Ramp shunting area vehicle queue three-stage collaborative lane changing method in intelligent network connection environment

Abstract

The invention discloses a three-stage cooperative lane changing method for a vehicle queue in a ramp diversion area in an intelligent network connection environment, and the method comprises the steps: firstly, determining a safety gap of a target lane based on a minimum safety distance and a maximum deceleration, and selecting a vehicle which can change the lane to the safety gap from a CAV queue for preferential lane changing; secondly, establishing a longitudinal trajectory planning model based on self-adaptive model predictive control, performing longitudinal distance adjustment on a lane-changed vehicle by using the model, and reversely calculating an initial position required by the vehicle to smoothly drive into a ramp based on a quintic polynomial trajectory planning method; embedding into adaptive model predictive control as a ramp tail end constraint to form a new safety gap of the target lane; and finally, the vehicles which do not change lanes are controlled to sequentially change lanes to the corresponding safety gaps, and lane changing of the CAV queue is achieved. According to the method, the CAV queue can efficiently and orderly drive into the exit ramp, the vehicle speed fluctuation of a shunting area can be reduced, and the overall passing efficiency of a road section is improved.

Description

Technical Field

[0001] This invention relates to the fields of connected autonomous vehicles and traffic control, and in particular to a three-stage coordinated lane-changing method for vehicle queuing in a ramp diversion area under an intelligent connected environment. Background Technology

[0002] Highway ramp merging zones, as crucial areas connecting main roads and ramps, are prone to frequent traffic delays and congestion. In this scenario, vehicles frequently need to change lanes to complete route transitions; especially near exits, the urgency of lane-changing needs clashes with lane resource constraints, easily leading to inappropriate lane-changing decisions and execution, causing speed fluctuations, traffic congestion, and even accidents, impacting the traffic efficiency and safety of ramp merging zones. In recent years, with the deep integration of vehicle-to-everything (V2X) and autonomous driving technologies, research involving vehicle platooning has emerged. Intelligent connected vehicle platooning has enormous potential in alleviating traffic congestion, improving traffic efficiency, enhancing driving safety, and improving fuel economy, providing a new path to solving this problem.

[0003] Existing lane-changing and control methods for ramp divergence zones primarily target single-vehicle lane changes in divergence zones, platoon lane changes in basic road sections, or merging zones, lacking research specifically on platoon lane changes in divergence zones. Furthermore, existing platoon lane-changing strategies are often based on the assumption of sufficient target clearance, failing to consider the fact that platoon lane changes require even greater clearance. In addition, most lane-changing models use fixed parameters, making it difficult to dynamically adjust according to real-time traffic conditions, resulting in poor adaptability to complex traffic environments.

[0004] Current research on vehicle platooning cooperative lane-changing methods: Patent application CN115830908A proposes a method and system for cooperative lane-changing of unmanned vehicle platoons in mixed traffic flow. When the target gap cannot accommodate the entire platoon to change lanes at the same time, the lane-changing process of the platoon is completed in steps. It also takes into account the uncertainty of the driving behavior of human-driven vehicles and the safety and time efficiency in the motion planning process of unmanned vehicles.

[0005] However, the current research scenario is a two-lane section of a highway. It only mentions that existing technologies can recommend the latest lane-changing start position for convoys under different traffic conditions, reminding them to change lanes in time, but it lacks specific methods and experiments. Furthermore, existing technologies use fixed parameters, making it difficult to dynamically adjust according to real-time traffic conditions. Summary of the Invention

[0006] The technical problem to be solved by this invention is: In view of the technical problems existing in the prior art, this invention provides a three-stage coordinated lane changing method for vehicle queuing in the ramp diversion area under intelligent connected environment, so as to realize the efficient and orderly entry of CAV queuing into the exit ramp, thereby reducing the speed fluctuation in the diversion area and improving the overall traffic efficiency of the road section.

[0007] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows: A three-stage coordinated lane-changing method for vehicle queuing in a ramp diversion area under an intelligent connected vehicle environment includes the following steps: The first safety gap of the target lane is determined based on the minimum safe distance and the maximum deceleration. CAV vehicles that can change lanes to the first safety gap are selected from the CAV queue to form a first sub-queue. The first sub-queue is controlled to change lanes to the first safety gap. CAV vehicles in the CAV queue before the first sub-queue form a second sub-queue, and CAV vehicles in the CAV queue after the first sub-queue form a third sub-queue. A longitudinal trajectory planning model based on adaptive model predictive control is established. The longitudinal distance of CAV vehicles in the first sub-queue is adjusted using the longitudinal trajectory planning model based on adaptive model predictive control. At the same time, the initial position required for the vehicle to smoothly enter the ramp is calculated in reverse based on the fifth-order polynomial trajectory planning method, and is embedded into the adaptive model predictive control as the ramp end constraint to form the second and third safety gaps of the target lane. The vehicles in the second sub-queue are controlled to change lanes sequentially to the corresponding second safety gap, and the vehicles in the third sub-queue are controlled to change lanes sequentially to the corresponding third safety gap and join the first sub-queue. At the same time, the longitudinal trajectory planning model based on adaptive model predictive control is used to adjust the longitudinal distance of the CAV vehicles in the updated first sub-queue, thereby updating the corresponding second and third safety gaps, until all CAV vehicles in the second and third sub-queues have completed the lane change.

[0008] Furthermore, the first safety gap is the maximum gap between the target lane that meets the lane-changing safety conditions and the HDV vehicle adjacent to the upper CAV queue. Specifically, the lane-changing safety conditions are that the actual gap between the current CAV vehicle and the vehicle in front of the gap is not less than the minimum safety gap, and when the current CAV vehicle performs a lane-changing action, the response acceleration of the vehicle behind the gap is not less than the maximum deceleration. Determining the first safety gap of the target lane based on the minimum safety gap and the maximum deceleration, and selecting CAV vehicles in the CAV queue that can change lanes to the first safety gap to form the first sub-queue, includes the following steps: Set up CAV queue The number of vehicles is ,Right now , As the lead vehicle in the queue, The second car in the queue, and so on. For the last vehicle in the queue, from the lead vehicle of the CAV queue Start vehicle-by-vehicle inspection, select the HDV preceding and following vehicles that are closest to the CAV vehicle being inspected in the target lane, and use the gap between the HDV preceding and HDV following vehicles as the target gap. Determine whether the inspected CAV vehicle simultaneously meets the lane-changing safety conditions for both the vehicle in front of the target gap HDV and the vehicle behind the target gap HDV. If it does, it is added to the first sub-queue, and the number of vehicles in the first sub-queue is [number missing]. The lead car is The first sub-queue is The first sub-queue performs lane changing synchronously; the remaining vehicles that have not changed lanes are then assigned to the second sub-queue based on their position relative to the first sub-queue. The second sub-queue is then... Vehicles that have not changed lanes will be assigned to the third sub-queue. The third sub-queue is... .

[0009] Furthermore, the mathematical expression for the lane-changing safety condition of the vehicle ahead of the target clearance HDV is as follows:

[0010]

[0011]

[0012] in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The recent example of HDV; This is the current time step; for and The actual distance; for and The minimum safe distance; When t The longitudinal position; When t The longitudinal position; L For vehicle body length, It is the time step during the entire lane change process. t A set; for hour The required parameters for performing a lane change action Minimum safe distance between; for hour speed; for The reaction time; for The maximum deceleration; for hour speed; for hour The maximum deceleration.

[0013] Furthermore, the mathematical expression for the lane-changing safety conditions of the vehicle following the HDV with respect to the target clearance is as follows:

[0014]

[0015]

[0016] in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The latest HDV follow-up, for hour When performing lane changing behavior The response acceleration; yes The maximum acceleration; yes speed; It is a vehicle and The actual distance; for The maximum deceleration; Minimum following distance; This represents the desired time interval.

[0017] Furthermore, the second safety gap is the target gap between the lead CAV vehicle of the first sub-queue and the HDV in front of it when the desired gap value is achieved, and the third safety gap is the target gap between the last two CAV vehicles of the first sub-queue when the desired gap value is achieved. The longitudinal trajectory planning model based on adaptive model predictive control includes the following steps: Let the prediction time domain be In the prediction time domain The objective function of the first sub-queue is constructed internally, which causes the lead CAV vehicle in the first sub-queue to increase the distance between itself and the HDV in front of it, forming the objective gap of the second sub-queue, and the two CAV vehicles at the end of the first sub-queue form the objective gap of the third sub-queue. In the prediction time domain The code constructs objective functions for the second and third sub-queues, ensuring that when the gap in the first sub-queue widens, the second and third sub-queues reach positions where they can enter the corresponding objective gaps. Specifically, this means the leader vehicle of the second sub-queue... Regulation and The spacing ensures the leader's vehicle The location satisfies the condition of The minimum safe clearance; the second sub-queue of following vehicles continues to move forward according to CACC, maintaining a safe distance. The same acceleration is used to maintain queue stability; the lead vehicle of the third sub-queue Adjustment in the first subqueue The spacing ensures the leader's vehicle The location satisfies the condition of The minimum safe distance; the following vehicles in the third sub-queue continue to move forward according to the CACC method, maintaining the minimum safe distance. Acceleration synchronization is used to maintain queue stability.

[0018] Furthermore, the mathematical expression for the objective function of the first sub-queue is as follows:

[0019] in, It is the total cost of the first sub-queue's driving comfort and the urgency of lane changes; It is the number of prediction steps; To control the input weights; for The set of accelerations of all vehicles in the first subqueue at time 1; For gap weights; yes The target gap is prepared for the second sub-queue; yes The target gap is prepared for the third sub-queue; This is the maximum safe clearance required for the second sub-queue to change lanes to the target lane, expressed mathematically as follows: when hour,

[0020] when hour,

[0021] This is the maximum safe clearance required for the third sub-queue to change lanes to the target lane, expressed mathematically as follows:

[0022] in, yes hour and Minimum safe distance between; yes hour and Minimum safe distance between; yes hour and The minimum safe distance; yes hour and Minimum safe distance between; yes hour and The minimum safe distance between them.

[0023] Furthermore, the mathematical expressions for the objective functions of the second and third sub-queues are as follows:

[0024]

[0025] in, and These represent the total costs of optimizing the driving comfort and lane-changing urgency of the second and third sub-queues, respectively. for The acceleration of the lead vehicle in the second sub-group at any given moment; for The acceleration of the lead vehicle in the third sub-group at any given moment; It is the lead vehicle in the second sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the second sub-queue. and The actual distance; It is the lead vehicle in the third sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the third sub-queue. and The actual spacing.

[0026] Furthermore, establishing a longitudinal trajectory planning model based on adaptive model predictive control also includes: establishing constraints, the mathematical expression of which is as follows:

[0027]

[0028]

[0029] in, , , For the current moment The longitudinal position, velocity, and acceleration; and They are respectively Maximum acceleration and deceleration; and These are the vehicle's maximum speed and minimum speed, respectively. yes The vehicles behind Current vertical position To minimize following distance, L This refers to the vehicle's length.

[0030] Furthermore, when the initial position required for the vehicle to smoothly enter the ramp is calculated by backpropagation based on the fifth-order polynomial trajectory planning method, and then embedded as the ramp end constraint into the adaptive model predictive control, the following steps are included: Trajectory planning is performed using a longitudinal trajectory planning model based on adaptive model predictive control without ramp end constraints. The longitudinal position, velocity, and acceleration of vehicles in each sub-queue changing lanes to the target lane are obtained and used as initial conditions. The mathematical expression is as follows:

[0031] in, The coefficients of the polynomial, , , For queue The longitudinal position, velocity, and acceleration; For lane changing time; Set the final state velocity to the same value as the initial velocity, set the acceleration to 0, and set the final state position to the end of the ramp deceleration lane. The mathematical expression is as follows:

[0032] in, For the CAV queue CAV vehicles The timing, longitudinal position, speed, and acceleration of the lane change to the target lane. They are respectively The starting moment of merging from the target lane to the deceleration lane, including the longitudinal position, speed, and acceleration at the start of merging; They are respectively The longitudinal position, velocity, and acceleration at the end of the merging process from the target lane to the deceleration lane; Location of the ramp divergence point. The velocity is the same as the initial velocity. Based on the velocity and acceleration constraints, the optimal initial position is obtained using a fifth-order polynomial iteration. As the minimum merging starting point, the position where the CAV vehicle ends its lane change in the target lane. If the position is less than the minimum merging starting point, the vehicle can successfully exit the ramp, generating the corresponding ramp end constraint and adding it to the AMPC controller.

[0033] Furthermore, when establishing the longitudinal trajectory planning model based on adaptive model predictive control, it also includes: dynamically adjusting the weights of the control input term and the backlash error term, as expressed mathematically below:

[0034]

[0035]

[0036]

[0037]

[0038] in, for time The distance between the location and the ramp divergence point; Location of the ramp divergence point; This represents the relative progress of vehicles from the ramp divergence point. To control the input, The weighting factor for the gap; It is a parameter that controls the growth rate of the gap weight; It is a parameter that controls the rate of change of driving comfort factors; For gap weights; The initial weights for the gaps; To control the input weights; To control the initial weights of the input.

[0039] Compared with the prior art, the advantages of the present invention are as follows: This invention divides the lane-changing process into three stages: "lane-changing of leading vehicles - active gap adjustment - lane-changing of remaining vehicles". It achieves collaborative optimization at the queue level, making up for the shortcomings of existing lane-changing and control methods in ramp diversion areas, which mostly focus on independent or local multi-vehicle coordination and lack a multi-stage collaborative and multi-objective real-time optimization scheme covering the entire lane-changing process.

[0040] This invention embeds ramp end constraints into adaptive model predictive control, enabling real-time coupling of longitudinal distance adjustment and lane-changing decisions. This significantly improves lane-changing efficiency, queue stability, and ride comfort, and compensates for the lack of applicability of existing research to the special scenario of the diversion zone where space is limited and lane-changing is urgent. Attached Figure Description

[0041] Figure 1 This is a flowchart of an embodiment of the present invention; Figure 2 This is a schematic diagram of a scene in an embodiment of the present invention; Figure 3 These are diagrams showing the changes in CAV queues and HDV trajectories under different methods in embodiments of the present invention. Figure 4 These are graphs showing the changes in CAV acceleration under different methods in the embodiments of the present invention; Figure 5 These are vehicle trajectories under different fleet sizes in embodiments of the present invention.

[0042] Figure 6 This is a diagram showing the trajectory changes of the CAV queue under different acceleration limits in an embodiment of the present invention.

[0043] Figure 7 This refers to the changes in CAV queue and HDV velocity under different acceleration limits in the embodiments of the present invention. Detailed Implementation

[0044] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0045] This embodiment focuses on the ramp diversion area of ​​highways. From the perspective of optimizing the coordinated lane-changing behavior of CAV queues, it proposes a three-stage coordinated lane-changing method for vehicle queues in ramp diversion areas under intelligent connected environments. The lane-changing process is decomposed into three stages: first, lane changing by vehicles that meet lane-changing safety conditions; second, gap adjustment; and third, lane changing by remaining vehicles. Through the adaptive model predictive control (AMPC) framework with ramp end constraints, multi-objective coordinated optimization of lane-changing urgency, queue stability, and comfort is achieved. Figure 1 As shown, the method includes the following steps: S1) Based on the minimum safe distance and maximum deceleration, determine the target lane safety gap, select the sequence of vehicles that will change lanes first to form a leading queue, and complete the lane change. Specifically: The safety gap of the target lane is determined based on the minimum safe distance and the maximum deceleration. In this embodiment, it is referred to as the first safety gap for distinction. CAV vehicles that can change lanes to the first safety gap are selected from the CAV queue to form a first sub-queue. The first sub-queue is controlled to change lanes to the first safety gap. CAV vehicles in the CAV queue before the first sub-queue form a second sub-queue, and CAV vehicles in the CAV queue after the first sub-queue form a third sub-queue. S2) Construct a trajectory planning model based on adaptive model predictive control, adjust the longitudinal distance between vehicles in the leading platoon, and actively create a safe gap in the target lane, specifically: A longitudinal trajectory planning model based on adaptive model predictive control is established. The longitudinal distance of CAV vehicles in the first sub-queue is adjusted using the longitudinal trajectory planning model based on adaptive model predictive control. At the same time, considering that the CAV queue must eventually enter the exit ramp within a limited distance, the initial position required for the vehicle to successfully enter the ramp is calculated in reverse based on the fifth-order polynomial trajectory planning method. This initial position is then embedded into the adaptive model predictive control as a ramp end constraint, thereby forming a new safety gap for the target lane. In this embodiment, the safety gap of the second sub-queue is referred to as the second safety gap for distinction, and the safety gap of the third sub-queue is referred to as the third safety gap for distinction. S3) Control the CAV vehicles in the CAV queue that have not yet changed lanes to change lanes sequentially into the new safe gap in the target lane, completing the lane change. Specifically: The vehicles in the second sub-queue are controlled to change lanes sequentially to the corresponding second safety gap, and the vehicles in the third sub-queue are controlled to change lanes sequentially to the corresponding third safety gap and join the first sub-queue. At the same time, the longitudinal trajectory planning model based on adaptive model predictive control is used to adjust the longitudinal distance of the CAV vehicles in the updated first sub-queue, thereby updating the corresponding second and third safety gaps, until all CAV vehicles in the second and third sub-queues have completed the lane change.

[0046] By following the steps above, we can ensure that the entire CAV queue changes lanes safely, orderly, and in formation to the target lane and can enter the exit ramp in a timely manner.

[0047] The following reference Figure 2 The above steps will be explained in detail. Figure 2 The scenario shown is a merging zone 900 meters upstream from the exit ramp of a two-lane highway. The CAVL lane, GL lane, and deceleration lane are adjacent in sequence, with one branch consisting of... m A convoy of CAVs When driving in the CAVL lane, the adjacent GL lane is used as the target lane, and the end of the GL lane is the ramp divergence point.

[0048] In this embodiment, step S1 involves finding the maximum gap between adjacent HDVs in lane GL during the CAV platoon's movement, which is then used as the current target gap. Since this target gap is insufficient to accommodate the entire platoon, a subset of CAVs in the platoon that can preferentially switch lanes to this target gap needs to be determined based on lane-changing safety conditions, forming a sub-platoon that will prioritize lane-changing. The number of vehicles is recorded as The leader's vehicle number is Sub-team Priority lane change to the GL lane, while the remaining vehicles continue in the CAVL lane, forming two sub-queues. and Set up a CAV queue. The number of vehicles is .Right now , As the lead vehicle in the convoy, The second car in the queue, and so on. This is the last vehicle in the queue. , , , .

[0049] At the same time, the lane-changing safety conditions that should be followed take into account both the minimum safe clearance and the maximum deceleration: 1) The vehicle in front of the target gap, if in the queue and The actual spacing is not less than and The minimum safe distance, then When changing lanes, you will not face the same situation as in the GL lane. The risk of collision. and The actual distance is Front end and The distance between backends. and The minimum safety clearance is obtained using the Gipps safety clearance algorithm to ensure... At the moment of changing lanes to lane GL and There exists a critical safety clearance, expressed as follows: (1) (2) (3) in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The recent example of HDV; This is the current time step; for and The actual distance; for and The minimum safe distance; for t hour The longitudinal position; for t hour The longitudinal position; L For vehicle body length, It is the time step during the entire lane change process. t A set; for hour The required parameters for performing a lane change action Minimum safe distance between; for hour speed; for The reaction time; for The maximum deceleration; for hour speed; for hour The maximum deceleration.

[0050] 2) The vehicle behind in the target lane, if When performing lane changing behavior The response acceleration is not less than The maximum deceleration, then The lane-changing behavior will not face the same situation as in the GL lane. The risk of a collision is expressed as follows: (4) time When performing lane changing behavior The response acceleration is calculated using the following formula. The IDM model can be used to describe the driving behavior of the HDV and predict the state of the CAV. The response acceleration. for according to The safe distance is dynamically calculated based on the motion state and the user's own speed: (5) (6) in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The latest HDV follow-up, for hour When performing lane changing behavior The response acceleration; yes The maximum acceleration; yes speed; It is a vehicle and The actual distance; for The maximum deceleration; Minimum following distance; This represents the desired time interval.

[0051] As can be seen from the above, the first safe gap in this embodiment is the maximum gap between the target lane that meets the lane-changing safety conditions and the HDV vehicle adjacent to the upper CAV queue. Specifically, the lane-changing safety conditions are that the actual gap between the current CAV vehicle and the vehicle in front of the gap is not less than the minimum safe distance, and when the current CAV vehicle performs a lane-changing action, the response acceleration of the vehicle behind the gap is not less than the maximum deceleration. Based on the minimum safe distance and the maximum deceleration, the first safe gap of the target lane is determined. Therefore, in step S1, when selecting CAV vehicles in the CAV queue that can change lanes to the first safe gap to form the first sub-queue, the following steps are included: S11) From the lead vehicle of the CAV queue Start vehicle-by-vehicle inspection, select the HDV preceding and following vehicles that are closest to the CAV vehicle being inspected in the target lane, and use the gap between the HDV preceding and HDV following vehicles as the target gap. S12) Determine whether the inspected CAV vehicle simultaneously meets the lane-changing safety conditions with both the vehicle in front of the target gap HDV and the vehicle behind the target gap HDV. If it does, add it to the first sub-queue. The lane-changing safety conditions of the vehicle ahead of the target gap HDV are satisfied by the aforementioned formulas (1) to (3), and the lane-changing safety conditions of the vehicle behind the target gap HDV are satisfied by the aforementioned formulas (4) to (6).

[0052] The first sub-queue can be expanded through the checks and judgments in step S12. Expand the first subqueue During the process, if a vehicle that does not meet the safety requirements for lane changing is encountered, the first sub-queue... Stop expanding; at this point, the first sub-queue... The number of vehicles is The lead car is , First subqueue Lane changing is performed synchronously; other vehicles in the CAV queue that have not changed lanes are then determined based on their position relative to the first queue. The position will be the first sub-team Vehicles that have not yet changed lanes will be assigned to the second sub-queue. Then the second subqueue The first sub-team Vehicles that did not change lanes behind them were assigned to the third sub-queue. The third sub-queue is .

[0053] This embodiment uses step S2 to precisely plan the longitudinal trajectory, creating sufficient safety gaps between the GL and CAVL lanes for sub-platforms that have not yet performed lane-changing operations, ensuring smooth subsequent lane changes. Specifically, the sub-platform closest to the GL lane... and Two gaps were selected as the target gaps: the second subqueue. Will enter the first sub-queue The gap between the lead vehicle and the HDV in front of it The third subqueue It will enter the end of the sub-queue. The gap between the two vehicles .when At that time, sub-queue and All entered the gap In this process, this embodiment constructs a longitudinal trajectory planning model based on adaptive model predictive control (AMPC), which enables the CAVs in the CAVL lane to adjust their longitudinal acceleration according to the real-time traffic conditions, ensuring that the target gap they form reaches the desired gap value in the prediction time domain, while also enabling vehicles that are about to change lanes to reach the position to enter the target gap.

[0054] As can be seen from the above, in this embodiment, the second safety gap is the target gap between the leader CAV vehicle of the first sub-queue and the HDV in front of it when the desired gap value is achieved, and the third safety gap is the target gap between the last two CAV vehicles of the first sub-queue when the desired gap value is achieved. The longitudinal trajectory planning model based on adaptive model predictive control includes the following steps: S21) Dynamically adjust the weights of the control input and clearance error terms to balance the safety and comfort of platoon driving. Specifically, as the lead vehicle in the platoon gets closer to the split point, the weight of the clearance error term increases to strengthen the control over the safe distance; at the same time, the weight of the control input term decreases, thereby providing greater flexibility for speed adjustment in critical areas. The mathematical expression is as follows: (7) (8) (9) (10) (11) in, for time The distance between the location and the ramp divergence point; Location of the ramp divergence point; This represents the relative progress of vehicles from the ramp divergence point. To control the input, The weighting factor for the gap; It is a parameter that controls the growth rate of the gap weight; It is a parameter that controls the rate of change of driving comfort factors; For gap weights; The initial weights for the gaps; To control the input weights; To control the initial weights of the input.

[0055] S22) Consider the first subqueue It needs to be adjusted vertically to form the second sub-queue. and the third subqueue The lane-changing execution creates sufficient safety gaps. Specifically, this manifests as: the first sub-queue The lead car Need to expand with The spacing forms a second sub-queue. target gap First subqueue The two CAVs at the middle and rear need to form a third sub-platform. target gap .

[0056] To achieve synergistic optimization of driving comfort and lane-changing urgency, this embodiment sets the prediction time domain as follows: In the prediction time domain The objective function for constructing the first sub-queue is such that the leader CAV vehicle in the first sub-queue... Increase the distance between it and the HDV in front to form the target gap of the second sub-queue, and the two CAV vehicles at the rear of the first sub-queue. and The objective gap for forming the third sub-queue; the mathematical expression for the objective function of the first sub-queue is as follows: (12) in, It is the total cost of the first sub-queue's driving comfort and the urgency of lane changes; It is the number of prediction steps; To control the input weights; for The set of accelerations of all vehicles in the first subqueue at time 1; For gap weights; yes The target gap is prepared for the second sub-queue; yes The target gap is prepared for the third sub-queue.

[0057] This is the maximum safe clearance required for the second sub-queue to change lanes to the target lane, expressed mathematically as follows: when hour, (13) when hour, (14) This is the maximum safe clearance required for the third sub-queue to change lanes to the target lane, expressed mathematically as follows: (15) in, yes hour and Minimum safe distance between; yes hour and Minimum safe distance between; yes CAV and CAV The minimum safe distance; yes hour and Minimum safe distance between; yes hour and The minimum safe distance between them.

[0058] S23) The lead CAV vehicle in the first sub-queue Expand its front HDV The spacing forms the target gap prepared for the second sub-queue, and the two CAV vehicles at the tail end of the first sub-queue and This creates a target gap for the third sub-queue. For the second and third sub-queues to enter the gap prepared by the first sub-queue, the vehicles in the second and third sub-queues also need to move forward to a suitable position so that they can switch lanes to the corresponding target gap.

[0059] To achieve synergistic optimization of driving comfort and lane-change urgency, this embodiment focuses on prediction in the time domain. The objective function for constructing the second and third sub-queues is defined to make the first sub-queue... When widening the gap, the second sub-queue and the third subqueue It is possible to reach and enter the corresponding target gap. , The location.

[0060] Specifically: the second sub-queue The lead car Regulation and The spacing ensures the leader's vehicle The position after changing lanes satisfies the condition of... Minimum safe distance The following vehicle continues to move forward according to CACC, maintaining a safe distance. Maintaining the same acceleration to keep the queue stable, reaching the point where it can enter the target gap. Location; Third subqueue The lead car Regulation and The spacing ensures the leader's vehicle The position after changing lanes satisfies the condition of... Minimum safe distance The following vehicle continues to move forward according to the CACC method, maintaining a safe distance from the vehicle. The acceleration is synchronized to maintain queue stability, and it also allows entry into the target gap. The location.

[0061] The mathematical expressions for the objective functions of the second and third sub-queues are as follows: (16) (17) in, and These represent the total costs of optimizing the driving comfort and lane-changing urgency of the second and third sub-queues, respectively. for The acceleration of the lead vehicle in the second sub-group at any given moment; for The acceleration of the lead vehicle in the third sub-group at any given moment; It is the lead vehicle in the second sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the second sub-queue. and The actual distance; It is the lead vehicle in the third sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the third sub-queue. and The actual spacing.

[0062] S24) Establish constraints regarding the sub-team , and In the longitudinal trajectory planning model, the constraints consist of kinematic constraints, acceleration and speed limits, safe distance between vehicles in the queue, and minimum safe exit position, as expressed mathematically below: (18) (19) (20) in, , , For the current moment The longitudinal position, velocity, and acceleration; and They are respectively Maximum acceleration and deceleration; and These are the vehicle's maximum speed and minimum speed, respectively. yes The vehicles behind The current vertical position.

[0063] In step S2 of this embodiment, to ensure that the coordinated lane-changing behavior of the CAV queue meets the timely lane-changing constraint from GL to the deceleration lane or diversion point, a lane-changing starting point calculation method based on the back-calculation of a fifth-order polynomial trajectory is proposed to obtain the minimum safe exit position. This position is then embedded into the adaptive MPC controller to construct an end-constraint optimization framework. Specifically, when the initial position required for the vehicle to smoothly enter the ramp is calculated back-calculated based on the fifth-order polynomial trajectory planning method and embedded as the ramp end constraint in the adaptive model predictive control, the following steps are included: Trajectory planning is performed using a longitudinal trajectory planning model based on adaptive model predictive control without ramp end constraints. This yields the longitudinal position, velocity, and acceleration of vehicles in each sub-queue changing lanes to the target lane, which serve as initial conditions. Based on the trajectory planning model, the sub-queue... and The sub-team can successfully change lanes to lane GL (the outermost lane). At this point, the sub-team... , and After changing lanes to lane GL, vehicles must then change lanes again from lane GL to the deceleration lane before the ramp divergence point to successfully exit the ramp. Each sub-vehicle group's position, velocity, and acceleration after changing lanes according to the trajectory planning model are known. Assuming that a vehicle in the queue completes its lane change from lane GL to the deceleration lane just before the end of the deceleration lane, the vehicle can successfully exit the ramp. The mathematical expressions for the longitudinal position, velocity, and acceleration of a vehicle in the queue upon changing lanes to the target lane are as follows: (twenty one) in, The coefficients of the polynomial, , , For queue The longitudinal position, velocity, and acceleration; This is the time for changing lanes.

[0064] Set the initial velocity of the ending state to the initial velocity, the acceleration to 0, and the position of the ending state to the end of the ramp deceleration lane. Construct the equations for the start and end states of the lane change, and the mathematical expression is as follows: (twenty two) in, For the CAV queue CAV vehicles The timing, longitudinal position, speed, and acceleration of the lane change to the target lane. They are respectively The starting moment of merging from the target lane to the deceleration lane, including the longitudinal position, speed, and acceleration at the start of merging; They are respectively The longitudinal position, velocity, and acceleration at the end of the merging process from the target lane to the deceleration lane; Location of the ramp divergence point. The speed is the same as the initial speed. In this embodiment, the location of the ramp divergence point... Lane change time is Set to 5 seconds.

[0065] Based on the velocity and acceleration constraints in formulas (18) and (19), the optimal initial position is obtained using the fifth-order polynomial iteration of formula (21). If CAV changes lanes from CAVL to the position where GL ends... If the position is less than the minimum merging starting point, the vehicle can successfully exit the ramp. Therefore, As the minimum merging starting point, the corresponding ramp end constraint is generated and added to the MPC controller, as shown in the following mathematical expression: (twenty three) in, The critical safe exit position required for the CAV to smoothly enter the ramp before the ramp divergence point; For vehicles changing lanes in the CAV queue The endpoint when changing lanes from CAVL lane to GL lane; for The moment to change lanes to lane GL.

[0066] In this embodiment, step S3 performs lane-changing control on the vehicles in the three sub-queues based on the longitudinal trajectory planning model established in step S2. Each time the MPC controller updates the control quantity, the vehicle states in the three sub-queues are updated accordingly, and the two target gaps formed in the second stage also change accordingly. Once the second sub-queue... and the third subqueue If a CAV meets the lane-changing safety conditions, it will enter the gap. and With the second sub-queue and the third subqueue CAVs successively switched lanes to the first sub-queue. During the creation interval, the first sub-queue The size of the first subqueue gradually increased, while the second subqueue... and the third subqueue The size of the queue will decrease accordingly. During this process, the queue position and index of each vehicle will also be dynamically updated. When the second sub-queue... and the third subqueue After all vehicles in the queue have completed lane changing and successfully joined the sub-queue, the second stage, the gap adjustment stage, and the third stage, the remaining vehicle lane changing stage, terminate simultaneously, and the queue... P Lane change complete.

[0067] The following study selects the ramp divergence area as the research scenario. Simulation comparison experiments and parameter sensitivity analyses are conducted using Python software to verify the effectiveness and stability of the method in this embodiment. In the experimental scenario, the main line has two lanes, CAVL and GL, with a minimum speed limit of 60 km / h and a maximum speed limit of 120 km / h. The origin of the main line is taken as the coordinate axis, and the length is 900 meters. The main line origin is 900 meters from the divergence point. In the CAVL lane, there is a queue of 5 CAVs. , , , , The initial positions are 411m, 381m, 351m, 321m, and 291m. There are two HDVs in lane GL, with positions of 411m and 311m respectively.

[0068] First, a comparative experiment was conducted to evaluate the lane-changing efficiency and comfort of different methods. Lane-changing efficiency was mainly assessed through lane-changing success rate and lane-changing completion time; comfort was analyzed using the root mean square error of acceleration. The following strategies were selected: Strategy 1: PID-based lane-changing control method, which optimizes the longitudinal trajectory of the CAV queue based on the PID controller to complete the lane change; Strategy Two: The uncontrolled method, where each vehicle in the convoy independently finds a suitable gap in the target lane and completes the lane change operation one by one. This method does not involve coordinated convoy behavior and does not have a phased lane-changing process; Strategy 3: A fleet coordination lane-changing method based on traditional MPC, which uses a traditional MPC controller to optimize the acceleration of the queue to adjust the lane-changing spacing during the longitudinal spacing adjustment phase; Strategy 4: The three-stage collaborative lane-changing method proposed in this embodiment uses an adaptive MPC controller that considers the constraints at the end of the ramp to optimize the acceleration of the queue in order to adjust the lane-changing spacing during the longitudinal spacing adjustment stage.

[0069] Figure 3 Showing when mWhen the queue size is 5, the longitudinal trajectories of the queue and the vehicles before and after the HDV are shown under four strategies. The blue solid line represents the longitudinal trajectory of the HDV on GL; the green solid line represents the longitudinal trajectory of the CAV in the queue on CAVL; and the orange dashed line represents the longitudinal trajectory of the CAV in the queue on GL. When the CAV changes lanes, the green solid line representing the longitudinal trajectory turns into an orange dashed line. Table 1 compares the queue lane-changing completion status under the four lane-changing strategies. Within the same simulation time, under strategy 1, 2 vehicles in the queue successfully changed lanes before the ramp divergence point, with a lane-changing completion rate of 40%; under strategy 2, 4 vehicles in the queue completed lane-changing before the ramp divergence point, with a lane-changing completion rate of 80%; under strategy 3, at 9.5s, all 5 vehicles in the queue completed lane-changing before the ramp divergence point, with a lane-changing success rate of 100%. Under strategy 4, at 8s, all 5 vehicles in the queue completed lane-changing before the ramp divergence point, with a lane-changing success rate of 100%. The lane change success rate was improved by 60% and 20% compared to Strategy 1 and Strategy 2, respectively; the lane change completion time was reduced by 27.2%, 27.2% and 15.8% compared to Strategy 1, Strategy 2 and Strategy 3, respectively.

[0070] Table 1 CAV Lane Change Information Table

[0071] Figure 4 The graph shows the CAV acceleration variation under different strategies. It can be seen that under all three strategies, the acceleration varies within the range of [-4, 4], and all exhibit some degree of fluctuation. Table 2 shows the root mean square (RMS) acceleration of each vehicle in the CAV queue under different lane-changing strategies. The results show that the RMS accelerations of the vehicles in the queue under strategies 1, 2, and 3 are 2.51, 2.79, 2.93, and 2.31, respectively. Under strategy 4, the RMS value is 21.19% lower than strategy 3, 17.14% lower than strategy 2, and 7.84% lower than strategy 1. Although strategy 1 uses a PID controller to optimize the longitudinal trajectory, its lane-changing process lacks lateral coordination within the queue. When entering the target lane, a significant adjustment to the acceleration of each vehicle is still required, resulting in a horizontally centered RMS acceleration. In strategy 2, each CAV independently seeks gaps and changes lanes one by one, ignoring overall queue coordination. The lack of synchronization between vehicles leads to frequent and large acceleration and braking actions, resulting in the highest acceleration fluctuation and the worst comfort. Strategy 4 dynamically adjusts gap weights through an adaptive MPC controller, allowing vehicles to flexibly adjust acceleration based on real-time traffic conditions during lane changes, avoiding excessive acceleration or deceleration and thus reducing acceleration fluctuations. In contrast, while Strategy 3 can also adjust acceleration in real time, it fails to adjust gap weights, resulting in poorer adaptability to traffic changes and larger acceleration fluctuations.

[0072] Table 2. Acceleration RMSE (Root Mean Square Error) values ​​for the CAV queue

[0073] Then, a systematic sensitivity analysis was conducted on the parameter platoon size and acceleration constraints to examine the model's response characteristics under different platoon sizes, dynamic constraints and surrounding traffic environments, and to further verify that the invention can maintain adaptability and stability under various conditions.

[0074] Figure 5 The diagram illustrates the three-dimensional trajectory changes of each CAV (Cross-Action Vehicle) when the platoon size is 5, 6, and 8 vehicles, respectively. It is evident that the vehicles complete an orderly lane change from the CAVL lane to the GL lane before the ramp divergence point, demonstrating the adaptability and stability of this control method in complex platoon coordination tasks. When =5, after all vehicles in the queue have changed lanes, the queue... , , , , The positions on the GL lane are 666m, 592m, 518m, 568m, and 554m respectively; When =6, after all vehicles in the queue have changed lanes, the queue... , , , , The positions on lane GL are 639m, 621m, 513m, 560m, 549m, and 538m respectively; When =8, after all vehicles in the queue have changed lanes, the queue... , , , , The positions on the GL lane are 743m, 725m, 592m, 661m, 651m, 639m, 629m, and 616m, respectively. Regardless of whether the queue size is 5, 6, or 8 vehicles, this strategy can still effectively coordinate the acceleration and deceleration of each vehicle, keeping the dynamic spacing above the safe threshold. All vehicles in the queue can complete the orderly lane change before the ramp divergence point, and the final position of each vehicle on the GL lane is always less than the minimum safe exit position (700m, 750m, 800m) required to smoothly enter the ramp before the divergence point.

[0075] Figure 6 and Figure 7 The paper demonstrates the changing trends of vehicle trajectory and velocity over time in the method of this embodiment under different acceleration constraints. The results show that when the acceleration range is [-4, 4], the vehicle that first meets the lane-changing safety conditions and changes lanes is... and , This is the last vehicle to change lanes; after all vehicles in the queue have changed lanes to lane GL, , , , , The positions are 666m, 592m, 518m, 568m, and 554m respectively. When the acceleration range is [-3, 6], the vehicle that first meets the lane-changing safety conditions and changes lanes is... and , This is the last vehicle to change lanes; after all vehicles in the queue have changed lanes to lane GL, , , , , The positions are 691m, 616m, 543m, 589m, and 567m respectively. When the acceleration range is [-2, 2], the vehicle that first meets the lane-changing safety conditions and changes lanes is... and , This is the last vehicle to change lanes; after all vehicles in the queue have changed lanes to lane GL, , , , , The positions were 704m, 621m, 534m, 573m, and 561m, respectively. The final position of each vehicle on lane GL was consistently less than the minimum safe exit position required to smoothly enter the ramp before the merging point. Although the upper and lower limits of acceleration affected lane-changing time and speed fluctuations, under three different dynamic constraints, all vehicles were able to complete lane changes in an orderly and smooth manner and meet their respective safe exit thresholds, without any abnormalities such as trajectory crossings or sudden acceleration / deceleration. These results fully verify the robustness and adaptability of the method in this embodiment under various acceleration constraints.

[0076] In summary, this invention proposes a three-stage collaborative lane-changing method for vehicle queuing in a ramp diversion area under intelligent connected vehicle conditions. In the first stage, the target lane safety gap is determined based on the minimum safe distance and maximum deceleration, and a sequence of vehicles prior to lane changing is selected to form a leading queue, completing the lane change. In the second stage, considering that the CAV queue ultimately needs to enter the exit ramp within a limited distance, the initial position required for smooth entry into the ramp is calculated backward using a fifth-order polynomial trajectory planning method. This constraint is embedded into adaptive model predictive control to adjust the longitudinal distance between vehicles in the queue, forming a new safety gap. In the third stage, vehicles in the queue that have not yet changed lanes sequentially change lanes to the safety gap in the target lane, completing the lane change. Simultaneously, the ramp diversion area is selected as the research scenario, and simulation comparison experiments and parameter sensitivity analyses are conducted using Python software to verify the effectiveness and stability of the method. Compared with existing technologies, the advantages of this invention are: 1. Existing technologies focus on a two-lane section of a highway, only mentioning that they can recommend the latest lane-changing start position for a convoy under different traffic conditions, reminding the convoy to change lanes in time, but without specific methods or experiments. This invention, however, focuses on a 900-meter upstream branching zone from the exit ramp of a two-lane highway. Based on a fifth-order polynomial backpropagation, it calculates the minimum safe exit position required for vehicles in the convoy to smoothly enter the ramp before the branching point. This position constraint is embedded into an MPC-based trajectory planning model. According to this constraint, it can be ensured that all vehicles in the convoy, after changing to the outermost lane, can smoothly enter the ramp before the branching point.

[0077] 2. Existing technologies use fixed parameters, making it difficult to dynamically adjust them according to real-time traffic conditions. Compared to existing technologies, this invention balances queuing safety and comfort by dynamically adjusting the weights of control input terms and gap error terms. In the MPC-based rolling optimization process, the weight settings of each part of the objective function have a significant impact on solving for the optimal control quantity.

[0078] 3. Compared with existing technologies, this invention conducts comparative experiments by analyzing the lane-changing efficiency and comfort of independent lane-changing methods, PID-based lane-changing control methods, and three-stage cooperative lane-changing methods. Lane-changing efficiency is mainly evaluated by lane-changing success rate; comfort is analyzed by the root mean square error of acceleration; at the same time, a systematic sensitivity analysis was carried out for parameter platoon size and acceleration constraints to examine the response characteristics of the model under different platoon sizes, dynamic constraints, and surrounding traffic environments, and further verify that this invention can maintain adaptability and stability under various conditions.

[0079] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A three-stage coordinated lane-changing method for vehicle queuing in a ramp diversion area under an intelligent connected environment, characterized in that, Includes the following steps: The first safety gap of the target lane is determined based on the minimum safe distance and the maximum deceleration. CAV vehicles that can change lanes to the first safety gap are selected from the CAV queue to form a first sub-queue. The first sub-queue is controlled to change lanes to the first safety gap. CAV vehicles in the CAV queue before the first sub-queue form a second sub-queue, and CAV vehicles in the CAV queue after the first sub-queue form a third sub-queue. A longitudinal trajectory planning model based on adaptive model predictive control is established. The longitudinal distance of CAV vehicles in the first sub-queue is adjusted using the longitudinal trajectory planning model based on adaptive model predictive control. At the same time, the initial position required for the vehicle to smoothly enter the ramp is calculated in reverse based on the fifth-order polynomial trajectory planning method, and is embedded into the adaptive model predictive control as the ramp end constraint to form the second and third safety gaps of the target lane. The vehicles in the second sub-queue are controlled to change lanes sequentially to the corresponding second safety gap, and the vehicles in the third sub-queue are controlled to change lanes sequentially to the corresponding third safety gap and join the first sub-queue. At the same time, the longitudinal trajectory planning model based on adaptive model predictive control is used to adjust the longitudinal distance of the CAV vehicles in the updated first sub-queue, thereby updating the corresponding second and third safety gaps, until all CAV vehicles in the second and third sub-queues have completed the lane change.

2. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 1, characterized in that, The first safety gap is the maximum gap between the target lane that meets the lane-changing safety conditions and the HDV vehicle adjacent to the upper CAV queue. Specifically, the lane-changing safety conditions are that the actual gap between the current CAV vehicle and the vehicle in front of the gap is not less than the minimum safety gap, and when the current CAV vehicle performs a lane-changing action, the response acceleration of the vehicle behind the gap is not less than the maximum deceleration. Based on the minimum safety gap and maximum deceleration, the first safety gap of the target lane is determined. When selecting CAV vehicles in the CAV queue that can change lanes to the first safety gap to form the first sub-queue, Includes the following steps: Set up CAV queue The number of vehicles is ,Right now , As the lead vehicle in the queue, The second car in the queue, and so on. For the last vehicle in the queue, from the lead vehicle of the CAV queue Begin vehicle-by-vehicle inspection, selecting the HDV vehicles closest to the inspected CAV in the target lane and the HDV vehicles following it. The gap between these HDV vehicles is taken as the target gap. Determine if the inspected CAV simultaneously meets the lane-changing safety conditions with both the HDV vehicle in front and the HDV vehicle behind it. If so, add it to the first sub-queue. The number of vehicles in the first sub-queue is [number missing]. The lead car is The first sub-queue is The first sub-queue performs lane changing synchronously; the remaining vehicles that have not changed lanes are then assigned to the second sub-queue based on their position relative to the first sub-queue. The second sub-queue is then... Vehicles that have not changed lanes will be assigned to the third sub-queue. The third sub-queue is... .

3. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 2, characterized in that, The mathematical expression for the lane-changing safety condition of the vehicle ahead of the target with a clearance of HDV is as follows: in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The recent example of HDV; This is the current time step; for and The actual distance; for and The minimum safe distance; When t The longitudinal position; When t The longitudinal position; L For vehicle body length, It is the set of time steps t throughout the entire lane-changing process; for hour The required parameters for performing a lane change action Minimum safe distance between; for hour speed; for The reaction time; for The maximum deceleration; for hour speed; for hour The maximum deceleration.

4. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 3, characterized in that, The mathematical expression for the lane-changing safety condition of the vehicle following the target with a clearance of HDV is as follows: in, It is the [number]th [item] in the CAV queue CAV; Is the target lane away The latest HDV follow-up, for hour When performing lane changing behavior The response acceleration; yes The maximum acceleration; yes speed; It is a vehicle and The actual distance; for The maximum deceleration; Minimum following distance; This represents the desired time interval.

5. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 4, characterized in that, The second safety gap is the target gap between the lead CAV vehicle of the first sub-queue and the HDV in front of it when the desired gap value is achieved. The third safety gap is the target gap between the last two CAV vehicles of the first sub-queue when the desired gap value is achieved. When establishing the longitudinal trajectory planning model based on adaptive model predictive control, the following steps are included: Let the prediction time domain be In the prediction time domain The objective function of the first sub-queue is constructed internally, which causes the lead CAV vehicle in the first sub-queue to increase the distance between itself and the HDV in front of it, forming the objective gap of the second sub-queue, and the two CAV vehicles at the end of the first sub-queue form the objective gap of the third sub-queue. In the prediction time domain The code constructs objective functions for the second and third sub-queues, ensuring that when the gap in the first sub-queue widens, the second and third sub-queues reach positions where they can enter the corresponding objective gaps. Specifically, this means the leader vehicle of the second sub-queue... Regulation and The spacing ensures the leader's vehicle The location satisfies the condition of The minimum safe clearance; the second sub-queue of following vehicles continues to move forward according to CACC, maintaining a safe distance. The same acceleration is used to maintain queue stability; the lead vehicle of the third sub-queue Adjustment in the first subqueue The spacing ensures the leader's vehicle The location satisfies the condition of The minimum safe distance; the following vehicles in the third sub-queue continue to move forward according to the CACC method, maintaining the minimum safe distance. Acceleration synchronization is used to maintain queue stability.

6. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 5, characterized in that, The mathematical expression for the objective function of the first sub-queue is as follows: in, It is the total cost of the first sub-queue's driving comfort and the urgency of lane changes; It is the number of prediction steps; To control the input weights; for The set of accelerations of all vehicles in the first subqueue at time 1; For gap weights; yes The target gap is prepared for the second sub-queue; yes The target gap is prepared for the third sub-queue; This is the maximum safe clearance required for the second sub-queue to change lanes to the target lane, expressed mathematically as follows: when hour, when hour, This is the maximum safe clearance required for the third sub-queue to change lanes to the target lane, expressed mathematically as follows: in, yes hour and Minimum safe distance between; yes hour and Minimum safe distance between; yes hour and The minimum safe distance; yes hour and Minimum safe distance between; yes hour and The minimum safe distance between them.

7. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 6, characterized in that, The mathematical expressions for the objective functions of the second and third sub-queues are as follows: in, and These represent the total costs of optimizing the driving comfort and lane-changing urgency of the second and third sub-queues, respectively. for The acceleration of the lead vehicle in the second sub-group at any given moment; for The acceleration of the lead vehicle in the third sub-group at any given moment; It is the lead vehicle in the second sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the second sub-queue. and The actual distance; It is the lead vehicle in the third sub-queue. Need and The minimum safe distance to maintain It is the lead vehicle in the third sub-queue. and The actual spacing.

8. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 5, characterized in that, When establishing a longitudinal trajectory planning model based on adaptive model predictive control, the following steps are also included: establishing constraints, the mathematical expressions of which are as follows: in, , , For the current moment The longitudinal position, velocity, and acceleration; and They are respectively Maximum acceleration and deceleration; and These are the vehicle's maximum speed and minimum speed, respectively. yes The vehicles behind Current vertical position To minimize following distance, L This refers to the vehicle's length.

9. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 5, characterized in that, When the initial position required for a vehicle to smoothly enter the ramp is calculated using a fifth-order polynomial trajectory planning method and embedded as a ramp end constraint in adaptive model predictive control, the following steps are included: Trajectory planning is performed using a longitudinal trajectory planning model based on adaptive model predictive control without ramp end constraints. The longitudinal position, velocity, and acceleration of vehicles in each sub-queue changing lanes to the target lane are obtained and used as initial conditions. The mathematical expression is as follows: in, The coefficients of the polynomial, , , For queue The longitudinal position, velocity, and acceleration; For lane changing time; Set the final state velocity to the same value as the initial velocity, set the acceleration to 0, and set the final state position to the end of the ramp deceleration lane. The mathematical expression is as follows: in, For the CAV queue CAV vehicles The timing, longitudinal position, speed, and acceleration of the lane change to the target lane. They are respectively The starting moment of merging from the target lane to the deceleration lane, including the longitudinal position, speed, and acceleration at the start of merging; They are respectively The longitudinal position, velocity, and acceleration at the end of the merging process from the target lane to the deceleration lane; This is the location of the ramp divergence point. The velocity is the same as the initial velocity. Based on the velocity and acceleration constraints, the optimal initial position is obtained using a fifth-order polynomial iteration. As the minimum merging starting point, the position where the CAV vehicle ends its lane change in the target lane. If the position is less than the minimum merging starting point, the vehicle can successfully exit the ramp, generating the corresponding ramp end constraint and adding it to the AMPC controller.

10. The three-stage coordinated lane-changing method for vehicle queuing in the ramp diversion area under intelligent connected environment according to claim 6 or 7, characterized in that, When establishing a longitudinal trajectory planning model based on adaptive model predictive control, it also includes: dynamically adjusting the weights of the control input term and the backlash error term, as expressed mathematically below: in, for time The distance between the location and the ramp divergence point; Location of the ramp divergence point; This represents the relative progress of vehicles from the ramp divergence point. To control the input, The weighting factor for the gap; It is a parameter that controls the growth rate of the gap weight; It is a parameter that controls the rate of change of driving comfort factors; For gap weights; The initial weights for the gaps; To control the input weights; To control the initial weights of the input.

Images