, when the vehicle role of the target vehicle that causes the anomaly is the lead vehicle in the single lead mode and the type of anomaly is active, the vehicle formation can be split into multiple sub-formations in the multi-lead mode, which can also be achieved in the following ways.

[0064] First, the vehicle formation is divided into multiple sub-formations, and the target vehicles that have anomalies are broadcast in the sub-formations as deweighted signals. The deweighted signals are used to identify the role of the vehicle that abandons the lead vehicle.

[0065] Here, when the anomaly of the lead vehicle is of the active type, the vehicle formation is still first divided into multiple sub-formations. The division can be based on equal distribution, and target vehicles that are not evenly divided are directly formed into a sub-formation. In single-leader mode, the lead vehicle can actively broadcast a "demotion signal" to each sub-formation, informing each sub-formation that it is about to relinquish its role as the lead vehicle.

[0066] After each sub-formation receives the downweighting signal, the communication quality index, control stability index, and navigation feasibility index of each target vehicle in the sub-formation are determined.

[0067] Once each sub-formation receives the "demotion signal" broadcast, it can then conduct an election for a lead vehicle within each sub-formation. The election is based on the comprehensive performance indicators of the target vehicle.

[0068] In practice, it still starts from a directed dynamic graph. In the selected control subgraph, all nodes with available states are identified for each sub-formation. Then, for each node in the sub-formation, communication quality indicators, control stability indicators, and navigation feasibility indicators are calculated based on node attributes and the attributes of incoming edges. The communication quality indicators are determined through inter-vehicle communication quality... Calculations show that, for example, if inter-vehicle communication quality is good, the communication quality index is 1; otherwise, it is 0. Control stability indices can be obtained through inter-vehicle perceived reachability. Calculations show that better inter-vehicle perceived reachability corresponds to a higher control stability index (valued at 1), otherwise it is set to 0. The navigation feasibility index can be determined through the driver status of node attributes. The calculation shows that the more fatigued the driver is, the lower the navigation feasibility index is. For example, if the driver's fatigue level exceeds the preset fatigue threshold, the navigation feasibility index is set to 0; otherwise, it is set to 1.

[0069] Next, the communication quality index, control stability index, and navigation feasibility index are weighted and calculated separately, and the weighted results are summed to obtain the comprehensive index. Here, weights can be set according to the importance of navigation feasibility, control stability, and communication, with navigation feasibility having the highest weight. The weights for control stability and communication can be set according to actual conditions; if equally important, they should be assigned the same weight. The comprehensive index is obtained by weighting the communication quality index, control stability index, and navigation feasibility index according to the set weights and summing the weighted results.

[0070] Finally, the target vehicle with the highest overall performance index is selected as the lead vehicle in the sub-formation, and the lead vehicle information is broadcast to the sub-formation. Here, from the nodes of each sub-formation in the control subgraph, the node with the highest overall performance index is selected, and the target vehicle corresponding to this node is selected as the lead vehicle of the sub-formation.

[0071] Therefore, by repeating the lead vehicle election process described above, a lead vehicle can be selected for each sub-formation. This selection is then broadcast within each sub-formation to inform other vehicles that a new lead vehicle has been elected to take over control. Furthermore, after each sub-formation completes its lead vehicle election, the attributes of nodes and directed edges in the control subgraph are updated to reflect the motion state of each target vehicle. This allows the vehicle formation to transition from a single-leader mode to a multi-leader mode, ensuring smooth operation even in the event of anomalies.

[0072] In this embodiment of the invention, when the lead vehicle abandons its lead role for reasons that cause an anomaly in the vehicle formation, an anomaly response mechanism is implemented. When the lead vehicle is in an anomaly, the formation is actively adjusted or the lead strategy is changed. The vehicle formation is immediately divided into multiple sub-formations, and a new lead vehicle is elected for each sub-formation. This allows the vehicle formation to transition from a single lead mode to a multi-lead mode, achieving stable operation of the entire vehicle formation in the event of an anomaly and improving the flexibility of vehicle formation operation.

[0073] Step 103: When each sub-formation meets the preset formation fusion conditions, the sub-formation is reorganized into a vehicle formation in single-navigation mode.

[0074] Through step 102 above, if an anomaly occurs in the target vehicle in the vehicle formation, and the vehicle formation is split into multiple sub-formations to achieve multi-navigation mode operation, it is necessary to monitor the operation status of multiple sub-formations in real time. When the sub-formations meet the preset formation fusion conditions, the sub-formations are reorganized into a vehicle formation in single-navigation mode.

[0075] like Figure 6 As shown, in multi-navigation mode, if each sub-formation meets the preset formation fusion conditions, a fusion start event can be triggered, reorganizing multiple sub-formations in multi-navigation mode back into vehicle formations under single-navigation mode. This process requires first transitioning from multi-navigation mode to fusion mode through the preset formation fusion conditions, and then finally achieving the conversion to single-navigation mode.

[0076] In some embodiments, the preset formation fusion conditions include the following three items: first, the real-time relative distance between target vehicles in each sub-formation is less than a preset distance threshold; second, the target vehicles in each sub-formation travel in the same direction; and third, the abnormal target vehicles have re-established the communication link.

[0077] Specifically, during the movement of multiple sub-formations, it is necessary to monitor whether the real-time relative distance between target vehicles in the sub-formations is less than a preset distance threshold. In the control subgraph of the directed dynamic graph, the node attributes of each node are acquired in real time, and the relative distance between vehicles in the node attributes is monitored. and make real-time judgments If the distance does not exceed a preset threshold, it indicates that the actual relative distance between the target vehicles is within a controllable range, meeting the platooning requirements. Secondly, during the movement of multiple sub-formations, it is necessary to monitor whether the driving direction of each target vehicle in the sub-formation is consistent. In the control subgraph of the directed dynamic graph, the node attributes of each node are acquired in real time, and the heading angle in the node attributes is monitored. Are they all the same? If they are the same, it means that the target vehicles are traveling in basically the same direction, avoiding the risk of collision during merging. Furthermore, during the movement of multiple sub-formations, it is necessary to monitor whether the target vehicles that have experienced anomalies have re-established their communication links. In the control subgraph of the directed dynamic graph, the edge attributes of the directed edges of the nodes experiencing anomalies are obtained in real time, and the inter-vehicle communication quality in the edge attributes is monitored. When the communication quality between vehicles on each directed edge All values ​​are greater than the preset communication quality threshold, indicating that the target vehicle that was malfunctioning has restored communication, re-established the communication link, and met the control requirements.

[0078] Therefore, when each sub-formation in the multi-navigation mode simultaneously meets the above three preset formation merging conditions, it indicates that each sub-formation has met the submission for merging and can trigger the merging start event. At this stage, each sub-formation begins to adjust its relative position and control strategy to prepare for the formation of a unified following structure, so as to reorganize the sub-formation into a vehicle formation in the single-navigation mode.

[0079] In this embodiment of the invention, under the multi-navigation mode of multiple sub-formations, by setting corresponding formation fusion conditions, the transition from the multi-navigation mode back to the single-navigation mode is realized under actual physical and information constraints, ensuring the feasibility and safety of vehicle formation fusion control.

[0080] In some embodiments, reorganizing a sub-formation into a vehicle formation in single-navigation mode can be achieved in the following ways, as detailed below.

[0081] First, determine the duration of stable state of the reorganized vehicle formation and the vehicle formation error.

[0082] like Figure 6 As shown, after triggering the fusion start event and entering fusion mode in multi-navigation mode, the lead vehicle in each sub-formation voluntarily relinquishes its lead vehicle role, relinquishes control, and enters fusion mode. Next, it is necessary to determine whether the vehicle formation has triggered the fusion completion event in fusion mode. The fusion completion event signifies whether the sub-formations have completed reorganization and reverted to a unified single-navigation structure. Triggering the fusion completion event is a stability confirmation process that must satisfy several structural and behavioral indicators, namely the duration of stable state of the vehicle formation and the vehicle formation error. The duration of stable state refers to the duration during which the vehicle states of each target vehicle, such as speed, acceleration, and control errors, remain stable without drastic fluctuations. The vehicle formation error specifically refers to the convergence degree of the vehicle formation error, including the relative position and speed errors between target vehicles.

[0083] When the stable state lasts for more than a preset time threshold, the vehicle formation error is less than a preset error range, and the lead vehicle in the single lead mode has been determined, it is determined that the reorganized vehicle formation has been converted from the multi-lead mode to the single lead mode.

[0084] Here, a judgment criterion is set for the indicators during the reorganization process to determine whether the fusion completion event is triggered. Three conditions are set: First, it needs to be determined whether the duration of state stability exceeds a preset time threshold; if so, it indicates that the vehicle formation has been stabilized. Second, it needs to be determined whether the vehicle formation error is less than a preset error range; if so, it indicates that the structural fluctuation of the vehicle formation is within a tolerable range, and the formation structure is stabilizing. Third, it needs to be determined whether the lead vehicle in single-leader mode has been determined. Because after the lead vehicle in each sub-formation voluntarily relinquishes its lead vehicle role and transfers control, a new lead vehicle needs to be elected in the reorganized vehicle formation to continue leading the vehicle formation in single-leader mode. The election method can be selecting the target vehicle with the highest state indicator or the target vehicle with the highest comprehensive indicator; it is not limited here. However, it is necessary to determine whether a lead vehicle has been elected after the vehicle formation is reorganized; only if it is confirmed that an election has been made can it be said that the sub-formation has completed reorganization.

[0085] like Figure 6 As shown, if the currently reorganized vehicle formation meets all three of the above conditions, it indicates that the fusion completion event has been triggered. At this time, the fusion mode naturally transitions to the single-navigation mode, which means that the reorganized vehicle formation has changed from the multi-navigation mode to the single-navigation mode.

[0086] Finally, the nodes and directed edges in the control subgraph are updated based on the reorganized vehicle formation. After the vehicle formation is reorganized, the nodes and directed edges in the control subgraph need to be updated. Specifically, the node attributes of each node and the edge attributes of the directed edges are updated to achieve dynamic updates of the directed dynamic graph, which will then be used for real-time monitoring of the vehicle formation. That is, the motion state of each target vehicle in the vehicle formation will continue to be determined in real time in single-navigation mode to determine whether a formation breakage event has been triggered.

[0087] Furthermore, in some embodiments, such as Figure 6 As shown, in fusion mode, if the currently reorganized vehicle formation does not meet the above three conditions—for example, the duration of stable state does not exceed a preset time threshold, the vehicle formation error is greater than a preset error range, and the lead vehicle in single-leader mode is not determined—then the fusion completion event can be considered not triggered, and the vehicle formation has not completed the transition to single-leader mode. This indicates that the vehicle formation is still in multi-leader mode. Therefore, each sub-formation continues to operate normally, and monitoring continues to determine whether the above three conditions are met to trigger the fusion completion event.

[0088] In this embodiment of the invention, by setting corresponding stability conditions to determine whether the vehicle formation has triggered a fusion completion event, and by using condition evaluation to realize sub-formations and automatic merging, the dynamic adjustment of the vehicle formation navigation structure is realized, so that the vehicle formation can return to unified control after the anomaly is recovered, ensuring a smooth switch of the vehicle formation from multi-navigation mode to single-navigation mode, and improving the coordination and operational efficiency of the vehicle formation.

[0089] In some embodiments, in complex scenarios where vehicle platooning faces dynamic changes in communication topology, inconsistent vehicle states, and continuous reconfiguration of platooning structure, traditional sequential car-following strategies are insufficient to meet the requirements for high reliability and high security control. In response, such as... Figure 2 As shown in the figure, this embodiment of the invention also provides a distributed model predictive control method. For each target vehicle, the desired trajectory is predicted by a state prediction model, and then motion control commands are generated to realize autonomous formation maintenance and smooth control of each target vehicle in the vehicle formation. The following is a detailed description.

[0090] When the vehicle formation is in single-navigation mode, the reference vehicle to follow for each target vehicle in the vehicle formation is first determined.

[0091] Considering that vehicle platooning typically operates in a single-navigation mode, this single-navigation mode can be either the original single-navigation mode or a single-navigation mode after switching from multiple navigation modes. When the vehicle platooning is in single-navigation mode, the following reference vehicle for each target vehicle in the platoon is determined. Specifically, from the control subgraph of the directed dynamic graph, for each node, the directly connected neighbor nodes of each node can be identified. These neighbor nodes may be the vehicle in front of the target vehicle, the vehicle to its side, or the leading vehicle.

[0092] There may be multiple neighboring nodes, so filtering is needed to identify the reference vehicle to follow. First, filter out neighboring nodes with valid communication or sensing connections—that is, neighboring nodes with high communication quality and inter-vehicle sensing reachability. Then, further, the neighboring node with the highest control edge weight—that is, the neighboring node with the highest communication quality and inter-vehicle sensing reachability—can be directly selected as the reference vehicle to follow, typically the vehicle ahead of the target vehicle. Alternatively, filtering can be based on the position and speed of each neighboring node. By setting corresponding weights, the position and speed of each neighboring node are weighted and fused to obtain the position and speed of a virtual neighboring node; this virtual neighboring node is the reference vehicle to follow.

[0093] Next, for each target vehicle, the expected offset between the target vehicle and the corresponding following reference vehicle is obtained, and the expected trajectory of the target vehicle is constructed based on the expected offset.

[0094] To maintain formation stability, simply determining the position of the target vehicle and the reference vehicle based on real-time relative distance is insufficient. Therefore, in this embodiment of the invention, a desired relative geometric position is preset for each target vehicle in the formation. This allows the determination of the deviation between the target vehicle's current position and its relative geometric position, which is the desired offset. Furthermore, based on the target vehicle's current speed, the desired trajectory of the target vehicle is planned, which is the final motion state that the target vehicle needs to achieve.

[0095] Furthermore, a preset state prediction model is invoked to optimize the desired trajectory and generate motion control commands for the target vehicle. Here, a local model predictive controller (MPC) can be installed in each target vehicle. The MPC controller operates independently and has a set control cycle. The state prediction model is deployed in the MPC controller. In each control cycle, the MPC controller invokes the state prediction model to optimize the desired trajectory and generate motion control commands for the target vehicle to control its operation.

[0096] Here, the state prediction model is a discrete-time dynamic model for the target vehicle. When optimizing, the objective function of the model must first be determined. Here, the vehicle state of each target vehicle is first determined. And the following vehicle status of the corresponding reference vehicle. .

[0097] Then determine the target value for the vehicle status. With the following vehicle status The difference between the reference value and the target value. Included target control input values Construct an objective function with the difference, denoted as The formula is expressed as follows: (2) In the above formula (2), N represents the total number of control cycles for the target vehicle. This represents the target vehicle state value of the i-th target vehicle within the k-th control cycle. This represents the following reference vehicle state reference value for the i-th target vehicle in the k-th control cycle. In actual calculations... Take the vehicle position and speed of movement Simply perform the calculation. Similarly, this allows the objective function to take into account the position and velocity errors, while This is to ensure smooth input control and avoid sudden acceleration or violent movements.

[0098] The state prediction model is optimized by minimizing the objective function. However, during the optimization process, it is also necessary to determine the constraints on the objective function. These constraints include: the objective value... Within the preset threshold range, the target control input value Within the preset control threshold range, the actual relative distance between the target vehicle and the vehicle in front is greater than the preset safe distance, and the fluctuation value of the target control input value is less than the preset difference range.

[0099] Specifically, the target value Within a preset threshold range, this constitutes a physical constraint on the target vehicle. The vehicle's state must be within permissible limits, such as speed and heading angle, as expressed as: (3) in, and These represent the vehicle status of the target vehicle. The upper and lower limits of the target value, such as the upper and lower limits of speed and heading angle, form the range of values ​​that is the preset threshold range.

[0100] Target control input value Within the preset control threshold range, this also constitutes a physical constraint on the target vehicle. The control input values ​​of the target vehicle must be within the vehicle's permissible range, such as acceleration and direction of travel, as expressed as: (4) in, and These represent the control input values ​​of the target vehicle. The upper and lower limits, such as the upper and lower limits of acceleration and driving direction, form the range of values ​​that is the preset threshold range.

[0101] The actual relative distance between the target vehicle and the vehicle in front is greater than the preset safe distance. In other words, for each target vehicle, its actual relative distance to the vehicle in front is greater than the preset safe distance, thus ensuring that a certain safe distance is always maintained between vehicles. This is expressed as: (5) in, Let represent the target value of the vehicle state of the i-th target vehicle. Let represent the target value of the vehicle state of the j-th target vehicle, which is also the target value of the vehicle state of the vehicle ahead of the target vehicle. In actual calculation, and All vehicle positions are taken. , Just do the calculation. This is the preset safe distance.

[0102] And target control input value The fluctuation value is less than the preset fluctuation threshold. This is to control changes in the control input value and prevent sudden changes in control commands when inputting control commands to the target vehicle, which could cause shaking or discomfort. Here, the difference between the control input values ​​in adjacent control cycles is calculated as the target control input value. The fluctuation values ​​are represented as follows: (6) in, This represents the control input value of the i-th target vehicle in the k-th control cycle, while This represents the control input value of the i-th target vehicle in the (k-1)-th control cycle. This indicates the preset fluctuation threshold.

[0103] Finally, under the constraints, a preset state prediction model is invoked to optimize the objective function, yielding the motion control commands for the target vehicle. Here, the current motion state of each target vehicle is first acquired and used as the initial value for the state prediction model. This initial value is then input into the state prediction model, which, under the aforementioned constraints, predicts the initial value with the goal of minimizing the objective function. This process is iterative, continuing until the initial value becomes increasingly close to the desired trajectory, at which point convergence is achieved, and the iteration stops.

[0104] After the state prediction model is optimized, the output prediction result is the value after the initial value is iteratively updated. Then, in the MPC controller, the corresponding motion control command for the target vehicle is generated based on the updated value and the initial value, such as how to adjust the speed, acceleration, heading angle, and change lanes.

[0105] In this embodiment of the invention, the motion state of each target vehicle is planned according to a predetermined desired trajectory under reasonable preset constraints, thereby realizing distributed control of vehicle platooning and improving the autonomy of target vehicle driving control.

[0106] Finally, motion control commands are sent to the corresponding target vehicle to control its movement. After the motion control commands are generated, they can be directly transmitted to the target vehicle through the MPC controller. Upon receiving the motion control commands, the target vehicle generates new control input values ​​to control it to travel according to these new values, enabling it to achieve the desired trajectory as much as possible.

[0107] In this embodiment of the invention, a distributed control method is designed to achieve individual control of each target vehicle in a vehicle platoon. By constructing a distributed model predictive control algorithm, coordinated control and formation maintenance among vehicles under topology change conditions are achieved, thereby improving the robustness, autonomy and cooperation of vehicle driving control as a whole.

[0108] The following describes the multi-leader vehicle formation control system based on formation graph connectivity judgment provided by the present invention. The multi-leader vehicle formation control system based on formation graph connectivity judgment described below can be referred to in correspondence with the multi-leader vehicle formation control method based on formation graph connectivity judgment described above.

[0109] See Figure 7 , Figure 7 This is a schematic diagram of the multi-leader vehicle formation control system based on formation graph connectivity judgment provided by the present invention, as shown below. Figure 7 As shown, the multi-leader vehicle platooning control system based on platooning graph connectivity judgment specifically includes: a determination module 701, a switching module 702, and a recovery module 703. Specifically, the determination module 701 is used to determine the motion state of each target vehicle in the vehicle platoon when the platooning operation mode is single-leader mode; the switching module 702, if it is determined from the motion state that a target vehicle in the vehicle platoon is abnormal, splits the vehicle platoon into multiple sub-platoons in multi-leader mode; the recovery module 703, when each sub-platoon meets the preset platooning fusion conditions, reassembles the sub-platoon into the vehicle platoon in single-leader mode.

[0110] like Figure 7 As shown, in some embodiments, the multi-leader vehicle formation control system based on formation graph connectivity judgment further includes a control module 704. Specifically, the control module 704 is used to determine the following reference vehicle for each target vehicle in the vehicle formation when the formation operation mode is single-leader mode; for each target vehicle, obtain the expected offset between the target vehicle and the corresponding following reference vehicle, and construct the expected trajectory of the target vehicle based on the expected offset; call a preset state prediction model to optimize the expected trajectory, generate motion control commands for the target vehicle, and send the motion control commands to the corresponding target vehicle to control the movement of the target vehicle.

[0111] It should be noted that the beneficial effects of the multi-leader vehicle formation control system based on formation graph connectivity judgment here correspond to those of the multi-leader vehicle formation control method based on formation graph connectivity judgment mentioned above. Therefore, the beneficial effects of the multi-leader vehicle formation control system based on formation graph connectivity judgment will not be elaborated here.

[0112] Figure 8 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 8 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a multi-leader vehicle formation control method based on formation graph connectivity judgment. The method includes: when the formation operation mode of the vehicle formation is a single-leader mode, determining the motion state of each target vehicle in the vehicle formation; if it is determined from the motion state that a target vehicle in the vehicle formation is abnormal, then splitting the vehicle formation into multiple sub-formations in the multi-leader mode; when each sub-formation meets a preset formation fusion condition, reassembling the sub-formation into the vehicle formation in the single-leader mode.

[0113] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0114] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the multi-leader vehicle formation control method based on formation graph connectivity judgment provided by the above methods. The method includes: when the formation operation mode of the vehicle formation is a single-leader mode, determining the motion state of each target vehicle in the vehicle formation; if it is determined from the motion state that a target vehicle in the vehicle formation is abnormal, then splitting the vehicle formation into multiple sub-formations in the multi-leader mode; when each sub-formation meets a preset formation fusion condition, reorganizing the sub-formation into the vehicle formation in the single-leader mode.

[0115] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the multi-leader vehicle formation control method based on formation graph connectivity judgment provided by the above methods. The method includes: when the formation operation mode of the vehicle formation is a single-leader mode, determining the motion state of each target vehicle in the vehicle formation; if it is determined from the motion state that a target vehicle in the vehicle formation is abnormal, then splitting the vehicle formation into multiple sub-formations in the multi-leader mode; when each sub-formation meets a preset formation fusion condition, recombining the sub-formation into the vehicle formation in the single-leader mode.

[0116] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-leader vehicle formation control method based on formation graph connectivity judgment, characterized in that, include: When the vehicle formation operation mode is single navigator mode, the motion state of each target vehicle in the vehicle formation is determined; If it is determined from the motion state that a target vehicle in the vehicle formation is abnormal, then the vehicle formation is split into multiple sub-formations in the multi-navigation mode; When each of the sub-formations meets the preset formation fusion conditions, the sub-formations are reorganized into a vehicle formation in the single-navigation mode.

2. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, Determining the motion state of each target vehicle in the vehicle formation includes: Each target vehicle in the vehicle formation is taken as a node, and the following relationship between the target vehicles is taken as a directed edge. The node attributes of the node include vehicle state and formation state. The vehicle state includes vehicle position, speed, heading angle, control input value and lane. The formation state includes vehicle role, vehicle control mode and driver state. The edge attributes of the directed edge include inter-vehicle communication quality, inter-vehicle perception reachability and relative distance between vehicles. Based on the nodes and the directed edges, construct a directed dynamic graph corresponding to the vehicle formation; A control subgraph is selected from the directed dynamic graph based on the edge attributes; During the vehicle formation movement, the motion state of each target vehicle in the vehicle formation is obtained based on the nodes and directed edges in the control subgraph.

3. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, The step of determining that a target vehicle in the vehicle formation has an anomaly based on the motion state includes: Based on the motion state, determine the communication status of each target vehicle with other vehicles, the distance between vehicles, and the driver status of each target vehicle; The vehicle formation is determined to have an anomaly if any of the following conditions are met: The communication status is that the communication link is interrupted; The distance between vehicles is greater than a preset vehicle distance threshold; The driver status indicates whether the driving duration exceeds a safety threshold or the driver requests to relinquish control.

4. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, When the target vehicle exhibiting the anomaly is a lead vehicle in single-leader mode and the anomaly type is passive, splitting the vehicle formation into multiple sub-formations in multi-leader mode includes: The vehicle formation is divided into multiple sub-formations; For each target vehicle in the sub-formation, the number of following vehicles corresponding to the target vehicle is counted, and the target vehicle with the largest number of following vehicles is determined as a candidate vehicle. When the number of candidate vehicles is 1, the candidate vehicle is determined as the lead vehicle of the sub-formation, and the lead vehicle is broadcast to the sub-formation. When the number of candidate vehicles is greater than 1, the candidate vehicle with the highest status index is selected from the candidate vehicles that are allowed to lead the way as the lead vehicle of the sub-formation, and the lead vehicle is broadcast to the sub-formation. The status index is positively correlated with the communication quality index and the control stability index.

5. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, When the target vehicle exhibiting the anomaly is a lead vehicle in single-leader mode and the anomaly type is active, splitting the vehicle formation into multiple sub-formations in multi-leader mode includes: The vehicle formation is divided into multiple sub-formations, and a deweighting signal for a target vehicle that has an anomaly is broadcast in the sub-formation. The deweighting signal is used to characterize the vehicle that abandons the role of the lead vehicle. After each sub-formation receives the downweighting signal, the communication quality index, control stability index, and navigation feasibility index of each target vehicle in the sub-formation are determined. The communication quality index, the control stability index, and the navigation feasibility index are weighted and calculated respectively, and the weighted results are summed to obtain a comprehensive index. The target vehicle with the highest comprehensive index is designated as the lead vehicle in the sub-formation, and the lead vehicle is broadcast to the sub-formation.

6. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, The preset formation fusion conditions include: The real-time relative distance between target vehicles in each sub-formation is less than a preset distance threshold; The target vehicles in each sub-formation travel in the same direction; The target vehicle that experienced the anomaly has re-established its communication link.

7. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 2, characterized in that, The step of reorganizing the sub-formation into a vehicle formation in the single-navigation mode includes: Determine the duration of stability of the reorganized vehicle formation and the vehicle formation error; When the duration of the stable state exceeds a preset time threshold, the vehicle formation error is less than a preset error range, and the lead vehicle in the single lead mode has been determined, it is determined that the reorganized vehicle formation has been converted from the multi-lead mode to the single lead mode. The nodes and directed edges in the control subgraph are updated based on the reorganized vehicle formation.

8. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 1, characterized in that, When the vehicle formation operation mode is single-navigation mode, the method further includes: Identify the reference vehicle to follow for each target vehicle in the vehicle platoon; For each target vehicle, the expected offset between the target vehicle and the corresponding following reference vehicle is obtained, and the expected trajectory of the target vehicle is constructed based on the expected offset. The desired trajectory is optimized by calling a preset state prediction model, a motion control command for the target vehicle is generated, and the motion control command is sent to the corresponding target vehicle to control the movement of the target vehicle.

9. The multi-leader vehicle formation control method based on formation graph connectivity judgment according to claim 8, characterized in that, The step of invoking a preset state prediction model to optimize the desired trajectory and generating motion control commands for the target vehicle includes: Determine the vehicle status of each target vehicle and the following vehicle status of the corresponding reference vehicle; Determine the difference between the target value of the vehicle state and the reference value of the following vehicle state, and construct a target function based on the target control input value included in the target value and the difference; Determine the constraints on the objective function, the constraints including: The target value is within a preset threshold range, the target control input value is within a preset control threshold range, the actual relative distance between the target vehicle and the vehicle in front is greater than a preset safe distance, and the fluctuation value of the target control input value is less than a preset fluctuation threshold. Under the constraints, a preset state prediction model is invoked to optimize the objective function, thereby obtaining the motion control command for the target vehicle.

10. A multi-leader vehicle platooning control system based on platooning graph connectivity judgment, characterized in that, The control system includes: The determination module is used to determine the motion state of each target vehicle in the vehicle formation when the formation operation mode is single navigator mode; The switching module, if it determines that a target vehicle in the vehicle formation is abnormal based on the motion state, then splits the vehicle formation into multiple sub-formations in the multi-navigation mode; The recovery module is used to reorganize the sub-formation into a vehicle formation in the single-navigation mode when each sub-formation meets the preset formation fusion conditions.

Images