Collaborative decision planning method and model for mixed traffic off-ramp scene

Abstract

The invention discloses a collaborative decision planning method for a mixed traffic ramp scene, and the method comprises the steps: building a highway ramp combination scene, formulating a road right distribution scheme based on vehicle information and constraint conditions, obtaining the time when a vehicle passes through a combination region under an ideal condition, and carrying out the calculation of the time when the vehicle passes through the combination region; and establishing an optimal control problem to obtain a control scheme and a prediction track of all vehicles, and updating the driving track of each vehicle in the highway ramp combined scene by using an event-driven track re-planning method in combination with the real-time driving state of the vehicle and the actual road condition. And obtaining a collaborative decision planning scheme of the highway ramp combination scene under the mixed traffic, and controlling the driving of all vehicles in the highway ramp combination scene. According to the method, the vehicle driving track planning is driven based on the track deviation event, and the self-adaptive deviation threshold is combined, so that the decision making efficiency is improved, the computing power consumption is reduced, and the making of an intelligent driving decision in a mixed traffic scene is guided.

Description

Technical Field

[0001] The present disclosure relates to the technical field of intelligent traffic control, and particularly relates to a collaborative decision-making and planning method and its model for a mixed traffic ramp scenario. Background Art

[0002] An intelligent connected vehicle refers to a new generation of vehicle that is equipped with advanced on-vehicle sensors, controllers, actuators and other devices, and through the integration of modern communication and network technologies, has functions such as complex environment perception, intelligent decision-making, and collaborative control, can achieve safe, efficient, comfortable, and energy-saving driving, and ultimately replace manual operation. With the development of communication technology, multi-vehicle collaborative control in intelligent connected scenarios can be realized. Compared with general traffic signal control methods, multi-vehicle collaborative control can achieve better traffic control performance through the coordination and motion planning of vehicle groups, and reduce fuel consumption and environmental pollution caused by traffic.

[0003] In the actual traffic environment, the mixed traffic scenario where human-driven vehicles and intelligent connected vehicles coexist will exist for a long time. Due to the uncertainty of the driving styles of human-driven vehicles and generally taking individual optimality as the starting point of decision-making, it is impossible to directly apply the collaborative control method based on global optimality to control intelligent connected vehicles, otherwise there will be a risk of causing traffic accidents.

[0004] In this case, how to observe human-driven vehicles based on roadside equipment, achieve real-time and efficient control of intelligent connected vehicles, and ensure the driving safety of intelligent connected vehicles and intelligent connected vehicles has become a challenging task. Therefore, there is an urgent need to propose a collaborative decision-making and planning method for a mixed traffic ramp scenario to guide the implementation of intelligent driving in the mixed traffic scenario. Summary of the Invention

[0005] The present invention provides a collaborative decision-making and planning method for a mixed traffic ramp scenario, including the following steps:

[0006] Obtain vehicle information in the highway ramp merging scenario at time t, and predict the driving trajectories of each vehicle based on the vehicle information in the highway ramp merging scenario at time t, where the vehicles include human-driven vehicles and intelligent connected vehicles;

[0007] And use an event-driven trajectory replanning method to determine the driving trajectories of each vehicle in the highway ramp merging scenario at time t + 1, and obtain the collaborative decision-making and planning scheme, where the event-driven trajectory replanning method is associated with the trajectory deviation degree of the target human vehicle, and the trajectory deviation degree is determined based on the vehicle information and driving prediction trajectory of the target human vehicle.

[0008] Preferably, predicting the driving trajectories of each intelligent connected vehicle based on the vehicle information of the highway ramp merging scenario at time t includes:

[0009] Determine a road right allocation function based on the vehicle information and the constraints of the vehicle in the highway ramp merging scenario, solve the road right allocation function to obtain a road right allocation plan, and the road right allocation plan is used to determine the order of each vehicle passing through the merging area;

[0010] Establish an optimal control plan based on the road right allocation plan, and the optimal control plan is used to optimize the energy consumption during vehicle driving;

[0011] And solve the road right allocation plan and the optimal control plan to obtain the driving prediction trajectories of all vehicles.

[0012] Preferably, the constraints in the highway ramp merging scenario include constraint one and constraint two;

[0013] Among them, constraint one is set based on the kinematic constraints of the vehicle itself, including:

[0014] u min ≤u i ≤u max ,v min ≤v i ≤v max

[0015] In the formula, u i is the acceleration of the i-th vehicle; u min is the minimum acceleration; u max is the maximum acceleration; v i is the speed of the i-th vehicle; v min is the minimum speed; v max is the maximum speed;

[0016] Among them, constraint two is set based on the headway when the vehicle passes through the merging area, and the constraint two includes a first sub-constraint and a second sub-constraint;

[0017] The first sub-constraint is set for adjacent vehicles traveling on the same lane, including:

[0018]

[0019] Among them, l and k are both vehicle numbers; is the moment when vehicle l reaches the merging area; is the moment when vehicle k reaches the merging area; Δt 1 is the preset minimum headway when adjacent vehicles on the same lane pass through the merging area;

[0020] The second sub-constraint is set for adjacent vehicles passing through the merging area successively on different lanes, and includes that for adjacent vehicles on different lanes, when the two vehicles pass through the merging area successively, the constraint two is set as:

[0021]

[0022] where Δt 2 is the preset minimum headway time when adjacent vehicles on different lanes pass through the merging area.

[0023] Preferably, determining the right-of-way allocation function includes:

[0024] Based on the constraints in the highway ramp merging scenario, with the global traffic efficiency as the optimization objective, the optimization objective function is obtained:

[0025]

[0026] where J eff is the optimization objective; ω 1 , ω 2 are both weight coefficients; max is the maximum value function; i is the vehicle number; is the time for the i-th vehicle to reach the merging area under a certain specific right-of-way allocation scheme; is the shortest time for vehicle i to reach the merging area under kinematic constraints; n 1 is the total number of human-driven vehicles; n 2 is the total number of intelligent connected vehicles;

[0027] The right-of-way allocation function F is determined as:

[0028]

[0029] where T assign is the set of times when all vehicles may reach the merging area under all right-of-way allocation schemes; B is the set of all binary variables b k,l ; n is the sum of the numbers of all vehicles, that is, n 1 +n 2 ; M is a preset positive number; N 1 is the set of all human-driven vehicles on the main road; N 2 is the set of all intelligent connected vehicles on the merging lane; b k,l is a binary variable used to represent the passing sequence of the vehicles on the main road and the vehicles on the merging lane through the merging area. When the vehicle on the main road passes through the merging area before the vehicle on the merging lane, b k,l =1, and when the vehicle on the main road passes through the merging area after the vehicle on the merging lane, b k,l =0;

[0030] According to the right-of-way allocation function, the right-of-way allocation scheme is obtained as follows:

[0031]

[0032] where V ctrl is the right-of-way order of the controlled intelligent connected vehicle; is the vehicle with the right-of-way in front of the intelligent connected vehicle; is the vehicle with the right-of-way behind the intelligent connected vehicle.

[0033] Preferably, establishing the optimal control scheme based on the right-of-way allocation scheme includes:

[0034] Taking the energy consumption during vehicle driving as the optimization objective, and combining the initial conditions, termination conditions, and kinematic constraints for the vehicle to enter the control area, the optimal control scheme is determined as follows:

[0035]

[0036] In the formula, min is the minimum value function; u(·) is the vehicle acceleration; v(·) is the vehicle speed; p(·) is the vehicle position; t is a specific moment; t 0 is the time when the vehicle enters the control area; t f is the time when the vehicle reaches the merging area after the right-of-way is allocated; v 0 is the initial speed of the vehicle; v f is the final speed of the vehicle.

[0037] Preferably, solving the right-of-way allocation scheme and the optimal control scheme to obtain the driving prediction trajectories of all vehicles includes:

[0038] By solving the global optimal solutions of the right-of-way allocation scheme and the optimal control scheme at time t, the driving prediction trajectories of all vehicles in the highway ramp merging scenario at time t are determined.

[0039] Preferably, using the event-driven trajectory replanning method to determine the driving trajectories of each vehicle in the highway ramp merging scenario at the (t + 1)-th moment, and obtaining the collaborative decision-making and planning scheme includes:

[0040] Judging whether the difference between the actual driving trajectory of the target human-driven vehicle at time t and the driving prediction trajectory is greater than a preset trajectory deviation threshold;

[0041] If the difference between the actual driving trajectory of the target human-driven vehicle and the predicted driving trajectory is greater than a preset trajectory deviation threshold, determine that the collaborative decision-making control method at the (t + 1)-th moment is to allocate road rights to all vehicles in the highway ramp merging scenario, update the driving trajectories of all connected and autonomous vehicles in the highway ramp merging scenario by executing the optimal control scheme, predict the predicted trajectory of the human-driven vehicle, and control the connected and autonomous vehicles to continuously perceive the vehicle information of the human-driven vehicle.

[0042] Preferably, the method further includes:

[0043] If the difference between the actual driving trajectory of the target human-driven vehicle and the predicted driving trajectory is less than or equal to the preset trajectory deviation threshold, determine whether the preset allocation condition is true. The preset allocation condition is: whether a new human-driven vehicle enters the main road at the t-th moment and the connected and autonomous vehicle is allocated the last road right.

[0044] If the preset allocation condition is true, determine that the execution collaborative decision-making control method at the (t + 1)-th moment is to allocate road rights to all vehicles in the highway ramp merging scenario.

[0045] Preferably, the method further includes:

[0046] If the preset allocation condition is false, determine that the execution collaborative decision-making control method at the (t + 1)-th moment is to control the connected and autonomous vehicle to continuously perceive the vehicle information of the human-driven vehicle, update the time step, and continue to monitor the driving trajectories of all vehicles in the highway ramp merging scenario.

[0047] Preferably, the trajectory deviation threshold can be adaptively adjusted according to the highway ramp merging scenario, and the calculation formula is:

[0048]

[0049] In the formula, τ(x) is the trajectory deviation threshold; x is the minimum distance between all connected and autonomous vehicles in the control area and the merging area; τ max is the maximum threshold; L is the length of the control area.

[0050] The present invention also provides an intelligent driver model based on diversified driving behaviors, which is applied to a collaborative decision-making and planning method for a ramp scenario under mixed traffic as described above, and is used to simulate the driving behavior of a driver to predict the driving trajectory of the human-driven vehicle.

[0051] Preferably, the intelligent driver model is:

[0052]

[0053] Among them,

[0054]

[0055] wherein, is the acceleration of the vehicle; a is the maximum acceleration of the vehicle; v is the longitudinal speed of the vehicle; v 0 is the desired vehicle speed; δ is the acceleration index; s′(v,Δv) is the desired distance; Δv is the relative speed between the vehicle and the vehicle in front; s is the relative distance between the vehicle and the vehicle in front; s 0 is the static safety distance; T is the safety headway; b is the desired deceleration;

[0056] The static safety distance, the safety headway, the desired deceleration and the acceleration index are IDM parameters;

[0057] By setting the IDM parameters and the occurrence probability of the behavior type of the intelligent driver model, the driving behavior of the driver is predicted by using the intelligent driver model.

[0058] The beneficial technical effects brought by the present invention are as follows:

[0059] (1) The present invention proposes a collaborative decision-making and planning method for the ramp scene under mixed traffic, solves the problem that it is difficult to accurately formulate the ramp merging decision plan in the mixed traffic scene, reduces the travel time and fuel consumption of vehicles, avoids the collision between autonomous vehicles and manually driven vehicles during the ramp merging process, and improves the traffic efficiency and safety of vehicles at the highway ramp merging area.

[0060] (2) The present invention proposes a collaborative decision-making and planning method for the ramp scene under mixed traffic. By introducing the trajectory deviation degree, the prediction result of the driving trajectory of the vehicle is more in line with the actual situation, improves the prediction accuracy of the vehicle trajectory on the premise of ensuring the merging efficiency and safety of the vehicle, weakens the influence of the prediction deviation on the control scheme and prediction trajectory of the vehicle, effectively ensures the accuracy of the trajectory deviation degree setting of the present disclosure method applied to different traffic scenes, and effectively avoids a large number of useless repeated calculations during the execution of the method. Description of the Drawings

[0061] Figure 1 is the flow chart of the traditional vehicle trajectory replanning method.

[0062] Figure 2 is the schematic diagram of the highway ramp merging scene of the present invention.

[0063] Figure 3 is the curve of the trajectory deviation degree threshold of the present invention changing with the minimum distance between the intelligent connected vehicle and the merging area.

[0064] Figure 4This is a flow chart of the centralized collaborative decision-making control method under mixed traffic of the present invention.

[0065] In the figure: 1. Manually driven vehicle; 2. Intelligent connected vehicle; 3. Perception system; 4. Merging area; 5. Control area. DETAILED DESCRIPTION

[0066] The present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0067] Example 1

[0068] In the prior art, for manually driven vehicles, drivers generally use individual driving efficiency as their driving goal, and their behavior is affected by many factors such as the driver's driving style, reaction time, and field of vision, resulting in a large difference between the vehicle's driving trajectory and the predicted trajectory planned based on the global optimal goal. In order to ensure the safety and efficiency of intelligent connected vehicles during driving, it is necessary to adjust and re-plan the trajectory of intelligent connected vehicles in real time.

[0069] Traditional vehicle trajectory replanning methods such as Figure 1 As shown in the figure, this method executes the planning method at a fixed time interval to update the planned trajectory. However, the traditional vehicle trajectory replanning method has certain limitations, namely how to determine the time interval between two adjacent operations. When the set time interval is too long, the update calculation frequency is too low, and it is easy for the driving trajectory of the manually driven vehicle to deviate too much from the predicted trajectory within the interval, causing potential risks; when the set time interval is too short, the frequency of execution of the calculation method is too high, which will bring frequent computing power requirements and additional resource consumption. In addition, factors such as traffic congestion in different traffic environments and differences in driving styles of human drivers will affect the execution interval of the calculation method, so the use of a time-driven centralized collaborative decision-making control method is not the best choice for vehicle trajectory planning.

[0070] In view of the above problems existing in the prior art, this embodiment proposes a collaborative decision-making planning method for a mixed traffic off-ramp scenario, comprising the following steps:

[0071] Based on the actual situation of highway ramp merging under mixed traffic, a highway ramp merging scenario is constructed.

[0072] The highway ramp merging scenario is as follows Figure 2As shown in the figure, it includes a main lane and an on-ramp lane. Among them, the vehicles driving on the main lane are human-driven vehicles (HDV for short), and the vehicles driving on the on-ramp lane are connected and autonomous vehicles (CAV for short). On one side of the main lane in the highway ramp merging scenario, there is a perception system for obtaining the driving state information of human-driven vehicles and connected and autonomous vehicles. The measurement area of the perception system is the control area, the merging point where the on-ramp lane merges into the main lane is the merging point, and the area where it merges into the main lane is the merging area. For the convenience of description, in the embodiments of the present disclosure, it is assumed that the merging point and the merging area are the same, both being the intersection of the main lane and the on-ramp lane. The perception system is connected to the central controller, and the central controller is used to perform trajectory planning for the connected and autonomous vehicles according to the measurement information of the perception system and send control commands to the connected and autonomous vehicles.

[0073] Obtain the vehicle information in the highway ramp merging scenario at time t (for example, it can be the initial time). The vehicles in the embodiments of the present disclosure include human-driven vehicles and connected and autonomous vehicles. The vehicle information includes the positions and speeds of each human-driven vehicle and each connected and autonomous vehicle. At time t, the speeds and positions of all human-driven vehicles and all connected and autonomous vehicles in the ramp merging scenario can be determined. It should be understood that when t = 0, the initial speeds and initial positions of all human-driven vehicles and all connected and autonomous vehicles are obtained.

[0074] Under ideal conditions, the behaviors of human-driven vehicles and connected and autonomous vehicles are the same. They both drive under the guidance of the central controller following the principle of global optimality. Under this condition, a global optimization problem can be established. Assume that the driving behaviors of human-driven vehicles in the highway ramp merging scenario conform to the criterion of maximizing global benefits. Use the event-driven trajectory replanning method to determine the driving trajectories of each vehicle in the highway ramp merging scenario at time t + 1, and obtain the collaborative decision-making and planning scheme. In the embodiments of the present disclosure, time t + 1 is the next moment after time t. By iteratively calculating the collaborative decision-making and planning method, the real-time driving trajectories of vehicles in the scenario of the present disclosure can be accurately predicted. Among them, the event-driven trajectory replanning method is associated with the trajectory deviation degree of the target human-driven vehicle, and the trajectory deviation degree is determined based on the vehicle information and driving prediction trajectory of the target human-driven vehicle. For each connected and autonomous vehicle, at least one target human-driven vehicle is matched. For each connected and autonomous vehicle, the human-driven vehicles passing through the merging area in front of and / or behind it may affect its merging behavior. Therefore, the human-driven vehicle passing through the merging area in front of and / or behind the i-th connected and autonomous vehicle is the target human-driven vehicle of the i-th connected and autonomous vehicle. The set of target human-driven vehicles of all connected and autonomous vehicles is the target human vehicle.

[0075] In this embodiment, a centralized collaborative decision-making and planning method under mixed traffic driven by events is adopted. When a specific event is detected, the trajectory is updated, and the execution frequency of the calculation method is reduced as much as possible without affecting the performance of the collaborative decision-making and planning method, so as to reduce unnecessary computing power requirements and resource consumption. In each time step, while executing the control of the connected and autonomous vehicle, the real-time trajectory of the human-driven vehicle is sensed. When a preset driving event occurs, the preset centralized collaborative decision-making and control method under mixed traffic is automatically re-executed.

[0076] Among them, the method for predicting the driving trajectories of each connected and autonomous vehicle based on the vehicle information in the highway ramp merging scenario at time t includes operations S1 to S3.

[0077] In operation S1, based on the vehicle information and the constraint conditions of the vehicle in the highway ramp merging scenario, a road right allocation function is determined, and the road right allocation function is solved to obtain a road right allocation scheme, which is used to determine the order of each vehicle passing through the merging area.

[0078] Preferably, the constraint conditions in the highway ramp merging scenario include constraint condition one and constraint condition two.

[0079] Among them, constraint condition one is set based on the kinematic constraints of the vehicle itself, including:

[0080] u min ≤u i ≤u max , v min ≤u i ≤u max

[0081] In the formula, u i is the acceleration of the i-th vehicle; u min is the minimum acceleration; u max is the maximum acceleration; v i is the speed of the i-th vehicle; v min is the minimum speed; v max is the maximum speed;

[0082] Among them, constraint condition two is set based on the time headway when the vehicle passes through the merging area. Constraint condition two includes a first sub-constraint condition and a second sub-constraint condition;

[0083] The first sub-constraint condition is set for adjacent vehicles traveling on the same lane, including:

[0084]

[0085] Among them, k refers to the number of human-driven vehicles, and l is the number of connected and automated vehicles; is the moment when vehicle l reaches the merging area; is the moment when vehicle k reaches the merging area; Δt 1 is the preset minimum headway when adjacent vehicles on the same lane pass through the merging area;

[0086] The second sub-constraint condition is set for adjacent vehicles passing through the merging area successively on different lanes. For adjacent vehicles traveling on different lanes, when the two vehicles pass through the merging area successively, the constraint condition 2 is set as:

[0087]

[0088] Among them, Δt 2 is the preset minimum headway when adjacent vehicles on different lanes pass through the merging area.

[0089] Among them, determining the right-of-way allocation function includes:

[0090] Based on the constraint conditions in the highway ramp merging scenario, with the global traffic efficiency as the optimization goal, the optimization objective function is obtained:

[0091]

[0092] Among them, J eff is the optimization goal; ω 1 , ω 2 are both weight coefficients; max is the maximum value function; i is the vehicle number; is the time when the i-th vehicle reaches the merging area under a certain specific right-of-way allocation scheme; is the shortest time for vehicle i to reach the merging area under kinematic constraints; n 1 is the total number of human-driven vehicles; n 2 is the total number of connected and automated vehicles;

[0093] Determine that the right-of-way allocation function F is:

[0094]

[0095] Among them, T assign is the set of times when all vehicles may reach the merging area under all right-of-way allocation schemes; B is the set of all binary variables b k,l ; n is the sum of the number of all vehicles, that is, n 1 +n 2 ; M is a preset positive number; N 1 is the set of all human-driven vehicles on the main road; N 2 is the set of all connected and automated vehicles on the merging lane; bk,l is a binary variable used to represent the passing order of vehicles traveling on the main road and vehicles traveling on the merging lane through the merging area. When the vehicle traveling on the main road passes through the merging area before the vehicle on the merging lane, b k,l = 1; when the vehicle traveling on the main road passes through the merging area after the vehicle on the merging lane, b k,l = 0.

[0096] In the embodiments of the present disclosure, the right-of-way allocation function is the optimal solution of the above optimization objective function. By solving the optimization objective function to obtain the value of bk,l (i.e., the set of values of b corresponding to each vehicle), the right-of-way allocation scheme can be obtained as follows: k,l where V

[0097]

[0098] is the right-of-way order of the controlled connected vehicle; ctrl is the vehicle with the right-of-way in front of the connected vehicle; is the vehicle with the right-of-way behind the connected vehicle. Typical solution methods may include general solution methods for mixed-integer linear programming problems, and may also be existing dynamic programming algorithms.

[0099] In operation S2, an optimal control scheme is established based on the right-of-way allocation scheme, and the optimal control scheme is used to optimize the energy consumption during vehicle driving. For vehicle V r , according to the time node when it enters the merging area, the optimal control scheme is used to obtain the planned trajectory from the current position to the merging area. For connected vehicles, the trajectory planning result is directly used for control; for manually driven vehicles, the trajectory planning result is used as a comparison reference for their actual trajectories and applied to the subsequent trajectory update process.

[0100] Preferably, establishing the optimal control scheme based on the right-of-way allocation scheme includes:

[0101] Taking the energy consumption during vehicle driving as the optimization objective, and combining the initial conditions, termination conditions and kinematic constraints for the vehicle to enter the control area, the optimal control scheme is determined as:

[0102]

[0103] where min is the minimum value function; u(·) is the vehicle's acceleration; v(·) is the vehicle's speed; p(·) is the vehicle's position; t is a specific moment; t 0 is the time when the vehicle enters the control area; t f is the time when the vehicle reaches the merging area after the right-of-way is allocated; v 0 is the initial speed of the vehicle; v f ​is the final speed of the vehicle.

[0104] In operation S3, solve the road right allocation scheme and the optimal control scheme to obtain the driving prediction trajectories of all vehicles. In an embodiment of the present disclosure, the speed, position, and control quantity (i.e., acceleration) information of the planned vehicle at each moment from the current moment until reaching the merging area can be obtained by solving the optimal control scheme and the driving prediction trajectory.

[0105] Specifically, by solving the global optimal solution of the road right allocation scheme and the optimal control scheme at time t, the driving prediction trajectories of all vehicles in the highway ramp merging scenario at time t can be determined.

[0106] In an embodiment of the present disclosure, the method for determining the driving trajectories of each vehicle in the highway ramp merging scenario at the (t + 1)-th moment by using the event-driven trajectory replanning method to obtain the collaborative decision-making and planning scheme includes operations S4 - S8.

[0107] In operation S4, determine whether the difference between the actual driving trajectory of the target human-driven vehicle at time t and the driving prediction trajectory is greater than a preset trajectory deviation threshold. Among them, the trajectory deviation threshold is the maximum deviation amount Δ_x between the actual trajectory and the predicted trajectory of the human-driven vehicle that needs to be recalculated before each update calculation.

[0108] In some embodiments, the trajectory deviation degree can be associated with the position information. For example, the trajectory deviation degree can be calculated by comparing the difference between the position of the target human-driven vehicle at time t and the predicted position value of the driving prediction trajectory of the vehicle at time t.

[0109] When setting the trajectory deviation threshold, if the value of the trajectory deviation threshold is too small, the execution frequency of the planning method adopted in this embodiment will not be much different from that of the time-driven planning method, and the computing power consumption cannot be effectively reduced. If the value of the trajectory deviation threshold is too large, the intelligent connected vehicle cannot adjust its own trajectory in time according to the driving behavior of the human-driven vehicle, especially when approaching the merging area, which is likely to bring potential safety problems.

[0110] Considering that when the intelligent connected vehicle is far from the merging area, the deviation between the actual trajectory and the predicted trajectory of the human-driven vehicle has little impact on the final merging efficiency of the intelligent connected vehicle. However, when the intelligent connected vehicle is very close to the merging area, a slight behavior deviation of the human-driven vehicle will have a great impact on the safety of vehicle lane-changing. Therefore, in this embodiment, a trajectory deviation threshold that can be adaptively adjusted according to the highway ramp merging scenario is adopted, and the trajectory deviation threshold is:

[0111]

[0112] In the formula, τ(x) is the trajectory deviation threshold; x is the minimum distance between all connected and automated vehicles in the control area and the merging area; τ max is the maximum threshold; L is the length of the control area.

[0113] Figure 3 The following shows the curve of the trajectory deviation threshold of the present disclosure varying with the minimum distance between the connected and automated vehicle and the merging area. From Figure 3 it can be seen that as the distance between the human-driven vehicle and the merging area gradually decreases, the trajectory deviation threshold decays quadratically until it drops to 0.

[0114] In an embodiment of the present disclosure, if the difference between the actual driving trajectory of the target human-driven vehicle and the driving prediction trajectory is greater than the preset trajectory deviation threshold, operation S5 is executed.

[0115] In operation S5, determine the cooperative decision-making control method at the (t + 1)-th moment to allocate road rights to all vehicles in the highway ramp merging scenario, update the driving trajectories of all connected and automated vehicles in the highway ramp merging scenario by executing the optimal control scheme, and predict the prediction trajectory of the human-driven vehicle, and control the connected and automated vehicle to perceive the vehicle information of the human-driven vehicle in real time.

[0116] In some other embodiments, if the difference between the actual driving trajectory of the target human-driven vehicle and the driving prediction trajectory is less than or equal to the preset trajectory deviation threshold, operation S6 is executed.

[0117] In operation S6, determine whether the preset allocation condition is true. The preset allocation condition is: whether a new human-driven vehicle enters the main road at the t-th moment and the connected and automated vehicle is allocated the last road right. Among them, the method for determining the last road right is: for each lane, the last vehicle is allocated the last road right.

[0118] If the preset allocation condition is true, then operation S7 is executed.

[0119] In operation S7, determine the execution cooperative decision-making control method at the (t + 1)-th moment to allocate road rights to all vehicles in the highway ramp merging scenario.

[0120] If the preset allocation condition is false, then operation S8 is executed.

[0121] In operation S8, determine the execution cooperative decision-making control method at the (t + 1)-th moment to control the connected and automated vehicle to perceive the vehicle information of the human-driven vehicle in real time, update the time step, and continue to monitor the driving trajectories of all vehicles in the highway ramp merging scenario.

[0122] Figure 4Exemplarily shown is a centralized collaborative decision-making control method under mixed traffic according to an example of the present disclosure. As Figure 4 shown, at the current time step, if the difference between the actual driving trajectory and the predicted trajectory of the target human-driven vehicle is greater than the preset trajectory deviation threshold τ, it is obtained that the difference between the actual driving trajectory and the predicted trajectory of the target human-driven vehicle is too large. Here, only the human-driven vehicles whose right of way is in front of and behind the target human-driven vehicle are considered. Since when the target human-driven vehicle travels according to the established trajectory planning result, only when there is a large deviation in the trajectories of the vehicles traveling in front of and behind the target human-driven vehicle, there is a certain probability that the target human-driven vehicle cannot enter the merging area according to the original control plan, and the trajectory deviations of the other vehicles will not affect the driving of the current connected and autonomous vehicle, that is, the physical positions of the other vehicles are restricted by the vehicles traveling in front of and behind the target vehicle. At this time, the collaborative decision-making control method is executed to allocate the right of way for all vehicles in the highway ramp merging scenario, update the driving trajectories of all connected and autonomous vehicles in the highway ramp merging scenario by executing the optimal control plan, predict the predicted trajectory of the human-driven vehicle, and control the connected and autonomous vehicle to continuously sense the vehicle information of the human-driven vehicle and update to the next time step.

[0123] Otherwise, it is judged whether a new human-driven vehicle has entered the main road and the connected and autonomous vehicle is allocated the last right of way. If a new human-driven vehicle has entered the main road and the connected and autonomous vehicle is allocated the last right of way, considering that the newly entered human-driven vehicle will not affect the previously entered human-driven vehicles, that is, only whether it will affect the original driving trajectory of the target human-driven vehicle needs to be considered. At this time, the collaborative decision-making control method is executed to allocate the right of way for all vehicles in the highway ramp merging scenario. Otherwise, the connected and autonomous vehicle is controlled to continuously sense the vehicle information of the human-driven vehicle, update to the next time step, and continue to monitor the driving trajectories of all vehicles in the highway ramp merging scenario.

[0124] According to the collaborative decision-making planning scheme for the highway ramp merging scenario under mixed traffic of the present disclosure, the central controller and the sensing system provided in the highway ramp merging scenario are used to control the driving of all vehicles in the highway ramp merging scenario.

[0125] The embodiment of the present disclosure also provides an intelligent driver model with diversified driving behaviors, which is applied to the collaborative decision-making planning method for the ramp scenario under mixed traffic described in the embodiment of the present disclosure. It can be preset in the central controller to simulate the driving behavior of the driver to predict the driving trajectory of the human-driven vehicle.

[0126] In real traffic driving scenarios, there may be differences in the driving behaviors of drivers of manually driven vehicles. The driving behaviors of drivers are classified into general type, aggressive type, and conservative type. In order to improve the authenticity and accuracy of simulation test results, during the process of simulation testing, it is necessary to predict the diverse driving behaviors or intentions of drivers, so as to ensure the robustness of the method disclosed in this application in complex and changeable traffic environments.

[0127] Considering influencing factors such as the distance to the vehicle ahead and relative speed during the driving process of a manually driven vehicle, and combining with human driving habits, the intelligent driver model is as follows:

[0128]

[0129] Among them,

[0130]

[0131] In the formula, is the acceleration of the vehicle; a is the maximum acceleration of the vehicle; v is the longitudinal speed of the vehicle; v 0 is the desired vehicle speed; δ is the acceleration index; s′(v,Δv) is the desired distance; Δv is the relative speed between the vehicle and the vehicle ahead; s is the relative distance between the vehicle and the vehicle ahead; s 0 is the static safety distance; T is the safe headway time; b is the desired deceleration;

[0132] The static safety distance, safe headway time, desired deceleration, and acceleration index are IDM parameters; the desired distance is the desired following distance (safety distance).

[0133] By setting the IDM parameters and occurrence probabilities of the intelligent driver model, the driving behaviors of drivers are predicted using the intelligent driver model.

[0134] Embodiment 2

[0135] In order to verify the effectiveness of a collaborative decision-making and planning method for a mixed traffic ramp scenario in Embodiment 1, especially to verify the centralized collaborative decision-making control method based on event-driven in mixed traffic in Embodiment 1, in this embodiment, through simulation, the centralized collaborative decision-making control method based on event-driven in mixed traffic in Embodiment 1 is compared with the traditional vehicle trajectory replanning method.

[0136] To verify the effectiveness of the centralized collaborative decision-making control method for event-driven mixed traffic in Embodiment 1, a simulation model of the highway ramp merging scenario is constructed using simulation software according to the highway ramp merging scenario, the vehicle merging process at the highway ramp is simulated, and the performance of this method is compared with that of the traditional time-driven vehicle trajectory replanning method in the simulation environment. The performance metrics for comparison include the number of executions, the number of collisions, the average delay time, the average energy consumption, and the average fuel consumption, and only the collisions, delay times, energy consumption, and fuel consumption of the connected and automated vehicles in the highway ramp merging scenario are considered.

[0137] Among them, the number of executions is the number of times the calculation method is executed at all times when there are connected and automated vehicles in the control area during the entire simulation process; the delay time is the difference Δ between the actual time when each vehicle arrives at the merging area and the minimum possible time for it to arrive at the merging area in the highway ramp merging scenario. t Δ t = t f - t min where t f is the time when the vehicle arrives at the merging area after the road rights are allocated, and t min is the minimum possible time for the vehicle to arrive at the merging area; the average delay time is the average of the delay times of all vehicles in the highway ramp merging scenario; the average energy consumption E is the integral of the square of the acceleration over time during the vehicle's driving process. The average fuel consumption F is the sum of the fuel consumption during the vehicle's uniform driving and the additional fuel consumption during acceleration. Among them, f cruise (t) is the fuel consumption function during the vehicle's uniform driving, f cruise (t) = b 0 + b 1 v(t) + b 2 v 2 (t) + b 3 v 3 (t), b 0 , b 1 , b 2 , b 3 are all fitting coefficients of the fuel consumption function during the vehicle's uniform driving, f accel (t) is the fuel consumption function during the vehicle's accelerating driving, f accel (t) = u(t)(c 0 + c 1 v(t) + c 2 v 2 (t)), c 0 , c 1 , c 2They are all fitting coefficients of the fuel consumption function when the vehicle is accelerating. u(t) is the forward acceleration during the vehicle's driving. When the vehicle is decelerating, u(t) = 0.

[0138] During the simulation process, the Poisson process is used to simulate the mixed traffic flow. Different vehicle arrival rates are set. The performance of the centralized cooperative decision-making control method based on event-driven under mixed traffic proposed in this disclosure is compared with the traditional time-driven vehicle trajectory replanning method. The simulation results are shown in Table 1.

[0139] Table 1 Performance simulation results

[0140]

[0141]

[0142] When the time interval Δt = 0.1, it is equivalent to making the optimal decision based on the current situation at each moment. That is, it can be considered that the performance obtained by the simulation is nearly optimal. Under the same vehicle arrival rate, using the method of this disclosure greatly reduces the number of calculations compared with the traditional time-driven vehicle trajectory replanning method and ensures a zero accident rate for the vehicle. When the time interval Δt ≥ 0.5, using the method of this disclosure compared with the traditional time-driven vehicle trajectory replanning method, under the condition of similar execution times, the method of this disclosure can effectively avoid vehicle accidents.

[0143] When the vehicle arrival rate is low, the average delay time of the method of this disclosure based on deviation feedback is similar to that of the traditional time-driven vehicle trajectory replanning method, that is, their traffic efficiencies are similar, but the corresponding energy consumption is relatively large. This is because when there are fewer vehicles, since the frequency of re-executing the decision-making planning method is low, it may be necessary to make relatively large decision adjustments at some moments, and the corresponding acceleration and deceleration amplitudes of the vehicle are large, so the energy consumption increases.

[0144] When the vehicle arrival rate is high, the average delay time obtained by the method of this disclosure based on deviation feedback is lower than that of the traditional time-driven vehicle trajectory replanning method. This is because in the first half of the driving, the trajectory deviation threshold is relatively large, but the traffic flow density is also large. If frequent planning is carried out, the intelligent connected vehicle is very likely to give way to the manually driven vehicle on the main road in advance, thus reducing its own speed. At this time, reducing the frequency of decision-making planning is beneficial to reducing the delay time.

[0145] To verify the effectiveness of the adaptive trajectory deviation threshold in the method of this disclosure, the deviation feedback method based on the adaptive threshold is compared with the deviation feedback method based on the fixed threshold. The thresholds are set to 0.1 m, 1 m, and 10 m respectively, and the other conditions remain unchanged. The simulation model of the highway ramp merging scenario is used for simulation respectively. The simulation results are shown in Table 2.

[0146] Table 2 Threshold Simulation Results

[0147]

[0148]

[0149] It can be analyzed from Table 2 that, compared with using a fixed threshold, the method of the present disclosure controls the execution times of the method reasonably by introducing an adaptive trajectory deviation threshold while ensuring the driving safety of the vehicle, reducing the execution times by 20% - 30% compared with the case of using a small threshold. Moreover, at a relatively high vehicle arrival rate, the adaptive threshold in the method of the present disclosure can flexibly arrange the time nodes of decision-making and planning, effectively avoiding unnecessary decisions and improving the driving performance of the intelligent connected vehicle.

[0150] In summary, through simulation in this embodiment, the advantages of the method of the present disclosure compared with the traditional time-driven vehicle trajectory replanning method are verified. The method of the present disclosure converts time-driven to event-driven, and by introducing an adaptive threshold, it greatly reduces the computing power consumption in the entire method calculation process while ensuring traffic safety and traffic efficiency. At the same time, through the ablation experiment of the adaptive threshold, it is also verified that introducing the adaptive threshold can balance the computing power consumption and the overall traffic performance, avoiding excessive execution times while obtaining better algorithm performance. In addition, it can be found from the analysis of the simulation results that the method of the present disclosure can achieve more excellent results under high traffic flow conditions, thus verifying the effectiveness and feasibility of the method of the present disclosure.

[0151] Embodiment 3

[0152] In order to simulate the driving behavior of drivers in an actual traffic scenario, in this embodiment, the gap between the simulation test and the real situation is reduced by randomly initializing the intelligent driver model. Considering that completely random parameter settings ignore the correlation between parameters, there may be situations where the behavior of the manually driven vehicle violates common sense during the simulation test. Therefore, this embodiment proposes a scheme for artificially constraining the IDM parameters of the intelligent driver model, where the IDM parameters include the static safety distance, the safety headway time, the desired deceleration, and the acceleration index.

[0153] The driving behaviors of human drivers are divided into general type, aggressive type, and conservative type, and different IDM parameters and occurrence probabilities are set respectively, as shown in Table 3. Among them, the occurrence probability refers to the probability of a certain type of driving behavior occurring.

[0154] Table 3 Driving Behavior Parameter Table

[0155]

[0156] During the simulation process based on the above intelligent driver model, the cooperative decision-making and planning method for the mixed traffic ramp scenario proposed in Embodiment 1 is used to predict the driving trajectory of the manually driven vehicle and determine the driving behavior of the manually driven vehicle:

[0157]

[0158] In the formula, B i = 0 represents a general driving behavior, B i = 1 represents an aggressive driving behavior, B i = 2 represents a conservative driving behavior; is the ideal trajectory obtained by solving the optimal control problem, is the actual driving trajectory of the vehicle; ∈ is a relatively small number, representing the tolerance for the deviation of the vehicle trajectory.

[0159] Apply the intelligent driver model to the highway ramp merging scenario simulation model in Embodiment 2, reconstruct the highway ramp merging scenario simulation model, and use the highway ramp merging scenario simulation model considering diversified driving behaviors to conduct simulation tests on the method of the present disclosure. The simulation results are shown in Table 4.

[0160] Table 4 Simulation results considering diversified driving behaviors

[0161]

[0162] Analyzing Table 4, it can be obtained that, under the condition of fully considering diversified driving behaviors, the cooperative decision-making and planning method for the mixed traffic ramp scenario adopted by the method of the present disclosure can still achieve approximately globally optimal performance by introducing an adaptive threshold, that is, it is similar to the traditional time-driven vehicle trajectory replanning method with a time interval of 0.1 s, and the calculation efficiency of the method of the present disclosure is greatly improved, and it has good robustness.

[0163] Certainly, the above description is not a limitation to the present disclosure, and the present disclosure is not limited to the above examples either. Changes, modifications, additions or substitutions made by those skilled in the art within the essence scope of the present disclosure should also belong to the protection scope of the present disclosure.

Claims

1. A collaborative decision-making planning method for mixed traffic off-ramp scenarios, characterized in that: The following steps are involved: Acquire vehicle information in a highway ramp merging scenario at time t, and predict the driving trajectory of each vehicle based on the vehicle information in the highway ramp merging scenario at time t, wherein the vehicles include manually driven vehicles and intelligent networked vehicles; And using an event-driven trajectory replanning method to determine the driving trajectory of each vehicle in the highway ramp merging scenario at time t+1, and obtain the collaborative decision-making planning scheme, wherein the event-driven trajectory replanning method is associated with the trajectory deviation of the target artificial vehicle, and the trajectory deviation is determined based on the vehicle information and driving prediction trajectory of the target artificial vehicle.

2. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 1 is characterized in that: The predicting of the driving trajectory of each intelligent connected vehicle based on the vehicle information of the highway ramp merging scenario at time t includes: Determine a road right allocation function based on the vehicle information and the constraint conditions of the vehicle in the highway ramp merging scenario, solve the road right allocation function to obtain a road right allocation scheme, and the road right allocation scheme is used to determine the order in which each vehicle passes through the merging area; Establishing an optimal control scheme based on the road right allocation scheme, wherein the optimal control scheme is used to optimize energy consumption during vehicle driving; And solve the road right allocation scheme and the optimal control scheme to obtain the predicted driving trajectories of all vehicles.

3. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 2 is characterized in that: The constraint conditions in the highway ramp merging scenario include constraint condition 1 and constraint condition 2; Among them, constraint condition 1 is set based on the vehicle's own kinematic constraints, including: u min ≤u i ≤u max ,v min ≤v i ≤v max In the formula, u i is the acceleration of the i-th vehicle; u min is the minimum acceleration; u max is the maximum acceleration; v i is the speed of the i-th vehicle; v min is the minimum speed; v max is the maximum speed; Wherein, a second constraint condition is set based on the headway time when the vehicle passes through the merging area, and the second constraint condition includes a first sub-constraint condition and a second sub-constraint condition; The first sub-constraint is set relative to adjacent vehicles traveling on the same lane, including: Among them, l and k are vehicle numbers; is the time when vehicle l reaches the merging zone; is the time when vehicle k reaches the merging area; Δt1 is the preset minimum headway time between adjacent vehicles on the same lane when they pass through the merging area; The second sub-constraint is set for adjacent vehicles traveling on different lanes that pass through the merging area successively. For adjacent vehicles traveling on different lanes, when two vehicles pass through the merging area successively, the second constraint is set as follows: Among them, Δt2 is the preset minimum headway time between adjacent vehicles on different lanes when they pass through the merging area.

4. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to any one of claims 2 or 3, characterized in that: Determine the road right allocation function including: Based on the constraints of the highway ramp merging scenario, the global traffic efficiency is taken as the optimization goal, and the optimization objective function is obtained: Among them, J eff is the optimization target; ω1 and ω2 are weight coefficients; max is the maximum value function; i is the vehicle serial number; is the time when the i-th vehicle arrives at the merge zone under a specific road right allocation scheme; is the shortest time for vehicle i to reach the merging area under kinematic constraints; n1 is the total number of manually driven vehicles; n2 is the total number of intelligent connected vehicles; Determine the road right allocation function F as: Among them, T assign is the set of possible arrival times of all vehicles at the merge zone under all road right allocation schemes; B is all zero-one variables b k，l The set of; n is the sum of all vehicles, that is, n1+n2; M is a preset positive number; N1 is the set of all manually driven vehicles on the main road; N2 is the set of all intelligent connected vehicles on the merging lane; b k，l is a zero-one variable, which is used to indicate the order in which vehicles on the main road and vehicles on the merging lane pass through the merging area. When the vehicle on the main road passes through the merging area before the vehicle on the merging lane, b k，l = 1, when the vehicle on the main lane passes the merging area later than the vehicle on the merging lane, b k，l =0; According to the road right allocation function, the road right allocation scheme is obtained as follows: Among them, V ctrl The right-of-way order of the controlled intelligent connected vehicles; The vehicle with the right of way is in front of the intelligent connected vehicle; The vehicle with the right of way behind the intelligent connected vehicle.

5. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 4 is characterized in that: The establishing of the optimal control scheme based on the road right allocation scheme comprises: Taking the energy consumption of the vehicle during driving as the optimization target, combined with the initial conditions, termination conditions and kinematic constraints of the vehicle entering the control area, the optimal control scheme is determined as follows: Where min is the minimum function; u(·) is the acceleration of the vehicle; v(·) is the speed of the vehicle; p(·) is the position of the vehicle; t is a specific time; t 0 is the time when the vehicle enters the control area; t f is the time it takes for the vehicle to reach the merge zone after the right of way is allocated; v0 is the initial speed of the vehicle; v f is the final velocity of the vehicle.

6. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 5 is characterized in that: Solving the road right allocation scheme and the optimal control scheme to obtain the predicted driving trajectories of all vehicles includes: By solving the global optimal solution of the road right allocation scheme and the optimal control scheme at time t, the predicted driving trajectories of all vehicles in the highway ramp merging scenario at time t are determined.

7. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 1 is characterized in that: The event-driven trajectory replanning method is used to determine the driving trajectory of each vehicle in the highway ramp merging scenario at time t+1, and the collaborative decision-making planning scheme is obtained, including: Determine whether the difference between the actual driving trajectory of the target manually driven vehicle at time t and the predicted driving trajectory is greater than a preset trajectory deviation threshold; If the difference between the actual driving trajectory of the target manually driven vehicle and the predicted driving trajectory is greater than a preset trajectory deviation threshold, the collaborative decision-making control method at time t+1 is determined to allocate road rights to all vehicles in the highway ramp merging scenario, and the driving trajectories of all intelligent connected vehicles in the highway ramp merging scenario are updated by executing the optimal control scheme, and the predicted trajectory of the manually driven vehicle is predicted, so as to control the intelligent connected vehicle to perceive the vehicle information of the manually driven vehicle in real time.

8. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 7 is characterized in that: The method further includes: If the difference between the actual driving trajectory of the target manually driven vehicle and the predicted driving trajectory is less than or equal to the preset trajectory deviation threshold, it is determined whether a preset allocation condition is true, and the preset allocation condition is: whether a new manually driven vehicle enters the main road at time t and the intelligent networked vehicle is allocated to the last right of way; If the preset allocation condition is true, it is determined that the execution collaborative decision control method at time t+1 allocates road rights to all vehicles in the highway ramp merging scenario.

9. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 7, characterized in that: The method further includes: If the preset allocation condition is false, the collaborative decision-making control method for executing at time t+1 is to control the intelligent connected vehicle to perceive the vehicle information of the manually driven vehicle in real time, update the time step, and continue to monitor the driving trajectories of all vehicles in the highway ramp merging scenario.

10. The collaborative decision-making planning method for mixed traffic off-ramp scenarios according to claim 7, characterized in that: The trajectory deviation threshold can be adaptively adjusted according to the highway ramp merging scenario, and the calculation formula is: Where τ(x) is the trajectory deviation threshold; x is the minimum distance between all intelligent connected vehicles in the control area and the merging area; τ max is the maximum threshold; L is the length of the control area.

11. Intelligent driver model based on diversified driving behaviors, characterized by: Applied to the collaborative decision-making planning method for mixed traffic off-ramp scenarios as described in any one of claims 1 to 10, to simulate the driver's driving behavior to predict the driving trajectory of the manually driven vehicle.

12. The intelligent driver model based on diversified driving behaviors according to claim 11, characterized in that: The intelligent driver model is: in, In the formula, is the acceleration of the vehicle; a is the maximum acceleration of the vehicle; v is the longitudinal speed of the vehicle; v0 is the expected speed; δ is the acceleration index; s′(v,Δv) is the expected distance; Δv is the relative speed between the vehicle and the vehicle in front; s is the relative distance between the vehicle and the vehicle in front; s0 is the static safety distance; T is the safe headway; v is the expected deceleration; The static safety distance, safe headway, expected deceleration and acceleration index are IDM parameters; By setting the IDM parameters of the intelligent driver model and the probability of occurrence of the behavior type, the intelligent driver model is used to predict the driver's driving behavior.

Images