Off-ramp connection area automatic driving exclusive lane and signal timing joint optimization method

Abstract

The present application relates to a kind of under-ramp connection area automatic driving special lane and signal timing joint optimization method, comprising the following steps: S1, the physical road of expressway under-ramp and intersection connection area is divided into two-dimensional space grid, is abstracted into directed network diagram;S2, the dynamic traffic capacity based on grid and section mapping attribute is established, and the improved space impedance calculation model considering lateral lane changing penalty is established;S3, the optimization upper decision model is established to minimize road network system total travel time and intersection signal delay;S4, under the premise of given special lane layout and signal timing, mixed vehicle seeks lower path selection distribution model;S5, obtain the optimization scheme of automatic driving special lane two-dimensional space layout and signal timing, and the special lane two-dimensional space layout scheme is released.The present application cooperates and optimizes intersection signal timing, effectively solves the complex congestion problem that special lane is set too early or too late to hinder ramp vehicle lateral interweave.

Description

Technical Field

[0001] This invention relates to the technical field of vehicle-to-everything (V2X) and autonomous driving traffic control, and more specifically, to a method for jointly optimizing the autonomous driving lane and signal timing in the off-ramp connection area. Background Technology

[0002] For a considerable period to come, mixed traffic flows consisting of human-driven vehicles (HDVs) and connected and automated vehicles (CAVs) will coexist on existing transportation infrastructure. The junctions between expressway off-ramps and intersections are typical traffic bottlenecks due to frequent lane changes and weaving. The mixing of HDVs and CAVs exacerbates right-of-way conflicts in these areas, easily causing queues to spill over onto the main road. Current technologies typically deploy dedicated lanes for automated CAVs statically across the entire road segment. However, in off-ramp merging areas, setting up dedicated lanes too early or too late hinders lateral weaving of vehicles from the ramps. Neither spatial isolation nor signal timing alone can resolve the complex congestion in these areas. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method for joint optimization of the timing of autonomous driving lanes and signals in the connection area of ​​off-ramp. This method effectively solves the complex congestion problem caused by setting up dedicated lanes too early or too late, which hinders the lateral weaving of vehicles on ramps. It has important theoretical significance and engineering value.

[0004] The technical solution adopted by this invention to solve its technical problem is: constructing a joint optimization method for autonomous driving lanes and signal timing in off-ramp connection areas, including the following steps: S1. Divide the physical roads in the connection area between the expressway off-ramp and the intersection into a two-dimensional spatiotemporal grid and abstract them into a directed network graph; S2. Establish an improved spatiotemporal impedance calculation model based on the dynamic traffic capacity of grid and road segment mapping attributes and considering the penalty for lateral lane changes. S3. Establish a high-level decision model that optimizes the one-dimensional time allocation of the two-dimensional spatial location of dedicated lanes and signal timing, with the goal of minimizing the total travel time of the road network system and the intersection signal delay. S4. Under the premise of given dedicated lane layout and signal timing, a lower-level path selection and allocation model for mixed vehicles seeks the minimum cost. S5. The above two-layer model is solved by nesting heuristic algorithm and successive averaging method to obtain the two-dimensional spatial layout and signal timing optimization scheme of the dedicated lane for autonomous driving. The signal timing scheme is then sent to the intersection signal controller, and the two-dimensional spatial layout scheme of the dedicated lane is sent to the roadside unit or connected vehicle terminal for publication.

[0005] According to the above scheme, the method of abstracting into a directed network graph in step S1 includes the following steps: S101. Perform lateral dimension partitioning, dividing the road into a set of lanes. The innermost lane is lane 1, and the outermost lane is lane 2. lane, and the first The lane includes the entrance / exit ramp; S102. Perform longitudinal dimension division. To ensure initial lane-changing space for vehicles merging into the main line from the off-ramp, a safety buffer section is reserved downstream of the physical merging point of the off-ramp. The distance from the end of this safety buffer section to the stop line of the downstream intersection is divided into... n A set of equal-length grids ; S103. Perform node and edge partitioning, treating each spatiotemporal grid in the road network as a node, defined as follows: , , The straight-through and lane-changing behaviors between units are defined as directed edges. Divided into from arrive The straight edge and from arrive The lane change side.

[0006] According to the above scheme, the method for establishing the improved spatiotemporal impedance calculation model in step S2 includes the following steps: S201, Define Mesh Whether it is a dedicated channel indicator variable ; S202, Calculate the edge Dynamic access capability ; S203, Calculate the edge On-the-go driving costs .

[0007] According to the above scheme, in step S201, a mesh is defined. Whether it is a dedicated channel indicator variable :

[0008] In the formula: This is a lateral decision variable; a value of 1 indicates a lane. When set to a dedicated lane, a value of 0 indicates a mixed-traffic lane. , Representing lanes The starting and ending numbers of the dedicated lane in the vertical grid; This is a conditional characteristic function, taking the value 1 if the condition within the parentheses is true, and 0 otherwise; define the edge. The dedicated channel attribute is This property depends on the edge. The attribute of the downstream mesh node it points to, i.e., when the edge Pointing to node hour, .

[0009] According to the above scheme, in step S202, the edges are calculated. Dynamic access capability :

[0010]

[0011] In the formula: The average headway of mixed traffic flow affected by different vehicle combinations on the mixed lane; The saturation headway for connected and autonomous vehicles following connected and autonomous vehicles; The saturation headway for connected autonomous vehicles (CAVs) to follow human-driven vehicles; The saturation headway between human-driven vehicles and connected autonomous vehicles; The saturation headway for human-driven vehicles following human-driven vehicles. This represents the current penetration rate of connected autonomous vehicles (CAVs).

[0012] According to the above scheme, in step S203, the edges are calculated. On-the-go driving costs :

[0013] In the formula: They are respectively assigned to the edges Traffic flow of human-driven vehicles and connected autonomous vehicles; For the edge The physical length; For free flow velocity; Additional delay penalty time will be added for lateral lane changes; The value is 1 if and only if the edge is a lateral lane-changing edge; otherwise, the value is 0. , The coefficients to be calibrated.

[0014] According to the above scheme, the specific mathematical expression of the upper-level spatiotemporal resource allocation model in step S3 is as follows:

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] In the formula: For the downstream intersection signal phase set, where This represents the total number of phases. Assign a signal phase index number; To merge into the first Total arrival flow of the phase; Belongs to phase The set of all allowed import edges; Queuing delay time should take into account signal period and phase green ratio; The signal period; For the first The green light ratio of the phase; For the first Phase saturation; Intersection phase The saturation flow rate (pcu / h) for the corresponding lane; , These are the minimum period and the maximum period, respectively. This results in a loss of time; The minimum green light duration for a single phase to ensure the safe passage of pedestrians or clearing of vehicles.

[0026] According to the above scheme, in step S4, the mathematical expression of the lower-level hybrid vehicle path selection model is:

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] In the formula: For the dimensionless virtual integral variable in the integral operation of the objective function; For the edge The driving cost function on; Let O be the set of all OD pairs in the road network; For set One of the characteristics of OD pairs; For OD The set of all feasible paths between; For set A certain path; , They represent OD pairs respectively Traveling between routes Traffic flow of human-driven HDV vehicles and connected autonomous CAV vehicles; , OD pairs respectively w Total demand for human-driven vehicles (HDVs) and OD (Original Demand) w Total demand for connected autonomous vehicles (CAVs) on the Internet; For path-segment association variables, if the edge In the path If it is above, it is 1; otherwise, it is 0. It is a very large positive number used for associativity. Construct constraint boundaries to forcibly and exclusively prevent human-driven HDV vehicles from being assigned to dedicated road segments for autonomous CAVs.

[0034] According to the above scheme, the method implemented in step S5 includes the following steps: S501. A genetic algorithm with real-integer hybrid encoding is used to solve the upper-level decision model. Integer encoding represents the state variables of each lane in the lateral direction and their independent longitudinal start-end point index, while real-integer encoding represents the signal period and green ratio. S502, Lower-level traffic assignment is solved using the successive average method, based on the data input from the upper level. The matrix adjusts the road segment impedance, and the shortest paths for human-driven vehicles (HDV) and connected autonomous vehicles (CAV) are searched based on Dijkstra's algorithm, respectively. S503, double-layer iteration until flow convergence, output the optimal dedicated channel horizontal and vertical two-dimensional spatial position, signal period and green signal ratio of each phase; S504. Send the signal timing scheme to the intersection signal controller and send the two-dimensional spatial layout scheme of the dedicated lane to the roadside unit or connected vehicle terminal for publication.

[0035] This invention also provides a joint optimization system for autonomous driving lanes and signal timing in off-ramp connection areas, comprising: The directed network graph module is used to divide the physical roads of the connection area between expressway off-ramp and intersection into a two-dimensional spatiotemporal grid and abstract them into a directed network graph. An improved spatiotemporal impedance calculation model module is used to establish an improved spatiotemporal impedance calculation model based on grid and road segment mapping attributes, which considers the dynamic traffic capacity and lateral lane change penalty. The upper-level decision model module is used to establish an upper-level decision model that optimizes the one-dimensional time allocation of the two-dimensional spatial location of dedicated lanes and signal timing with the goal of minimizing the total travel time of the road network system and the intersection signal delay. The lower-level route selection and allocation model module is used to establish a lower-level route selection and allocation model for mixed vehicles to seek the minimum cost under the premise of given dedicated lane layout and signal timing. The signal timing scheme module is used to solve the above two-layer model by nesting heuristic algorithms and successive averaging methods to obtain the two-dimensional spatial layout and signal timing optimization scheme of the dedicated lane for autonomous driving. The signal timing scheme is then sent to the intersection signal controller, and the two-dimensional spatial layout scheme of the dedicated lane is sent to the roadside unit or connected vehicle terminal for publication.

[0036] The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to the present invention has the following beneficial effects: This invention breaks through the limitations of traditional static dedicated lanes across the entire road segment. Based on spatiotemporal grid modeling, it transforms the setting of dedicated lanes into a two-dimensional layout problem, allowing longitudinal space for vehicles to change lanes and weave on off-ramp. Through a two-layer collaborative optimization model, it realizes the dynamic joint configuration of space (dedicated lane layout) and time (signal timing), which can adaptively adjust according to the penetration rate of different connected autonomous vehicles, maximizing the solution to the bottleneck congestion problem on off-ramp. Attached Figure Description

[0037] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a flowchart of the method for joint optimization of autonomous driving lanes and signal timing in the off-ramp connection area according to the present invention; Figure 2 This is a schematic diagram of the layout of dedicated lanes for autonomous driving (CAV) in the grid-based road segmentation of this invention. Detailed Implementation

[0038] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0039] Example The area connecting the off-ramp of a certain expressway and the intersection includes, laterally... 3 lanes on the main road The ramp has one lane and is longitudinally divided into... A grid. Setting basic headway parameters, the connected autonomous vehicle follows the connected autonomous vehicle. s, Connected autonomous vehicles follow human-driven vehicles s, Human-driven vehicles and connected autonomous vehicles are s, human-driven vehicles follow human-driven vehicles as The proposed method is optimized for different penetration rates of connected autonomous vehicles.

[0040] like Figure 1-2 As shown, the method for joint optimization of autonomous driving lanes and signal timing in the off-ramp connection area of ​​the present invention includes the following steps: S1. Divide the physical roads in the connection area between the expressway off-ramp and the intersection into a two-dimensional spatiotemporal grid, and abstract it into a directed network graph. Specifically, this includes the following steps: S101, Lateral Dimension Division: Dividing the road into sets of lanes. The innermost lane is lane 1, and the outermost lane is lane 4, with lane 4 including the off-ramp merging lane. S102. Longitudinal Dimension Division: To ensure initial lane-changing space for vehicles merging into the main line from the off-ramp, a safety buffer section is reserved downstream of the physical merging point of the off-ramp. The distance from the end of this safety buffer section to the stop line of the downstream intersection is divided into... n A set of equal-length grids ; S103, Node and Edge Partitioning: Each spatiotemporal grid in the road network is considered a node, defined as follows: , , The straight-through and lane-changing behaviors between units are defined as directed edges. Divided into from arrive The straight edge and from arrive For the lane change side, the longitudinal length of a grid is set to 20 meters.

[0041] S2. Establish an improved spatiotemporal impedance calculation model based on the dynamic traffic capacity of grid and road segment mapping attributes, and considering lateral lane-changing penalties. This includes the following steps: S201, Define Mesh Whether it is a dedicated channel indicator variable :

[0042] In the formula: This is a lateral decision variable; a value of 1 indicates a lane. When set to a dedicated lane, a value of 0 indicates a mixed-traffic lane. , Representing lanes The starting and ending numbers of the dedicated lane in the vertical grid; This is a conditional characteristic function; it takes the value 1 if the condition within the parentheses is true, and 0 otherwise. Define the edge. The dedicated channel attribute is This property depends on the edge. The attribute of the downstream mesh node it points to, i.e., when the edge Pointing to node hour, .

[0043] S202, calculate the edge using the following formula. Dynamic access capability :

[0044]

[0045] In the formula: The average headway of mixed traffic flow affected by different vehicle combinations on the mixed lane; This represents the current penetration rate of connected and autonomous vehicles.

[0046] S203, Calculate the edge On-the-go driving costs :

[0047] In the formula: They are respectively assigned to the edges Traffic from human-driven vehicles (HDV) and connected autonomous vehicles (CAV) on the internet; For the edge The physical length; For the free flow velocity, this embodiment takes... 60km / h; Add a delay penalty time for lateral lane changes, set to 0.25s; The value is 1 if and only if the edge is a lateral lane-changing edge; otherwise, the value is 0. , For the coefficients to be calibrated, take... , .

[0048] S3. Establish a higher-level decision-making model that optimizes the one-dimensional time allocation of the two-dimensional spatial location and signal timing of dedicated lanes, with the goal of minimizing the total travel time of the road network system and intersection signal delays. The specific mathematical expression of the higher-level spatiotemporal resource allocation model is as follows:

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059] In the formula: For the downstream intersection signal phase set, where This represents the total number of phases. Assign a signal phase index number; To merge into the first Total arrival flow of the phase; Belongs to phase The set of all allowed import edges; Queuing delay time should take into account signal period and phase green ratio; The signal period; For the first The green light ratio of the phase; For the first Phase saturation; Intersection phase The saturation flow rate (pcu / h) for the corresponding lane; , These are the minimum period and the maximum period, respectively. This results in a loss of time; The minimum green light duration for a single phase to ensure the safe passage of pedestrians or clearing of vehicles.

[0060] S4. Based on the given dedicated lane layout and signal timing, a lower-level path selection and allocation model for mixed vehicles seeks the minimum cost. The specific mathematical expression is as follows:

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067] In the formula: For the dimensionless virtual integral variable in the integral operation of the objective function; For the edge The driving cost function on; Let O(OD) be the set of all OD pairs in the road network; For set One of the characteristics of OD pairs; For OD The set of all feasible paths between; For set A certain path; , They represent OD pairs Traveling between routes Traffic flow of human-driven HDV vehicles and connected autonomous CAV vehicles; , OD pairs respectively w Total demand for human-driven vehicles (HDVs) and OD (Original Demand) w Total demand for connected autonomous vehicles (CAVs) on the Internet; For path-segment association variables, if the edge In the path If it is above, it is 1; otherwise, it is 0. It is a very large positive number used for associativity. Construct constraint boundaries to forcibly and exclusively prevent human-driven HDV vehicles from being assigned to dedicated road segments for autonomous CAVs.

[0068] S5. The above two-layer model is solved using a nested heuristic algorithm and successive averaging method to obtain the optimized two-dimensional spatial layout and signal timing scheme for the dedicated autonomous driving lanes. The signal timing scheme is then distributed to the intersection signal controller, and the two-dimensional spatial layout scheme for the dedicated lanes is sent to the roadside unit or connected vehicle terminal for publication. Specifically, this includes the following steps: S501. Set the downstream intersection to use typical four-phase control for main and secondary flows (phase 1 for eastbound entry, phase 2 for westbound entry, phase 3 for southbound entry, and phase 4 for northbound entry, i.e., the total number of phases). In the MATLAB environment, the upper layer uses a genetic algorithm (GA) with mixed real-integer encoding, and the chromosome design is as follows:

[0069] S502, The lower layer uses the successive average method (MSA) to solve the problem, with a step size of [missing information] for each iteration. .

[0070] S503. Given a 70% CAV (Consumer-Aided Vehicle) penetration rate, output the optimal two-dimensional spatial positions of the dedicated lane, signal period, and green light ratio of each phase. The joint optimization scheme obtained is shown in Tables 1 and 2. Table 1 Deployment Plan for Dedicated Automated Driving Lanes

[0071] Table 2 Signal Timing Scheme

[0072] S504. Send the signal timing scheme to the intersection signal controller and send the two-dimensional spatial layout scheme of the dedicated lane to the roadside unit or connected vehicle terminal for publication.

[0073] This invention also provides a joint optimization system for autonomous driving lanes and signal timing in off-ramp connection areas, comprising: a directed network graph module for dividing the physical roads of the connection area between expressway off-ramp and intersection into a two-dimensional spatiotemporal grid and abstracting it into a directed network graph; an improved spatiotemporal impedance calculation model module for establishing an improved spatiotemporal impedance calculation model based on the dynamic capacity of the grid and road segment mapping attributes and considering lateral lane-changing penalties; an upper-level decision model module for establishing an upper-level decision model that optimizes the one-dimensional time allocation of the two-dimensional spatial position of the dedicated lane and signal timing with the goal of minimizing the total travel time of the road network system and intersection signal delay; a lower-level path selection and allocation model module for establishing a lower-level path selection and allocation model for mixed vehicles seeking the minimum cost under the premise of given dedicated lane layout and signal timing; and a signal timing scheme module for using a heuristic algorithm and a successive averaging method to solve the above two-level model, obtaining an optimized scheme for the two-dimensional spatial layout and signal timing of the autonomous driving lane, and distributing the signal timing scheme to the intersection signal controller and sending the two-dimensional spatial layout scheme of the dedicated lane to the roadside unit or connected vehicle terminal for publication.

[0074] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas, characterized in that, Includes the following steps: S1. Divide the physical roads in the connection area between the expressway off-ramp and the intersection into a two-dimensional spatiotemporal grid and abstract them into a directed network graph; S2. Establish an improved spatiotemporal impedance calculation model based on the dynamic traffic capacity of grid and road segment mapping attributes and considering the penalty for lateral lane changes. S3. Establish a high-level decision model that optimizes the one-dimensional time allocation of the two-dimensional spatial location of dedicated lanes and signal timing, with the goal of minimizing the total travel time of the road network system and the intersection signal delay. S4. Under the premise of given dedicated lane layout and signal timing, a lower-level path selection and allocation model for mixed vehicles seeks the minimum cost. S5. The above two-layer model is solved by nesting heuristic algorithm and successive averaging method to obtain the two-dimensional spatial layout and signal timing optimization scheme of the dedicated lane for autonomous driving. The signal timing scheme is then sent to the intersection signal controller, and the two-dimensional spatial layout scheme of the dedicated lane is sent to the roadside unit or connected vehicle terminal for publication.

2. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 1, characterized in that, In step S1, the method of abstracting into a directed network graph includes the following steps: S101. Perform lateral dimension partitioning, dividing the road into a set of lanes. The innermost lane is lane 1, and the outermost lane is lane 2. lane, and the first The lane includes the entrance / exit ramp; S102. Perform longitudinal dimension division. To ensure initial lane-changing space for vehicles merging into the main line from the off-ramp, a safety buffer section is reserved downstream of the physical merging point of the off-ramp. The distance from the end of this safety buffer section to the stop line of the downstream intersection is divided into... n A set of equal-length grids ; S103. Perform node and edge partitioning, treating each spatiotemporal grid in the road network as a node, defined as follows: , , The straight-through and lane-changing behaviors between units are defined as directed edges. Divided into from arrive The straight edge and from arrive The lane change side.

3. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 1, characterized in that, In step S2, the method for establishing the improved spatiotemporal impedance calculation model includes the following steps: S201, Define Mesh Whether it is a dedicated channel indicator variable ; S202, Calculate the edge Dynamic access capability ; S203, Calculate the edge On-the-go driving costs .

4. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 3, characterized in that, In step S201, a mesh is defined. Whether it is a dedicated channel indicator variable : In the formula: This is a lateral decision variable; a value of 1 indicates a lane. When set to a dedicated lane, a value of 0 indicates a mixed-traffic lane. , Representing lanes The starting and ending numbers of the dedicated lane in the vertical grid; This is a conditional characteristic function, taking the value 1 if the condition within the parentheses is true, and 0 otherwise; define the edge. The dedicated channel attribute is This property depends on the edge. The attribute of the downstream mesh node it points to, i.e., when the edge Pointing to node hour, .

5. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 4, characterized in that, In step S202, the edges are calculated. Dynamic access capability : In the formula: The average headway of mixed traffic flow affected by different vehicle combinations on the mixed lane; The saturation headway for connected and autonomous vehicles following connected and autonomous vehicles; The saturation headway for connected autonomous vehicles (CAVs) to follow human-driven vehicles; The saturation headway between human-driven vehicles and connected autonomous vehicles; The saturation headway for human-driven vehicles following human-driven vehicles. This represents the current penetration rate of connected autonomous vehicles (CAVs).

6. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 5, characterized in that, In step S203, the edges are calculated. On-the-go driving costs : In the formula: They are respectively assigned to the edges Traffic flow of human-driven vehicles and connected autonomous vehicles; For the edge The physical length; For free flow velocity; Additional delay penalty time will be added for lateral lane changes; The value is 1 if and only if the edge is a lateral lane-changing edge; otherwise, the value is 0. , The coefficients to be calibrated.

7. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 1, characterized in that, In step S3, the specific mathematical expression of the upper-level spatiotemporal resource allocation model is as follows: In the formula: For the downstream intersection signal phase set, where This represents the total number of phases. Assign a signal phase index number; To merge into the first Total arrival flow of the phase; Belongs to phase The set of all allowed import edges; Queuing delay time should take into account signal period and phase green ratio; The signal period; For the first The green light ratio of the phase; For the first Phase saturation; Intersection phase The saturation flow rate (pcu / h) for the corresponding lane; , These are the minimum period and the maximum period, respectively. This results in a loss of time; The minimum green light duration for a single phase to ensure the safe passage of pedestrians or clearing of vehicles.

8. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 1, characterized in that, In step S4, the mathematical expression for the lower-level hybrid vehicle path selection model is: In the formula: For the dimensionless virtual integral variable in the integral operation of the objective function; For the edge The driving cost function on; Let O be the set of all OD pairs in the road network; For set One of the characteristics of OD pairs; For OD The set of all feasible paths between; For set A certain path; , They represent OD pairs respectively Traveling between routes Traffic flow of human-driven HDV vehicles and connected autonomous CAV vehicles; , OD pairs respectively w Total demand for human-driven vehicles (HDVs) and OD (Original Demand) w Total demand for connected autonomous vehicles (CAVs) on the Internet; For path-segment association variables, if the edge In the path If it is above, it is 1; otherwise, it is 0. It is a very large positive number used for associativity. Construct constraint boundaries to forcibly and exclusively prevent human-driven HDV vehicles from being assigned to dedicated road segments for autonomous CAVs.

9. The method for joint optimization of autonomous driving lanes and signal timing in off-ramp connection areas according to claim 1, characterized in that, The method implemented in step S5 includes the following steps: S501. A genetic algorithm with real-integer hybrid encoding is used to solve the upper-level decision model. Integer encoding represents the state variables of each lane in the lateral direction and their independent longitudinal start-end point index, while real-integer encoding represents the signal period and green ratio. S502, Lower-level traffic assignment is solved using the successive average method, based on the data input from the upper level. The matrix adjusts the road segment impedance, and the shortest paths for human-driven vehicles (HDV) and connected autonomous vehicles (CAV) are searched based on Dijkstra's algorithm, respectively. S503, double-layer iteration until flow convergence, output the optimal dedicated channel horizontal and vertical two-dimensional spatial position, signal period and green signal ratio of each phase; S504. Send the signal timing scheme to the intersection signal controller and send the two-dimensional spatial layout scheme of the dedicated lane to the roadside unit or connected vehicle terminal for publication.

10. A joint optimization system for autonomous driving lanes and signal timing in off-ramp connection areas, characterized in that, include: The directed network graph module is used to divide the physical roads of the connection area between expressway off-ramp and intersection into a two-dimensional spatiotemporal grid and abstract them into a directed network graph. An improved spatiotemporal impedance calculation model module is used to establish an improved spatiotemporal impedance calculation model based on grid and road segment mapping attributes, which considers the dynamic traffic capacity and lateral lane change penalty. The upper-level decision model module is used to establish an upper-level decision model that optimizes the one-dimensional time allocation of the two-dimensional spatial location of dedicated lanes and signal timing with the goal of minimizing the total travel time of the road network system and the intersection signal delay. The lower-level route selection and allocation model module is used to establish a lower-level route selection and allocation model for mixed vehicles to seek the minimum cost under the premise of given dedicated lane layout and signal timing. The signal timing scheme module is used to solve the above two-layer model by nesting heuristic algorithms and successive averaging methods to obtain the two-dimensional spatial layout and signal timing optimization scheme of the dedicated lane for autonomous driving. The signal timing scheme is then sent to the intersection signal controller, and the two-dimensional spatial layout scheme of the dedicated lane is sent to the roadside unit or connected vehicle terminal for publication.

Images