Vehicle formation method of vehicle and queue two-dimensional FSM model

Abstract

The invention provides a vehicle formation method of a vehicle and queue two-dimensional FSM model, relates to the field of automatic driving, and aims to solve the problems of insufficient dynamic environment adaptability, vehicle intelligent screening and formation state coordination mechanism in the prior art. S2, monitoring an interaction event between the vehicle and the queue in real time, and triggering state transition based on the two-dimensional state library of the vehicle and the queue; and S3, performing a state execution action based on the state transition, and performing speed coordination and lane change control through the state execution action, thereby optimizing the time headway and expanding the formation scale. The method has a good application prospect in the field of cooperative traffic management of the automatic driving vehicles, and can significantly improve the formation passing efficiency, reduce the energy consumption and relieve the traffic jam.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving, and specifically to a vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons. Background Technology

[0002] With the rapid development of autonomous vehicles, how to better utilize them to improve traffic flow has become a key research focus. Vehicle platooning, due to its ability to travel with smaller headway, plays a significant role in reducing wind resistance and improving traffic capacity. With ongoing research into platooning operations, vehicle control within a platoon has gradually matured. Car-following models such as Adaptive Cruise Control (ACC) and Cooperative Adaptive Cruise Control (CACC) have been widely applied and proven to achieve the goals of reducing fuel consumption and improving traffic capacity.

[0003] However, significant limitations remain in platooning formation efficiency and vehicle selection mechanisms. Traditional methods rely on fixed paths or manual rules to select members, lacking real-time response capabilities in dynamic environments. The vehicle selection mechanism suffers from adaptability deficiencies, relying solely on path overlap matching while ignoring differences in power characteristics. The absence of state management and coordination mechanisms, and the lack of smooth transition strategies when vehicles join or leave, restricts platooning stability and throughput. These issues expose the shortcomings of existing platooning technologies in dynamic adaptability and global coordination, necessitating breakthroughs through intelligent modeling and adaptive control. Summary of the Invention

[0004] To address the shortcomings of existing technologies in dynamic environment adaptability, intelligent vehicle selection, and platooning state coordination mechanisms, this invention provides the following solution: a vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons, comprising the following steps: S1. Establish a two-dimensional state database for vehicles and queues; S2. Monitor the interaction events between vehicles and queues in real time, and trigger state transitions based on the two-dimensional state database of the vehicles and queues; S3. Based on the state transition, perform state execution actions, and use the state execution actions to perform speed coordination and lane change control, thereby optimizing the headway and expanding the formation size.

[0005] Furthermore, the two-dimensional state library of vehicles and queues includes a vehicle state library and a queue state library. The vehicle queue library is divided into states based on the vehicle state and the state relationship between the vehicle and the queue. The queue state library is divided into states based on the queue state and the state relationship between the queue and the vehicle.

[0006] Furthermore, the vehicle status database is divided into three categories: The first category is the bicycle status, including: search queue, waiting to join queue, and create queue; The second category is the queue entry status, including: entry at the head of the queue, entry at the tail of the queue, and insertion in the middle of the queue; The third category is the state in the queue, including: the state of the lead vehicle and the states of the following vehicles, where: The head vehicle status includes: maintaining stability, the tail of the convoy separating, and the tail of the convoy merging; The following vehicle states include: maintaining stability, leaving the queue, and increasing the distance between vehicles.

[0007] Furthermore, the queue state library is divided into six categories of queue reorganization operations: The first type of operation is to maintain stability, including: maintaining stability of the lead vehicle and maintaining stability of the following vehicles; The second type of operation is merging vehicles, including: keeping the lead vehicle stable, keeping the following vehicles stable, and increasing the distance between the following vehicles; The third type of operation is vehicle separation, including: keeping the lead vehicle stable, keeping the following vehicle stable, and separating from the rear of the convoy; The fourth type of operation is merging queues, including: keeping the lead car stable and keeping the following cars stable; The fifth type of operation is the merged queue, including: keeping the lead car stable, keeping the following cars stable, and merging the lead and tail of the queue; The sixth type of operation is to separate the queue, including: keeping the lead vehicle stable, keeping the following vehicles stable, and separating the lead vehicle from the tail of the queue.

[0008] Furthermore, the state transition includes vehicle state transition and queue state transition; The vehicle state transition is divided into three cases: The first scenario: If no vehicle can be added to the queue or a vehicle on the same path, the current state remains unchanged. The second scenario: When the vehicle is in the search queue, if the target queue is found, the vehicle status of the vehicle will be triggered to join at the head, join at the tail, or insert in the middle according to the insertion position. At the same time, the target queue status will be driven to change from maintaining stability to merging vehicles. The status of the vehicles immediately following the insertion position in the target queue will change from maintaining stability to increasing the spacing. At the same time, the binding relationship between the vehicle and the target queue will be established. The third scenario: When the vehicle is in the search for a vehicle ahead, the preceding vehicle's status is triggered to create a queue to generate a new queue. After the new queue is created, the vehicle's status changes to join the queue and it joins from the end. The queue state transition is divided into five scenarios: vehicle merging, vehicle separation, queue separation, queue discovery and queue merging, and no change. Vehicle merger scenarios include: S2.1.1: When a vehicle merging event is triggered, the queue state changes from stable to merged vehicles. When a vehicle is inserted in the middle position, the state of the vehicle immediately following the insertion position changes from stable to increased spacing. S2.1.2: When the vehicle merging is completed, the state of this queue changes from merging vehicles to remaining stable; Vehicle separation scenarios include: S2.2.1: When the event of separating a vehicle is triggered, the state of this queue changes from "Stable" to "Separated Vehicle". The state of the vehicle to be separated changes from "Stable" to "Leave Queue". S2.2.2: When vehicle separation is complete, the queue state changes from separated vehicles to stable. Queue separation scenarios include: S2.3.1: When the queue separation event is triggered, the state of this queue changes from "maintaining stability" to "separated queue". The state of the lead car in the queue to be separated changes from "following car - maintaining stability" to "lead car and tail car separated", and the queue creation event is triggered at the same time. S2.3.2: When separation is complete, the state of this queue changes from separated queue to stable. If a new queue is created from the queue to be separated, the state of the new queue is stable. Queue and queue merging scenarios found include: S2.4.1: When the event of finding a queue is triggered, the state of the queue to be merged changes from stable to being merged, the head car of the queue to be merged changes from stable to being merged at the tail, and the queue merging event is triggered at the same time, the state of the main queue changes from stable to merging queue. S2.4.2: When the merge is complete, the two queues are merged into one queue, the queue to be merged is cancelled, the state of the lead car in the queue to be merged changes to following car - stable, and the state of the main queue changes from merged queue to stable. No change: When there are no separated vehicles in the queue and no other queues or vehicles request to merge, the queue state remains unchanged.

[0009] Furthermore, the state execution actions include vehicle state execution actions and queue state execution actions; The vehicle status execution action is divided into three stages: Phase 1: Search the queue; if possible, add it to the queue dictionary. If the value is not empty, the event will be searched in the queue. The specific steps are as follows: S1: Obtain surrounding vehicles, search for vehicles within a certain range around the current vehicle, and generate a dictionary of adjacent vehicles. Adjacent vehicle dictionary The key represents the adjacency relationship with the current vehicle, including: front vehicle, rear vehicle, left front, left rear, right front, and right rear, and the value is the vehicle ID; S2: Obtain queue information for surrounding vehicles, based on the adjacent vehicle dictionary. Given the vehicle ID, query the queue ID to which the vehicle belongs and generate an adjacent queue dictionary. ; The key is The keys are determined, including: front queue, back queue, left front, left back, right front, and right back, and the value is the queue ID; S3: Filter the queues that can be joined, and query the dictionary of adjacent queues. For each queue, if the next segment of the current vehicle's path coincides with the next segment of the queue's head vehicle's path, and the queue size has not reached its upper limit, then in the adjacent queue dictionary... If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove key-value pairs from the queue; S4: Determine if there is a possibility of adding to the queue. If there is a possibility, check the adjacent queue dictionary. If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove the key-value pairs from the queue. The front and back queues can be added directly. For other queues, check for lane-changing space. If vehicles to the left and right are in the same queue, lane-changing space can be created for the current vehicle, or it can be inserted in the middle. If vehicles to the left front and left rear, or right front and right rear are not in the same queue, check for lane-changing gaps; if no gaps exist, the vehicle cannot be added. Classify the queues that can be added and generate a dictionary of eligible queues. The keys of the dictionary represent the method of addition, categorized as adding to the head of the queue, adding to the tail of the queue, and inserting into the middle of the queue. Phase Two: Search for the vehicle ahead; if the vehicle ahead is in the dictionary... If the value is not empty, an event is triggered to find the preceding vehicle. The specific steps are as follows: Step 1: Obtain surrounding vehicles, search for vehicles within a certain range in front of the current vehicle, and generate a dictionary of vehicles ahead. Vehicle dictionary ahead The key represents the adjacency relationship with the current vehicle, including: directly in front, left front, and right front; Step 2: Filter vehicles whose next path segments overlap and whose vehicle status is also "single vehicle_search queue"; determine whether the next path segments of the vehicles directly in front, to the left, and to the right overlap with the next path segments of the current vehicle's path, and whether the status of the vehicles in front is also "single vehicle_search queue". If so, add them to the dictionary of vehicles in front. Reserved in the dictionary; otherwise, it will be from the vehicle dictionary ahead. Remove the key-value pair of the preceding vehicle; Phase 3: No change. If the search fails and the previous vehicle in the search queue is not found, the current state is maintained. The queue state execution action is divided into three stages: The first step involves basic speed and lane-changing operations: The lead vehicle in the convoy drives according to its own vehicle control system, and the following vehicles in the convoy perform speed and lane-changing operations based on the movement status of the lead vehicle and themselves. The following operations are performed for each following vehicle: First, determine whether the speeds of the lead vehicle and the currently following vehicle are not zero. If not, drive according to the vehicle's own control system without giving any suggestions. If so, determine whether the currently following vehicle belongs to the same queue as the lead vehicle. Then, if the following vehicle belongs to the same queue as the vehicle in front, the lane of the leading vehicle is used as the target lane for lane changing, and the distance between the current vehicle and the vehicle in front is calculated. If the distance is greater than the maximum value of the safe distance range, the vehicle accelerates; if the distance is less than the minimum value of the safe distance range, the vehicle decelerates. If the following vehicle does not belong to the same queue as the vehicle in front, the lane of the vehicle immediately in front of the current following vehicle in the queue is used as the target lane for lane changing, and the speed is controlled by the vehicle's own control system. The second step is to maintain a stable state: First, determine if the distance between vehicles in the queue exceeds the search range. If it does, determine if the vehicle whose distance from the vehicle in front exceeds the search range is the last vehicle in the queue. If it is the last vehicle, change the state of the vehicle to "Left Queue" and change the state of the queue to "Separated Vehicle". If it is not the last vehicle, change the state of the vehicle to "Lead Vehicle - Tail of Queue Separated" and change the state of the queue to "Separated Queue". Then, if the queue state remains stable, it is determined whether there are any vehicles in the queue that want to leave the queue, and whether the next segment of the following vehicle's path overlaps with the next segment of the lead vehicle's path. If they do not overlap, the state of the following vehicle is changed to leave the queue. Secondly, if the queue state remains stable, search for vehicles ahead that have joined the queue; search for vehicles within the search range ahead of the current vehicle, query the vehicle's queue ID, and determine whether the next segment of the path of the lead vehicle in this queue overlaps with the next segment of the path of the lead vehicle in the searched queue and whether the searched queue is in a stable state. If so, they can be merged. The searched queue state changes to merged queue, the current queue state changes to merged queue, and the current queue's lead vehicle state changes to _lead vehicle_tail of the queue merged. Finally, if none of the above situations occur, the queue state remains unchanged; The third step involves a state transition: When merging vehicles, the vehicle to be merged changes lanes with the lead vehicle in the queue as the target lane. The distance between the vehicle to be merged and the vehicle immediately in front of it in the queue is calculated. If the distance exceeds the safe distance range, the vehicle accelerates; if the distance is less than the safe distance range, the vehicle decelerates; if the distance is within the safe distance range and the vehicle to be merged and the lead vehicle in the queue are in the same lane, the merge completion event is triggered. When separating a vehicle, this queue removes the vehicle to be separated from the vehicle management list and triggers a separation completion event; When merging queues, the lead car in the queue to be merged changes lanes with the lead car in the current queue as the target lane. The distance between the lead car in the queue to be merged and the last car in the current queue is calculated. If the distance exceeds the safe distance range, the lead car in the queue to be merged accelerates; if the distance is less than the safe distance range, the lead car in the queue to be merged decelerates; if the distance is within the safe distance range and the lead car in the queue to be merged and the lead car in the current queue are in the same lane, the merge completion event is triggered. When separating a queue, the vehicle to be separated is removed from the vehicle management list, triggering a separation completion event. The queue to be separated then triggers a queue creation event, generating a new queue.

[0010] Furthermore, search for vehicles passing within a certain range ahead of the current vehicle: ; To achieve, among which, The current speed of the vehicle, The search range is 100, which represents the static safety distance.

[0011] Compared with the prior art, the beneficial effects of the present invention are: by constructing a two-dimensional finite state machine (FSM) model of vehicles and queuing and a dynamic cooperative control mechanism, the present invention significantly improves queuing efficiency and stability, thereby increasing traffic volume; This invention enhances adaptability to complex traffic environments through dynamic search range and vehicle filtering logic; This invention shortens headway, reduces energy consumption, and minimizes traffic disturbances caused by sudden acceleration and deceleration by using state transition rules and cooperative control strategies. Attached Figure Description

[0012] Figure 1 A flowchart of a vehicle queuing method based on a two-dimensional FSM model of vehicles and queues; Figure 2 This is a flowchart of the vehicle state transition process; Figure 3 Here is a flowchart of the queue state transition process; Figure 4 This is a comparison chart showing the improvement in traffic efficiency in the SUMO simulation environment.

[0013] in, Figure 4 Figure 'a' shows the throughput simulation diagram for the no-formation method, with a throughput of 835.2 pcu / h. b is a simulation graph of the throughput of the open-source baseline formation algorithm, with a throughput of 1612.8 pcu / h; c is a simulation diagram of the throughput of this invention, with a throughput of 2059.2 pcu / h. Detailed Implementation

[0014] Example 1, combined with Figure 1 This embodiment describes a vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons, comprising the following steps: S1. Establish a two-dimensional state database for vehicles and queues; S2. Monitor the interaction events between vehicles and queues in real time, and trigger state transitions based on the two-dimensional state database of the vehicles and queues; S3. Based on the state transition, perform state execution actions, and use the state execution actions to perform speed coordination and lane change control, thereby optimizing the headway and expanding the formation size.

[0015] The two-dimensional state database of vehicles and queues includes a vehicle state database and a queue state database. The vehicle queue database is divided into states based on the vehicle state and the state relationship between the vehicle and the queue. The queue state database is divided into states based on the queue state and the state relationship between the queue and the vehicle.

[0016] The vehicle status database is divided into three categories: The first category is the bicycle status, including: search queue, waiting to join queue, and create queue; The second category is the queue entry status, including: entry at the head of the queue, entry at the tail of the queue, and insertion in the middle of the queue; The third category is the state in the queue, including: the state of the lead vehicle and the states of the following vehicles, where: The head vehicle status includes: maintaining stability, the tail of the convoy separating, and the tail of the convoy merging; The following vehicle states include: maintaining stability, leaving the queue, and increasing the distance between vehicles.

[0017] Specifically, the vehicle status database is shown in Table 1 below: Table 1

[0018] The queue state database is divided into six categories of queue reorganization operations: The first type of operation is to maintain stability, including: maintaining stability of the lead vehicle and maintaining stability of the following vehicles; The second type of operation is merging vehicles, including: keeping the lead vehicle stable, keeping the following vehicles stable, and increasing the distance between the following vehicles; The third type of operation is vehicle separation, including: keeping the lead vehicle stable, keeping the following vehicle stable, and separating from the rear of the convoy; The fourth type of operation is merging queues, including: keeping the lead car stable and keeping the following cars stable; The fifth type of operation is the merged queue, including: keeping the lead car stable, keeping the following cars stable, and merging the lead and tail of the queue; The sixth type of operation is to separate the queue, including: keeping the lead vehicle stable, keeping the following vehicles stable, and separating the lead vehicle from the tail of the queue.

[0019] Specifically, the queue state library is shown in Table 2 below: Table 2

[0020] The state transitions include vehicle state transitions and queue state transitions; The vehicle state transition is divided into three cases: The first scenario: If no vehicle can be added to the queue or a vehicle on the same path, the current state remains unchanged. The second scenario: When the vehicle is in the search queue, if the target queue is found, the vehicle status of the vehicle will be triggered to join at the head, join at the tail, or insert in the middle according to the insertion position. At the same time, the target queue status will be driven to change from maintaining stability to merging vehicles. The status of the vehicles immediately following the insertion position in the target queue will change from maintaining stability to increasing the spacing. At the same time, the binding relationship between the vehicle and the target queue will be established. The third scenario: When the vehicle is in the search for a vehicle ahead, the preceding vehicle's status is triggered to create a queue to generate a new queue. After the new queue is created, the vehicle's status changes to join the queue and it joins from the end. Specifically, through Figure 2 It is known how vehicles drive their own state changes through event triggers, and synchronously trigger a chain reaction of queue states through a binding mechanism.

[0021] In the second scenario, when the vehicle is at the head or tail of the queue, the head and tail joining event is triggered, and the queue state changes from remaining stable to merging vehicles, while simultaneously establishing a binding relationship between the vehicle and the target queue. When the vehicle is inserted in the middle of the queue, the middle insertion event is triggered, and the queue state changes from remaining stable to merging vehicles. The state of the vehicle immediately following the insertion position in the target queue changes from remaining stable to increasing the spacing, while simultaneously establishing a binding relationship between the vehicle and the target queue.

[0022] The detailed state transition list for the second scenario is shown in Table 3 below: Table 3

[0023] The queue state transition is divided into five scenarios, including vehicle merging, vehicle separation, queue separation, queue discovery and queue merging, and no change. Figure 3 It can be seen that there are three types of operations for dynamic reorganization of queue structure and how the queue is formed through state transition.

[0024] Vehicle merger scenarios include: S2.1.1: When a vehicle merging event is triggered, the queue state changes from stable to merged vehicles. When a vehicle is inserted in the middle position, the state of the vehicle immediately following the insertion position changes from stable to increased spacing. S2.1.2: When the vehicle merging is completed, the state of this queue changes from merging vehicles to remaining stable; When this vehicle is in a vehicle merger situation, the detailed state transition list is shown in Table 5 below: Table 5

[0025] Vehicle separation scenarios include: S2.2.1: When the event of separating a vehicle is triggered, the state of this queue changes from "Stable" to "Separated Vehicle". The state of the vehicle to be separated changes from "Stable" to "Leave Queue". S2.2.2: When vehicle separation is complete, the queue state changes from separated vehicles to stable. When this vehicle is in a disengaged state, the detailed state transition list is shown in Table 6 below: Table 6

[0026] Queue separation scenarios include: S2.3.1: When the queue separation event is triggered, the state of this queue changes from "maintaining stability" to "separated queue". The state of the lead car in the queue to be separated changes from "following car - maintaining stability" to "lead car and tail car separated", and the queue creation event is triggered at the same time. S2.3.2: When separation is complete, the state of this queue changes from separated queue to stable. If a new queue is created from the queue to be separated, the state of the new queue is stable. When this vehicle is in a queue separation situation, the detailed state transition list is shown in Table 7 below: Table 7

[0027] Queue and queue merging scenarios found include: S2.4.1: When the event of finding a queue is triggered, the state of the queue to be merged changes from stable to being merged, the head car of the queue to be merged changes from stable to being merged at the tail, and the queue merging event is triggered at the same time, the state of the main queue changes from stable to merging queue. S2.4.2: When the merge is complete, the two queues are merged into one queue, the queue to be merged is cancelled, the state of the lead car in the queue to be merged changes to following car - stable, and the state of the main queue changes from merged queue to stable. When this vehicle is in the queue search and queue merging phases, the detailed state transition list is shown in Table 8 below: Table 8

[0028] No change: When there are no separated vehicles in the queue and no other queues or vehicles request to merge, the queue state remains unchanged.

[0029] The state execution actions include vehicle state execution actions and queue state execution actions; The vehicle status execution action is divided into three stages: Phase 1: Search the queue; if possible, add it to the queue dictionary. If the value is not empty, the event will be searched in the queue. The specific steps are as follows: S1: Obtain surrounding vehicles, search for vehicles within a certain range around the current vehicle, and generate a dictionary of adjacent vehicles. Adjacent vehicle dictionary The key represents the adjacency relationship with the current vehicle, including: front vehicle, rear vehicle, left front, left rear, right front, and right rear, and the value is the vehicle ID; S2: Obtain queue information for surrounding vehicles, based on the adjacent vehicle dictionary. Given the vehicle ID, query the queue ID to which the vehicle belongs and generate an adjacent queue dictionary. ; The key is The keys are determined, including: front queue, back queue, left front, left back, right front, and right back, and the value is the queue ID; S3: Filter the queues that can be joined, and query the dictionary of adjacent queues. For each queue, if the next segment of the current vehicle's path coincides with the next segment of the queue's head vehicle's path, and the queue size has not reached its upper limit, then in the adjacent queue dictionary... If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove key-value pairs from the queue; S4: Determine if there is a possibility of adding to the queue. If there is a possibility, check the adjacent queue dictionary. If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove the key-value pairs from the queue. The front and back queues can be added directly. For other queues, check for lane-changing space. If vehicles to the left and right are in the same queue, lane-changing space can be created for the current vehicle, or it can be inserted in the middle. If vehicles to the left front and left rear, or right front and right rear are not in the same queue, check for lane-changing gaps; if no gaps exist, the vehicle cannot be added. Classify the queues that can be added and generate a dictionary of eligible queues. The keys of the dictionary represent the method of addition, categorized as adding to the head of the queue, adding to the tail of the queue, and inserting into the middle of the queue. Phase Two: Search for the vehicle ahead; if the vehicle ahead is in the dictionary... If the value is not empty, an event is triggered to find the preceding vehicle. The specific steps are as follows: Step 1: Obtain surrounding vehicles, search for vehicles within a certain range in front of the current vehicle, and generate a dictionary of vehicles ahead. Vehicle dictionary ahead The key represents the adjacency relationship with the current vehicle, including: directly in front, left front, and right front; Step 2: Filter vehicles whose next path segments overlap and whose vehicle status is also "single vehicle_search queue"; determine whether the next path segments of the vehicles directly in front, to the left, and to the right overlap with the next path segments of the current vehicle's path, and whether the status of the vehicles in front is also "single vehicle_search queue". If so, add them to the dictionary of vehicles in front. Reserved in the dictionary; otherwise, it will be from the vehicle dictionary ahead. Remove the key-value pair of the preceding vehicle; Phase 3: No change. If the search fails and the previous vehicle in the search queue is not found, the current state is maintained. The queue state execution action is divided into three stages: The first step involves basic speed and lane-changing operations: The lead vehicle in the convoy drives according to its own vehicle control system, and the following vehicles in the convoy perform speed and lane-changing operations based on the movement status of the lead vehicle and themselves. The following operations are performed for each following vehicle: First, determine whether the speeds of the lead vehicle and the currently following vehicle are not zero. If not, drive according to the vehicle's own control system without giving any suggestions. If so, determine whether the currently following vehicle belongs to the same queue as the lead vehicle. Then, if the following vehicle belongs to the same queue as the vehicle in front, the lane of the leading vehicle is used as the target lane for lane changing, and the distance between the current vehicle and the vehicle in front is calculated. If the distance is greater than the maximum value of the safe distance range, the vehicle accelerates; if the distance is less than the minimum value of the safe distance range, the vehicle decelerates. If the following vehicle does not belong to the same queue as the vehicle in front, the lane of the vehicle immediately in front of the current following vehicle in the queue is used as the target lane for lane changing, and the speed is controlled by the vehicle's own control system. The second step is to maintain a stable state: First, determine if the distance between vehicles in the queue exceeds the search range. If it does, determine if the vehicle whose distance from the vehicle in front exceeds the search range is the last vehicle in the queue. If it is the last vehicle, change the state of the vehicle to "Left Queue" and change the state of the queue to "Separated Vehicle". If it is not the last vehicle, change the state of the vehicle to "Lead Vehicle - Tail of Queue Separated" and change the state of the queue to "Separated Queue". Then, if the queue state remains stable, it is determined whether there are any vehicles in the queue that want to leave the queue, and whether the next segment of the following vehicle's path overlaps with the next segment of the lead vehicle's path. If they do not overlap, the state of the following vehicle is changed to leave the queue. Secondly, if the queue state remains stable, search for vehicles ahead that have joined the queue; search for vehicles within the search range ahead of the current vehicle, query the vehicle's queue ID, and determine whether the next segment of the path of the lead vehicle in this queue overlaps with the next segment of the path of the lead vehicle in the searched queue and whether the searched queue is in a stable state. If so, they can be merged. The searched queue state changes to merged queue, the current queue state changes to merged queue, and the current queue's lead vehicle state changes to _lead vehicle_tail of the queue merged. Finally, if none of the above situations occur, the queue state remains unchanged; The third step involves a state transition: When merging vehicles, the vehicle to be merged changes lanes with the lead vehicle in the queue as the target lane. The distance between the vehicle to be merged and the vehicle immediately in front of it in the queue is calculated. If the distance exceeds the safe distance range, the vehicle accelerates; if the distance is less than the safe distance range, the vehicle decelerates; if the distance is within the safe distance range and the vehicle to be merged and the lead vehicle in the queue are in the same lane, the merge completion event is triggered. When separating a vehicle, this queue removes the vehicle to be separated from the vehicle management list and triggers a separation completion event; When merging queues, the lead car in the queue to be merged changes lanes with the lead car in the current queue as the target lane. The distance between the lead car in the queue to be merged and the last car in the current queue is calculated. If the distance exceeds the safe distance range, the lead car in the queue to be merged accelerates; if the distance is less than the safe distance range, the lead car in the queue to be merged decelerates; if the distance is within the safe distance range and the lead car in the queue to be merged and the lead car in the current queue are in the same lane, the merge completion event is triggered. When separating a queue, the vehicle to be separated is removed from the vehicle management list, triggering a separation completion event. The queue to be separated then triggers a queue creation event, generating a new queue.

[0030] Furthermore, search for vehicles passing within a certain range ahead of the current vehicle: ; To achieve, among which, The current speed of the vehicle, The search range is 100, which represents the static safety distance.

[0031] Example 2, combined with Figure 4This embodiment illustrates that, as shown in the comparison chart of traffic efficiency improvement in the SUMO simulation environment, the present invention can increase the traffic capacity of roads. Taking an accident on a three-lane highway that renders two lanes impassable as an example, in the comparative experiment conducted in the SUMO simulation environment, the present invention increases traffic capacity by 147% compared to the case without platooning; compared to the open-source baseline platooning method, the present invention increases traffic capacity by 28%.

[0032] Although the invention has been described with reference to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and instructional purposes, and not for the purpose of interpreting or limiting the subject matter of the invention. Therefore, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of the invention is illustrative and not restrictive, and the scope of the invention is defined by the appended claims.

Claims

1. A vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons, characterized in that, Includes the following steps: S1. Establish a two-dimensional state database for vehicles and queues; S2. Monitor the interaction events between vehicles and queues in real time, and trigger state transitions based on the two-dimensional state database of the vehicles and queues; S3. Based on the state transition, perform state execution actions, and use the state execution actions to perform speed coordination and lane change control, thereby optimizing the headway and expanding the formation size.

2. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 1, characterized in that, The two-dimensional state database of vehicles and queues includes a vehicle state database and a queue state database. The vehicle queue database is divided into states based on the vehicle state and the state relationship between the vehicle and the queue. The queue state database is divided into states based on the queue state and the state relationship between the queue and the vehicle.

3. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 2, characterized in that, The vehicle status database is divided into three categories: The first category is the bicycle status, including: search queue, waiting to join queue, and create queue; The second category is the queue entry status, including: entry at the head of the queue, entry at the tail of the queue, and insertion in the middle of the queue; The third category is the state in the queue, including: the state of the lead vehicle and the states of the following vehicles, where: The head vehicle status includes: maintaining stability, the tail of the convoy separating, and the tail of the convoy merging; The following vehicle states include: maintaining stability, leaving the queue, and increasing the distance between vehicles.

4. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 3. It is characterized in that The queue status is divided into six categories of queue reorganization operations: The first type of operation is to maintain stability, including: maintaining stability of the lead vehicle and maintaining stability of the following vehicles; The second type of operation is merging vehicles, including: keeping the lead vehicle stable, keeping the following vehicles stable, and increasing the distance between the following vehicles; The third type of operation is vehicle separation, including: keeping the lead vehicle stable, keeping the following vehicle stable, and separating from the rear of the convoy; The fourth type of operation is merging queues, including: keeping the lead car stable and keeping the following cars stable; The fifth type of operation is the merged queue, including: keeping the lead car stable, keeping the following cars stable, and merging the lead and tail of the queue; The sixth type of operation is queue separation, including: keeping the lead vehicle stable, keeping the following vehicles stable, and the lead convoy. Tail separation.

5. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 4, characterized in that, The state transitions include vehicle state transitions and queue state transitions; The vehicle state transition is divided into three cases: The first scenario: If no vehicle can be added to the queue or a vehicle on the same path, the current state remains unchanged. The second scenario: When the vehicle is in the search queue, if the target queue is found, the vehicle status of the vehicle will be triggered to join at the head, join at the tail, or insert in the middle according to the insertion position. At the same time, the target queue status will be driven to change from maintaining stability to merging vehicles. The status of the vehicles immediately following the insertion position in the target queue will change from maintaining stability to increasing the spacing. At the same time, the binding relationship between the vehicle and the target queue will be established. The third scenario: When the vehicle is in the search for a vehicle ahead, the preceding vehicle's status is triggered to create a queue to generate a new queue. After the new queue is created, the vehicle's status changes to join the queue and it joins from the end. The queue state transition is divided into five scenarios: vehicle merging, vehicle separation, queue separation, queue discovery and queue merging, and no change. Vehicle merger scenarios include: S2.1.1: When a vehicle merging event is triggered, the queue state changes from remaining stable to merging vehicles. When a vehicle is inserted in the middle, the state of the vehicle immediately following the insertion position changes from remaining stable to increasing the spacing. S2.1.2: When the vehicle merging is completed, the state of this queue changes from merging vehicles to remaining stable; Vehicle separation scenarios include: S2.2.1: When the event of separating a vehicle is triggered, the state of this queue changes from "Stable" to "Separated Vehicle"; the state of the vehicle to be separated changes from "Stable" to "Leave Queue". S2.2.2: When vehicle separation is complete, the queue state changes from separated vehicles to stable. Queue separation scenarios include: S2.3.1: When the queue separation event is triggered, the state of this queue changes from "maintaining stability" to "separated queue"; the state of the lead car in the queue to be separated changes from "following car - maintaining stability" to "lead car and tail car separated", and the queue creation event is triggered at the same time. S2.3.2: When separation is complete, the state of this queue changes from separated queue to stable. If a new queue is created from the queue to be separated, the state of the new queue is stable. Queue and queue merging scenarios found include: S2.4.1: When the event of finding a queue is triggered, the state of the queue to be merged changes from stable to being merged, the head car of the queue to be merged changes from stable to being merged at the tail, and the queue merging event is triggered at the same time, the state of the main queue changes from stable to merging queue. S2.4.2: When the merge is complete, the two queues are merged into one queue, the queue to be merged is cancelled, the state of the lead car in the queue to be merged changes to following car - stable, and the state of the main queue changes from merged queue to stable. No change: When there are no separated vehicles in the queue and no other queues or vehicles request to merge, the queue state remains unchanged.

6. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 5, characterized in that, The state execution actions include vehicle state execution actions and queue state execution actions; The vehicle status execution action is divided into three stages: Phase 1: Search the queue; if possible, add it to the queue dictionary. If the value is not empty, the event will be searched in the queue. The specific steps are as follows: S1: Obtain surrounding vehicles, search for vehicles within a certain range around the current vehicle, and generate a dictionary of adjacent vehicles. Adjacent vehicle dictionary The key represents the adjacency relationship with the current vehicle, including: front vehicle, rear vehicle, left front, left rear, right front, and right rear, and the value is the vehicle ID; S2: Obtain queue information for surrounding vehicles, based on the adjacent vehicle dictionary. Given the vehicle ID, query the queue ID to which the vehicle belongs and generate an adjacent queue dictionary. ; The key is The keys are determined, including: front queue, back queue, left front, left back, right front, and right back, and the value is the queue ID; S3: Filter the queues that can be joined, and query the dictionary of adjacent queues. For each queue, if the next segment of the current vehicle's path coincides with the next segment of the queue's head vehicle's path, and the queue size has not reached its upper limit, then in the adjacent queue dictionary... If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove key-value pairs from the queue; S4: Determine if there is a possibility of adding to the queue. If there is a possibility, check the adjacent queue dictionary. If a character is retained in the dictionary, it is otherwise taken from the adjacent queue. Remove the key-value pairs from the queue. The front and back queues can be added directly. For other queues, check for lane-changing space. If vehicles to the left and right are in the same queue, lane-changing space can be created for the current vehicle, or it can be inserted in the middle. If vehicles to the left front and left rear, or right front and right rear are not in the same queue, check for lane-changing gaps; if no gaps exist, the vehicle cannot be added. Classify the queues that can be added and generate a dictionary of eligible queues. The keys of the dictionary represent the method of addition, categorized as adding to the head of the queue, adding to the tail of the queue, and inserting into the middle of the queue. Phase Two: Search for the vehicle ahead; if the vehicle ahead is in the dictionary... If the value is not empty, an event is triggered to find the preceding vehicle. The specific steps are as follows: Step 1: Obtain surrounding vehicles, search for vehicles within a certain range in front of the current vehicle, and generate a dictionary of vehicles ahead. Vehicle dictionary ahead The key represents the adjacency relationship with the current vehicle, including: directly in front, left front, and right front; Step 2: Filter vehicles whose next path segments overlap and whose vehicle status is also "single vehicle_search queue"; determine whether the next path segments of the vehicles directly in front, to the left, and to the right overlap with the next path segments of the current vehicle's path, and whether the status of the vehicles in front is also "single vehicle_search queue". If so, add them to the dictionary of vehicles in front. Reserved in the dictionary; otherwise, it will be from the vehicle dictionary ahead. Remove the key-value pair of the preceding vehicle; Phase 3: No change. If the search fails and the previous vehicle in the search queue is not found, the current state is maintained. The queue state execution action is divided into three stages: The first step involves basic speed and lane-changing operations: The lead vehicle in the convoy drives according to its own vehicle control system, and the following vehicles in the convoy perform speed and lane-changing operations based on the movement status of the lead vehicle and themselves. The following operations are performed for each following vehicle: First, determine whether the speeds of the lead vehicle and the currently following vehicle are not zero. If not, drive according to the vehicle's own control system without giving any suggestions. If so, determine whether the currently following vehicle belongs to the same queue as the lead vehicle. Then, if the following vehicle belongs to the same queue as the vehicle in front, the lane of the leading vehicle is used as the target lane for lane changing, and the distance between the current vehicle and the vehicle in front is calculated. If the distance is greater than the maximum value of the safe distance range, the vehicle accelerates; if the distance is less than the minimum value of the safe distance range, the vehicle decelerates. If the following vehicle does not belong to the same queue as the vehicle in front, the lane of the vehicle immediately in front of the current following vehicle in the queue is used as the target lane for lane changing, and the speed is controlled by the vehicle's own control system. The second step is to maintain a stable state: First, determine if the distance between vehicles in the queue exceeds the search range. If it does, determine if the vehicle whose distance from the vehicle in front exceeds the search range is the last vehicle in the queue. If it is the last vehicle, change the state of the vehicle to "Left Queue" and change the state of the queue to "Separated Vehicle". If it is not the last vehicle, change the state of the vehicle to "Lead Vehicle - Tail of Queue Separated" and change the state of the queue to "Separated Queue". Then, if the queue state remains stable, it is determined whether there are any vehicles in the queue that want to leave the queue, and whether the next segment of the following vehicle's path overlaps with the next segment of the lead vehicle's path. If they do not overlap, the state of the following vehicle is changed to leave the queue. Secondly, if the queue state remains stable, search for vehicles ahead that have joined the queue; search for vehicles within the search range ahead of the current vehicle, query the vehicle's queue ID, and determine whether the next segment of the path of the lead vehicle in this queue overlaps with the next segment of the path of the lead vehicle in the searched queue and whether the searched queue is in a stable state. If so, they can be merged. The searched queue state changes to merged queue, the current queue state changes to merged queue, and the current queue's lead vehicle state changes to _lead vehicle_tail of the queue merged. Finally, if none of the above situations occur, the queue state remains unchanged; The third step involves a state transition: When merging vehicles, the vehicle to be merged changes lanes with the lead vehicle in the queue as the target lane. The distance between the vehicle to be merged and the vehicle immediately in front of it in the queue is calculated. If the distance exceeds the safe distance range, the vehicle accelerates; if the distance is less than the safe distance range, the vehicle decelerates; if the distance is within the safe distance range and the vehicle to be merged and the lead vehicle in the queue are in the same lane, the merge completion event is triggered. When separating a vehicle, this queue removes the vehicle to be separated from the vehicle management list and triggers a separation completion event; When merging queues, the lead car in the queue to be merged changes lanes with the lead car in the current queue as the target lane. The distance between the lead car in the queue to be merged and the last car in the current queue is calculated. If the distance exceeds the safe distance range, the lead car in the queue to be merged accelerates; if the distance is less than the safe distance range, the lead car in the queue to be merged decelerates; if the distance is within the safe distance range and the lead car in the queue to be merged and the lead car in the current queue are in the same lane, the merge completion event is triggered. When separating a queue, the vehicle to be separated is removed from the vehicle management list, triggering a separation completion event. The queue to be separated then triggers a queue creation event, generating a new queue.

7. The vehicle platooning method based on a two-dimensional FSM model of vehicles and platoons according to claim 6, characterized in that, Search for vehicles passing within a certain range ahead of the current vehicle: ； To achieve, among which, The current speed of the vehicle, The search range is 100, which represents the static safety distance.

Images