Optimization Layout Method for Horizontal and Vertical Positions of Exclusive Lanes for Autonomous Driving in the Scenario of Human-Machine Mixed Driving

Abstract

The present invention relates to a method for optimizing the horizontal and vertical layout of an exclusive automatic driving lane in a human-machine co-driving scenario, which includes the following steps: S1. Abstract the sections and ramps of an expressway into a directed network graph, considering the basic lane sections, lane-changing connection sections, on-ramps, and off-ramps; S2. Divide the movement behaviors of vehicles on the expressway into car-following and lane-changing, and determine the cost functions of the two behaviors; S3. Obtain the OD matrix of vehicles on the on-ramps and off-ramps of the expressway; S4. Establish a mixed traffic flow distribution model considering the exclusive automatic driving lane; S5. Establish an optimization layout decision model for the exclusive automatic driving lane; S6. Solve to obtain the optimized horizontal and vertical layout scheme of the exclusive automatic driving lane. Based on the lane-level modeling of the expressway network, the present invention fully considers the car-following and lane-changing behaviors of vehicles on the expressway, and can jointly optimize the horizontal and vertical layout of the exclusive automatic driving lane, providing a new and scientific quantitative decision-making method.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle-road networking and autonomous driving traffic design, and particularly to a method for optimizing the horizontal and vertical layout of exclusive lanes for autonomous driving in a human-machine mixed driving scenario. Background Art

[0002] For a long time in the future, a traffic flow consisting of human-driven vehicles and autonomous driving vehicles will co-run on existing traffic infrastructure, such as urban expressways. As an important means of road right allocation in a human-machine mixed driving scenario, optimizing the layout of exclusive lanes for autonomous driving on expressways is the key to further improving the operation performance of the urban expressway system in a human-machine mixed driving scenario.

[0003] In the existing technologies for optimizing the layout of exclusive lanes for autonomous driving, the matching degree of traffic capacity and traffic flow is usually calculated to determine the number of exclusive lanes for autonomous driving. The conclusion only gives the relationship between the number of exclusive lanes for autonomous driving required for the basic sections of expressways and the penetration rate of autonomous driving vehicles, without considering the topological structure of the on-ramps and off-ramps on expressways and the influence of OD demand on the horizontal layout of exclusive lanes for autonomous driving. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method for optimizing the horizontal and vertical layout of exclusive lanes for autonomous driving in a human-machine mixed driving scenario, which can jointly optimize the horizontal and vertical layout of exclusive lanes for autonomous driving and provide a new and scientific quantitative decision-making method for setting exclusive lanes for autonomous driving on expressways.

[0005] The technical solution adopted by the present invention to solve its technical problems is to provide a method for optimizing the horizontal and vertical layout of exclusive lanes for autonomous driving in a human-machine mixed driving scenario, including the following steps:

[0006] S1. Abstract the sections and ramps of the expressway into a directed network graph, taking into account the basic lane sections, lane-changing connection sections, on-ramps, and off-ramps;

[0007] S2. Classify the movement behaviors of vehicles on the expressway into following and lane-changing, and determine the cost functions of the two movement behaviors respectively;

[0008] S3. Obtain the OD matrix of vehicles on the on-ramps and off-ramps of the expressway;

[0009] S4. Establish a mixed traffic flow distribution model considering exclusive lanes for autonomous driving;

[0010] S5. Establish a decision-making model for optimizing the layout of exclusive lanes for autonomous driving;

[0011] S6. Obtain the optimized layout scheme for the horizontal and vertical positions of the dedicated lane for autonomous driving.

[0012] According to the above scheme, step S1 includes the following steps:

[0013] S101. Abstract the basic lane sections, on-ramps, and off-ramps as one-way edges in the network graph, where each lane has two attributes: horizontal position l and longitudinal position a.

[0014] S102. Abstract the lane-changing connection sections between ramps and lanes, and between lanes as two-way edges in the network graph, numbered b.

[0015] S103. Abstract the starting point, ending point of the expressway, and the intersections of the on-ramps, off-ramps and the basic lane sections as nodes in the network graph.

[0016] S104. Connect the nodes and edges to form a directed network graph, which completes the modeling of the expressway facilities.

[0017] According to the above scheme, step S2 includes the following steps:

[0018] S201. Divide the behaviors of vehicles on the expressway into two categories: following and lane-changing. The following behavior refers to the driving behavior of a vehicle following the vehicle in front in a lane and maintaining a certain safety distance, and the following behavior corresponds to the driver's operations on the accelerator and brake. The lane-changing behavior refers to the behavior of a vehicle changing from one lane to another by finding a safe interval in the adjacent lane, and the lane-changing behavior corresponds to the driver's operation on the steering wheel.

[0019] S202. C a，l，MTF represents the single-lane passing capacity of the mixed human-machine traffic flow on lane a, l, and the calculation method is given by the following formula:

[0020] C a，l，MTF = C HV + (C AV - C HV )(P AV ) 2

[0021] In the formula: P AV is the penetration rate of autonomous vehicles in the lane; C HV represents the lane passing capacity when the penetration rate of human-driven cars is 100%, h HV is the saturated headway of human-driven vehicles; C AV represents the lane passing capacity when P AV is 100%, h AVis the saturated headway of an autonomous vehicle;

[0022] S203. Calculate the travel cost of lane following using the following formula:

[0023]

[0024] where: l and a are the numbers of the lateral and longitudinal positions of the lane respectively; t a，l is the travel cost of traveling on lanes a and l; t a，l，free represents the travel time of traveling on lanes a and l under free flow conditions; x a，l，HV represents the traffic flow of human-driven vehicles on lanes a and l; x a，l，AV represents the traffic flow of autonomous vehicles on lanes a and l; e and f are coefficients to be calibrated;

[0025] S204. Calculate the travel cost of lane changing using the following formula:

[0026] t b = K1(v out - v in ) 2 + K2d in + K3

[0027] where: v out is the driving speed of the lane before lane change; v in is the driving speed of the target lane; d in is the density of the target lane; K1, K2, and K3 are coefficients to be calibrated.

[0028] According to the above solution, step S3 includes the following steps:

[0029] S301. Use the video detectors installed at the on-ramps and off-ramps to record the numbers of the on-ramps and off-ramps of the vehicles;

[0030] S302. Statistically analyze the vehicle on-ramp and off-ramp conditions during the morning and evening rush hours on weekdays and on weekends respectively, to obtain the OD matrix of the expressway on weekdays and on weekends, where the OD matrix includes the starting point, ending point, on-ramps, and off-ramps of the expressway;

[0031] S303. Split the OD matrix into the OD HV of human-driven vehicles and the OD AV of autonomous vehicles according to the penetration rate of autonomous vehicles, and the calculation formula is as follows:

[0032] OD HV = OD total × (1 - P AV )

[0033] ODAV = OD total × PA V

[0034] According to the above solution, step S4 includes the following steps:

[0035] S401. Establish a mixed traffic flow distribution model considering dedicated lanes for autonomous driving. The specific mathematical expression is:

[0036]

[0037] Subject to:

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045] In the formula: represents the traffic flow of human-driven vehicles on the i-th path between OD pair r-s; represents the traffic flow of autonomous driving vehicles on the i-th path between OD pair r-s; is a decision variable, indicating whether a dedicated lane for autonomous driving is set up on lane a, l. When it is set up, it takes 1, and when it is not set up, it takes 0; M is a penalty term, taking a sufficiently large number.

[0046] According to the above solution, step S5 includes the following steps:

[0047] Establish an optimal layout decision model for dedicated lanes for autonomous driving. The specific mathematical expression is:

[0048]

[0049] Subject to:

[0050]

[0051] t b = K1(v out - v in ) 2 + K2d in + K3

[0052]

[0053]

[0054] where: x a，l，HV , x a，l，AV obtained from the mixed traffic flow distribution model in step S401, where TTT is the system travel time, is the non - dedicated road path count variable, is the path count variable after setting the dedicated lane, taking 1 when the path passes through the lane change numbered b, and 0 when not established.

[0055] According to the above solution, step S6 includes the following steps:

[0056] S601. Use a heuristic algorithm to solve the optimization layout decision model for the autonomous driving dedicated lane, where the mixed traffic flow distribution model is solved using the MATLAB optimization toolbox;

[0057] S602. Obtain the optimal Set the sections corresponding to the parts with equal to 1 as the dedicated lanes for autonomous driving vehicles, and the rest as mixed - traffic lanes.

[0058] According to the above solution, the heuristic algorithm is a genetic algorithm or a simulated annealing algorithm.

[0059] Implementing the method for optimizing the horizontal and vertical layout of the autonomous driving dedicated lane in the human - machine mixed driving scenario of the present invention has the following beneficial effects:

[0060] Based on the lane - level expressway network modeling, the present invention fully considers the car - following and lane - changing behaviors of vehicles on the expressway, avoiding the drawback that previous methods can only determine the number of autonomous driving lanes on the basic sections of the expressway, and can jointly optimize the horizontal and vertical layout of the autonomous driving dedicated lane, providing a new and scientific quantitative decision - making method for setting up the autonomous driving dedicated lane on the expressway. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is the flowchart of the method for optimizing the horizontal and vertical layout of the autonomous driving dedicated lane in the human - machine mixed driving scenario of the present invention;

[0062] Figure 2 is the schematic diagram of the expressway modeling method in the present invention;

[0063] Figure 3 is the schematic diagram of the layout of the autonomous driving dedicated lane on the expressway in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To have a clearer understanding of the technical features, objectives, and effects of the present invention, the specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0065] A section of an urban expressway contains a total of 7 on-ramps and 7 off-ramps. When the proportion of autonomous vehicles reaches a certain level, the method proposed in this application is adopted. Under the given condition of autonomous vehicle penetration rate, autonomous dedicated lanes are arranged in both horizontal and vertical directions to obtain the optimal road network operation performance, as Figure 1 shown, and specifically includes the following steps:

[0066] S1. Abstract the sections and ramps of the expressway into a directed topological network, and this process needs to consider the basic lane sections, lane-changing connection sections, on-ramps, and off-ramps simultaneously;

[0067] S101. Abstract the basic lane sections, on-ramps, and off-ramps into one-way edges in the network diagram, where each lane has two attributes, namely the lateral position l and the longitudinal position a;

[0068] S102. Abstract the lane-changing connection sections between ramps and lanes, and between lanes into two-way edges in the network diagram, numbered b;

[0069] S103. Abstract the starting point, ending point of the expressway, and the intersections of on-ramps, off-ramps with the basic lane sections into nodes in the network diagram;

[0070] S104. Connect the above-mentioned nodes and edges to form a directed network diagram, that is, the modeling of the expressway facilities is completed, as shown in the appendix Figure 2 shown.

[0071] S2. Divide the behaviors of vehicles on the expressway into two categories: following and lane-changing, and determine the cost functions of the two behaviors respectively;

[0072] S201. Divide the behaviors of vehicles on the expressway into two categories: following and lane-changing. Following means that the vehicle follows the vehicle in front in the lane and maintains a certain safety distance. The following behavior corresponds to the driver's operations on the accelerator and brake. Lane-changing means that the vehicle changes from one lane to another by finding a safe gap in the adjacent lane. The lane-changing behavior corresponds to the driver's operation on the steering wheel;

[0073] S202. C a，l，MTF represents the single-lane passing capacity of the mixed human-machine traffic flow on lane a, l, and its calculation method is given by the following formula:

[0074] C a，l，MTF = C HV +(C AV - CHV )(P AV ) 2

[0075] Where: P AV is the penetration rate of autonomous vehicles in the lane; C HV represents the lane capacity when the penetration rate of human-driven vehicles is 100%, h HV is the saturated headway of human-driven vehicles, which is 1 s; C AV represents the lane capacity when P AV is 100%, h AV is the saturated headway of autonomous vehicles, taking 2 s.

[0076] S203. Calculate the travel cost of lane following using the following formula:

[0077]

[0078] Where: l and a are the numbers of the lateral and longitudinal positions of the lane respectively; t a，l is the travel cost of lane a, l; t a，l，free represents the travel time on lane a, l under free flow conditions; x a，l，HV represents the flow of human-driven vehicles on lane a, l; x a，l，AV represents the flow of autonomous vehicles on lane a, l; e and f are coefficients to be calibrated.

[0079] S204. Calculate the travel cost of lane changing using the following formula:

[0080] t b = K1(v out - v in ) 2 + K2d in + K3

[0081] Where: v out is the driving speed of the lane before lane change; v in is the driving speed of the target lane; d in is the density of the target lane; K1, K2, and K3 are coefficients to be calibrated.

[0082] S3. Obtain the OD matrix of vehicles on the on-ramps and off-ramps of the expressway;

[0083] S301. Use the video detectors installed at the on-ramps and off-ramps to record the numbers of the on-ramps and off-ramps of the vehicles;

[0084] S302. Separately count the vehicle on-ramp and off-ramp situations during the morning and evening peak hours on weekdays (e.g., 7:00 - 8:30, 17:00 - 19:30) and on weekend rest days (e.g., 10:00 - 12:00) to obtain the OD matrices of the expressway on weekdays and weekend rest days. This OD matrix includes the starting point, ending point, on-ramp, and off-ramp of the expressway. The morning peak OD matrix is shown in Table 1:

[0085] Table 1 OD Matrix in the Morning Peak

[0086] <![CDATA[OD total > 3 5 6 9 11 13 15 16 1 180 206 95 474 377 380 318 1108 2 413 280 280 194 213 186 216 380 4 - 198 110 501 368 313 216 315 7 - - - 526 381 409 425 594 8 - - - 511 376 368 430 568 10 - - - - 632 398 325 500 12 - - - - - 735 545 966 14 - - - - - - 484 741

[0087] S303. Split the OD matrix according to the penetration rate of autonomous vehicles (in this example, P AV = 0.4) into the OD of human-driven vehicles HV and the OD of autonomous vehicles AV , and the calculation formula is as follows:

[0088] OD HV = OD total × (1 - P AV )

[0089] OD AV = OD total × P AV

[0090] The split morning peak OD matrix is shown in Tables 2 - 3:

[0091] Table 2 Split Morning Peak OD HV Matrix

[0092] <![CDATA[OD HV > 3 5 6 9 11 13 15 16 1 108 123.6 57 284.4 226.2 228 190.8 664.8 2 247.8 168 168 116.4 127.8 111.6 129.6 228 4 - 118.8 66 300.6 220.8 187.8 129.6 189 7 - - - 315.6 228.6 245.4 255 356.4 8 - - - 306.6 225.6 220.8 258 340.8 10 - - - - 379.2 238.8 195 300 12 - - - - - 441 327 579.6 14 - - - - - - 290.4 444.6

[0093] Table 3 Split Morning Peak OD AV Matrix

[0094] <![CDATA[OD AV > 3 5 6 9 11 13 15 16 1 72 82.4 38 189.6 150.8 152 127.2 443.2 2 165.2 112 112 77.6 85.2 74.4 86.4 152 4 - 79.2 44 200.4 147.2 125.2 86.4 126 7 - - - 210.4 152.4 163.6 170 237.6 8 - - - 204.4 150.4 147.2 172 227.2 10 - - - - 252.8 159.2 130 200 12 - - - - - 294 218 386.4 14 - - - - - - 193.6 296.4

[0095] S304. Establish a mixed traffic flow distribution model considering the setting of exclusive lanes for autonomous vehicles or exclusive lanes for human-driven vehicles;

[0096]

[0097] Subject to:

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105] Wherein: represents the traffic flow of human-driven vehicles on the i-th path between OD pairs r-s; represents the traffic flow of autonomous vehicles on the i-th path between OD pairs r-s; is a decision variable, indicating whether a dedicated lane for autonomous driving is set up in lane a, l. It takes 1 when set up and 0 when not set up; M is a penalty term, taking a sufficiently large number. Other parameters have the same meaning as those described above.

[0106] S4. Establish an optimization model for dedicated lanes for autonomous driving;

[0107] Establish an optimization layout decision model for dedicated lanes for autonomous driving. The specific mathematical expression is:

[0108]

[0109] Subject to:

[0110]

[0111] t b = K1(v out - v in ) 2 + K2d in + K3

[0112]

[0113]

[0114] (x a,l,HV , x a,l,AV ) is obtained from the mixed traffic flow distribution model (51)

[0115] Wherein: TTT is the system travel time, is the path count variable without setting up a dedicated road, is the path count variable after setting up a dedicated lane, taking 1 when the path passes through a lane change numbered b and 0 when not set up.

[0116] S5. Solve to obtain the combined optimization layout plan for the horizontal and vertical positions of dedicated lanes for autonomous driving.

[0117] S501. Solve the optimization layout decision model of the dedicated lane for autonomous driving using heuristic algorithms (such as genetic algorithms and simulated annealing algorithms), and solve the mixed traffic flow distribution model using the MATLAB optimization toolbox.

[0118] S502. Obtain the optimal Set the sections corresponding to the parts with a value of 1 as dedicated lanes for autonomous vehicles, and the rest as mixed traffic lanes. As shown in the attached

[0119]

[0120] figure, the positions of the corresponding dedicated lanes for autonomous driving. Figure 3 As shown in the attached figure, the positions of the corresponding dedicated lanes for autonomous driving.

[0121] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit and scope protected by the present invention and the claims. All of these are within the protection scope of the present invention.

Claims

1. An optimization layout method for the horizontal and vertical positions of an exclusive automatic driving lane in a human-machine co-driving scenario, characterized in that It includes the following steps: S1. Abstract the sections and ramps of the expressway into a directed network graph, and at the same time consider the basic lane sections, lane-changing connection sections, on-ramps, and off-ramps; S2. Classify the movement behaviors of vehicles on the expressway into two categories: car-following and lane-changing, and determine the cost functions of the two movement behaviors respectively; The step S2 includes the following steps: S201. Classify the behaviors of vehicles on the expressway into two major categories: car-following and lane-changing: The car-following behavior refers to the driving behavior of a vehicle following the vehicle in front in the lane and maintaining a certain safety distance, and the car-following behavior corresponds to the driver's operations on the accelerator and brake; The lane-changing behavior refers to the behavior of a vehicle changing from one lane to another by finding a safe interval existing in the adjacent lane, and the lane-changing behavior corresponds to the driver's operation on the steering wheel; S202, represents the single-lane passing capacity of the mixed human-machine traffic flow on the lane, and the calculation method is given by the following formula: ​ Wherein: is the penetration rate of autonomous vehicles in the lane; represents the lane capacity when the penetration rate of human-driven vehicles is 100%; , is the saturated headway of human-driven vehicles; represents is the lane capacity of 100%; , is the saturated headway of autonomous vehicles; S203. Calculate the travel cost of lane car-following using the following formula: Wherein: are the numbers of the lateral position and longitudinal position of the lane respectively; is the lane travel cost; represents the travel time of the lane under the free flow scenario; represents the traffic flow of human-driven vehicles on the lane ; represents the traffic flow of autonomous vehicles on the lane ; is the coefficient to be calibrated; S204. Calculate the travel cost of lane-changing using the following formula: t b = K1(v out - v in ) 2 + K2d in + K3 where: v out is the driving speed of the lane before lane change; v in is the driving speed of the target lane; d in is the density of the target lane; K1, K2, K3 are coefficients to be calibrated; S3. Obtain the OD matrix of vehicles on the on-ramps and off-ramps of the expressway; S4. Establish a mixed traffic flow distribution model considering the exclusive lane for autonomous driving; S5. Establish an optimization layout decision model for the exclusive lane for autonomous driving; S6. Solve to obtain the optimized layout scheme of the horizontal and vertical positions of the exclusive lane for autonomous driving.

2. The method for optimizing the horizontal and vertical positions of the dedicated lane for autonomous driving in the human-machine co-driving scenario according to claim 1, wherein, The step S1 includes the following steps: S101. Abstract the basic lane sections, on-ramps, and off-ramps as one-way edges in the network graph, where each lane contains two attributes: lateral position and longitudinal position. These are two attributes. S102. Abstract the lane-changing connection sections between ramps and lanes, and between lanes as two-way edges in the network graph, numbered as ; S103. Abstract the starting point, ending point, on-ramps and off-ramps of the expressway and the intersections of the basic lane sections as nodes in the network graph; S104. Connect the nodes and edges to form a directed network graph, that is, complete the modeling of the expressway facilities.

3. The method for optimizing the horizontal and vertical positions of the dedicated lane for autonomous driving in the human-machine co-driving scenario according to claim 2, wherein The following steps are included in the step S3: S301. Use the video detectors installed at the on-ramps and off-ramps to record the numbers of the on-ramps and off-ramps of the vehicles; S302. Statistically analyze the vehicle on-ramp and off-ramp conditions during the morning and evening peak hours on weekdays and on weekends and rest days respectively, and obtain the OD matrix of the expressway on weekdays and on weekends and rest days. The OD matrix includes the starting point, ending point, on-ramps, and off-ramps of the expressway; S303. Split the OD matrix into that of human-driven vehicles and that of autonomous vehicles according to the penetration rate of autonomous vehicles, and the calculation formula is as follows: 。 4. The method for optimizing the horizontal and vertical positions of the dedicated lane for autonomous driving in the human-machine co-driving scenario according to claim 3, wherein, The step S4 includes the following steps: S401. Establish a mixed traffic flow distribution model considering the exclusive lane for autonomous driving, and the specific mathematical expression is: Wherein: represents the traffic flow of human-driven vehicles on the r-s th path between OD pairs i ; represents the traffic flow of autonomous vehicles on the r-s th path between OD pairs i ; is a decision variable, indicating whether a dedicated lane for autonomous vehicles is set up. It takes 1 when set up and 0 when not set up; a,l M is a penalty term and takes a sufficiently large number.​ 5. The method for optimizing the horizontal and vertical positions of the dedicated lane for autonomous driving in the human-machine co-driving scenario according to claim 4, wherein The step S5 includes the following steps: Establish an optimization layout decision model for the exclusive lane for autonomous driving, and the specific mathematical expression is: Wherein: Obtained from the mixed traffic flow distribution model in the step S401 Is the system travel time Is the non - dedicated road path count variable Is the path count variable after setting dedicated lanes. When the path passes through the lane change numbered b Take 1, otherwise take 0 6. The method for optimizing the horizontal and vertical layout of the exclusive automatic driving lane in the human-machine co-driving scenario according to claim 5, wherein The step S6 includes the following steps: S601. Use a heuristic algorithm to solve the optimization layout decision model for the exclusive lane for autonomous driving, and the mixed traffic flow distribution model is solved using the MATLAB optimization toolbox; S602. Obtain the optimal , and set the road sections corresponding to the parts with being 1 as dedicated lanes for autonomous vehicles, and the rest as mixed traffic lanes.

7. The method for optimizing the horizontal and vertical position layout of the dedicated lane for autonomous driving in the human-machine co-driving scenario according to claim 6, wherein The heuristic algorithm is a genetic algorithm or a simulated annealing algorithm.

Images