CN110867097A - An autonomous decision-making method for conflict risk avoidance in expressway merging areas - Google Patents

An autonomous decision-making method for conflict risk avoidance in expressway merging areas Download PDF

Info

Publication number
CN110867097A
CN110867097A CN201911115216.1A CN201911115216A CN110867097A CN 110867097 A CN110867097 A CN 110867097A CN 201911115216 A CN201911115216 A CN 201911115216A CN 110867097 A CN110867097 A CN 110867097A
Authority
CN
China
Prior art keywords
vehicle
vehicles
automatic driving
cost
adjustment
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911115216.1A
Other languages
Chinese (zh)
Inventor
王立超
杨敏
汪林
李烨
张健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southeast University
Original Assignee
Southeast University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southeast University filed Critical Southeast University
Priority to CN201911115216.1A priority Critical patent/CN110867097A/en
Publication of CN110867097A publication Critical patent/CN110867097A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computational Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Evolutionary Biology (AREA)
  • Operations Research (AREA)
  • Probability & Statistics with Applications (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Algebra (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Traffic Control Systems (AREA)

Abstract

The invention discloses an autonomous decision-making method for collision avoidance in a highway confluence area, which comprises the following specific steps: initially arranging the conflict risk avoiding sequence of the automatic driving vehicles by using the time demand intensity of the vehicles and the priority levels of the types of the vehicles; when the time demand intensity and the vehicle type priority level are the same, a vehicle driving intention interaction model is established based on a cooperative game theory, and the collision risk avoiding sequence of the automatically driven vehicles is finally arranged; establishing a vehicle driving intention quantification model based on probability subtraction, and realizing intention quantification on the automatic driving vehicle needing driving behavior adjustment; and realizing the automatic decision of automatically driving the vehicle for collision and risk avoidance on the basis of the collision and risk avoidance sequencing and the vehicle driving intention quantification result. The invention has the beneficial effects that: the method actually provides decision-making suggestions for the collision avoidance process of the automatically driven vehicles and provides technical service support for further development of the networked automatically driven vehicles.

Description

高速公路合流区冲突避险自主决策方法An autonomous decision-making method for conflict risk avoidance in expressway merging areas

技术领域technical field

本发明涉及网联自动驾驶车辆技术领域，特别是涉及一种高速公路合流区冲突避险自主决策方法。The invention relates to the technical field of network-connected automatic driving vehicles, in particular to an autonomous decision-making method for conflict avoidance in a highway confluence area.

背景技术Background technique

智能交通系统当前主要发展方向为自动驾驶智能车路协同技术，其发展重心已然由单体智能阶段过渡到群体智能及环境智能交互的进程中，先进的无线通信及互联网技术为自动驾驶车辆间的互联互通提供了保障，也大大提高了自动驾驶车辆间的信息共享程度。在最新的自动驾驶车辆自动避险技术中未能充分利用上述技术的优势和支持实现自动驾驶车辆在避险过程中的相互协商决策，并且已有的自主避险技术中未能考虑到车辆类型、时间需求强度及不同车辆行驶偏好等因素对自动驾驶车辆在自主避险过程中所产生的影响，考虑片面，实用性差，无法满足自动驾驶车辆对于相互合作自主决策避险的要求。The current main development direction of the intelligent transportation system is autonomous driving, intelligent vehicle-road coordination technology, and its development focus has shifted from the single intelligence stage to the process of swarm intelligence and environmental intelligence interaction. Interconnection provides a guarantee and greatly improves the degree of information sharing between autonomous vehicles. In the latest autonomous vehicle avoidance technology, the advantages and support of the above technologies are not fully utilized to realize the mutual consultation decision-making of the autonomous vehicle in the process of risk avoidance, and the vehicle type is not considered in the existing autonomous vehicle avoidance technology. The impact of factors such as time demand intensity and different vehicle driving preferences on autonomous vehicles in the process of autonomous risk avoidance is one-sided and has poor practicability, which cannot meet the requirements of autonomous vehicles for mutual cooperation and autonomous decision-making for risk avoidance.

发明内容SUMMARY OF THE INVENTION

本发明的目的是提供一种高速公路合流区冲突避险自主决策方法，充分考虑自动驾驶车辆在实际运行过程的时间需求强度、车辆类型特性差异，实现以时间需求强度优先等级、车辆类型优先等级为基础的车辆冲突自主避险决策，结合现实条件，实用性好，能够满足自动驾驶车辆面对冲突自主决策避险的要求。The purpose of the present invention is to provide an autonomous decision-making method for conflict and risk avoidance in a highway junction area, which fully considers the time demand intensity and vehicle type characteristics differences of autonomous driving vehicles in the actual operation process, and realizes the time demand intensity priority level and vehicle type priority level. Based on vehicle conflict autonomous risk avoidance decision-making, combined with realistic conditions, it has good practicability and can meet the requirements of autonomous vehicle decision-making and risk avoidance in the face of conflict.

为了实现以上目的，本发明提供了如下方案：In order to achieve the above object, the present invention provides the following scheme:

步骤1：基于车辆时间需求强度对两车辆避险解脱优先等级进行比较，若两自动驾驶车辆时间需求强度相同则转入步骤2，若两自动驾驶车辆时间需求强度不同则确定两车辆避险解脱先后顺序，然后转入步骤4。Step 1: Compare the priority levels of the two vehicles based on the time demand intensity of the two vehicles. If the time demand intensity of the two automatic driving vehicles is the same, go to Step 2. If the time demand intensity of the two automatic driving vehicles is different, determine the two vehicles. in order, then go to step 4.

步骤2：基于车辆类型对两车辆避险解脱优先等级进行比较，若两自动驾驶车辆的车辆类型一致则转入步骤3，若车辆类型不同则按照车辆类型优先等级确定两车辆避险解脱的先后顺序，然后转入步骤4。Step 2: Compare the priority levels of the two vehicles based on the vehicle type. If the vehicle types of the two autonomous vehicles are the same, go to Step 3. If the vehicle types are different, determine the priority of the two vehicles according to the vehicle type priority. sequence, and then go to step 4.

步骤3：两自动驾驶车辆进行行驶意图交互，根据车辆行驶意图交互结果确定两自动驾驶车辆避险解脱方案，然后转入步骤5。Step 3: The two autonomous driving vehicles interact with each other on their driving intentions, and determine an escape plan for the two autonomous driving vehicles according to the interaction results of the driving intentions of the two autonomous driving vehicles, and then go to Step 5.

步骤4：根据两自动驾驶车辆的避险解脱先后顺序确定两车辆是否需要进行调整，确定各车辆是否需要调整后转入步骤5。Step 4: Determine whether the two vehicles need to be adjusted according to the sequence of escape and avoidance of the two autonomous driving vehicles, and then go to step 5 after determining whether each vehicle needs to be adjusted.

步骤5：需进行行为调整的自动驾驶车辆确定本车辆的调节方式，转入步骤 6。Step 5: The self-driving vehicle that needs to perform behavior adjustment determines the adjustment method of its own vehicle, and goes to step 6.

步骤6：自动驾驶车辆行驶行为调整。Step 6: Adjust the driving behavior of the autonomous vehicle.

所述步骤1中基于车辆时间需求强度对两车辆避险解脱优先等级进行比较，在此将第i辆自动驾驶车辆的时间需求强度表示τ，并且有

Figure RE-GDA0002364112880000021
(i＝A、B，A代表自动驾驶车辆A，B代表自动驾驶车辆B，l代表自动驾驶车辆i距离潜在冲突点的长度，v代表自动驾驶车辆i的当前形势速度)。按照τ对存在潜在冲突的两自动驾驶车辆进行时间需求强度优先等级比较。In the step 1, the priority levels of the two vehicles are compared based on the time demand intensity of the vehicle, where the time demand intensity of the i-th autonomous vehicle is expressed as τ, and there are
Figure RE-GDA0002364112880000021
(i=A, B, A represents the autonomous vehicle A, B represents the autonomous vehicle B, l represents the length of the autonomous vehicle i from the potential conflict point, and v represents the current situational speed of the autonomous vehicle i). According to τ, the priority level of time demand intensity is compared between two autonomous vehicles with potential conflict.

所述步骤2中基于车辆类型对两车辆避险解脱优先等级进行比较是在步骤1 中两车辆时间需求强度相同的基础上进行的。在此车辆类型主要从其运输功能和作业任务进行划分，对潜在冲突的两自动驾驶车辆进行车辆类型优先等级确定。The comparison of the priority levels of the two vehicles based on the vehicle type in the step 2 is performed on the basis that the time demand intensity of the two vehicles is the same in the step 1. In this case, the vehicle types are mainly divided from their transportation functions and operational tasks, and the vehicle type priority is determined for the two potentially conflicting autonomous vehicles.

所述步骤3中两自动驾驶车辆进行行驶意图交互是在步骤1中时间需求强度相同且步骤2中车辆类型优先等级相同的基础上进行的，两自动驾驶车辆意图交互的量化过程主要以合作博弈为理论基础。合作博弈的基本目标是实现团队利益的最大化，或者一方利益最大化且他方利益不会受到威胁甚至损失。根据自动驾驶车辆在解决潜在冲突决策过程中的“利益博弈”特点，将合作博弈理论引入到自动驾驶车辆行驶意图交互过程中具有一定的合理性和较强的实践性。The interaction between the two autonomous vehicles in the step 3 is based on the same time demand intensity in step 1 and the same vehicle type priority in step 2. The quantification process of the intention interaction between the two autonomous vehicles is mainly based on cooperative games. for the theoretical basis. The basic goal of cooperative games is to maximize the interests of the team, or to maximize the interests of one party and the interests of the other party will not be threatened or even lost. According to the "game of interest" characteristics of autonomous vehicles in the decision-making process of resolving potential conflicts, it is reasonable and practical to introduce cooperative game theory into the interaction process of autonomous vehicles' driving intentions.

自动驾驶车辆行驶意图交互过程主要分为以下步骤。The interaction process of autonomous vehicle driving intention is mainly divided into the following steps.

步骤3-1：输入各联盟支付成本。Step 3-1: Enter the payment cost for each alliance.

步骤3-2：利用随机函数量化自动驾驶车辆的调整意图，将自动驾驶车辆调整意图和各联盟支付成本输入到虚拟成本总函数计算中。Step 3-2: Use the random function to quantify the adjustment intention of the self-driving vehicle, and input the adjustment intention of the self-driving vehicle and the payment cost of each alliance into the calculation of the virtual cost total function.

步骤3-3：计算各联盟虚拟成本总量，比较各联盟虚拟成本总量的大小。Step 3-3: Calculate the total virtual cost of each alliance, and compare the size of the total virtual cost of each alliance.

步骤3-4：输出联盟虚拟成本总量最小值并确定虚拟成本总量最小值对应的联盟调整方案。Step 3-4: Output the minimum value of the total virtual cost of the alliance and determine the alliance adjustment scheme corresponding to the minimum value of the total virtual cost.

步骤3-5：判断虚拟成本总量最小值对应调整方案是否需进行两车辆同时调整，如果不需要两车辆同时调整则转入步骤3-6，如果需要两车辆同时调整转入步骤3-7。Step 3-5: Determine whether the adjustment plan corresponding to the minimum value of the total virtual cost requires simultaneous adjustment of two vehicles. If it is not necessary to adjust both vehicles at the same time, go to step 3-6. If it is necessary to adjust both vehicles at the same time, go to step 3-7 .

步骤3-6：不需要调整的车辆确定冲突解脱顺序排列靠前，需要调整的车辆确定为冲突解脱顺序排列靠后，冲突解脱顺序确定后转入步骤3-9。Step 3-6: The vehicles that do not need to be adjusted are determined to be ranked in the front of the conflict resolution order, and the vehicles that need to be adjusted are determined to be ranked lower in the conflict resolution sequence. After the conflict resolution sequence is determined, go to Step 3-9.

步骤3-7：计算虚拟成本总量最小值对应联盟调整方案中各联盟成员贡献值。Step 3-7: Calculate the contribution value of each alliance member in the alliance adjustment plan corresponding to the minimum total virtual cost.

步骤3-8：比较调整方案中各联盟成员贡献值的大小，确定较小贡献值所对应的自动驾驶车辆冲突解脱顺序排列靠前，确定较大贡献值所对应的自动驾驶车辆冲突解脱顺序排列靠后，冲突解脱顺序确定后转入步骤3-9。Step 3-8: Compare the contribution value of each alliance member in the adjustment plan, determine the order of the self-driving vehicle conflict resolution corresponding to the smaller contribution value, and determine the order of the self-driving vehicle conflict resolution corresponding to the larger contribution value. Later, after the conflict resolution sequence is determined, go to steps 3-9.

步骤3-9：转入到两车辆冲突解脱逻辑步骤4。Step 3-9: Go to step 4 of the two-vehicle conflict resolution logic.

其中所述步骤3-1中各联盟成员的支付成本xi(i＝A、B，A代表自动驾驶车辆A，B代表自动驾驶车辆B)根据两自动驾驶车辆之间的位置关系及行驶速度，考虑到自动驾驶车辆距离潜在冲突点的长度li越远则自动驾驶车辆越安全，则有xi随着li的增大而降低；行驶速度vi越大则自动驾驶车辆越将会在短时间内靠近冲突点，则有xi随着vi的增大而增大，结合两参数的作用权重k1,k2，确定联盟成员支付成本的来源方程为

Figure RE-GDA0002364112880000041
Wherein the payment cost x i of each alliance member in the step 3-1 (i=A, B, A represents the automatic driving vehicle A, B represents the automatic driving vehicle B) according to the positional relationship between the two automatic driving vehicles and the driving speed , considering that the farther the autonomous vehicle is from the length li of the potential conflict point, the safer the autonomous vehicle is, then x i decreases with the increase of li ; the greater the driving speed vi , the more likely the autonomous vehicle will be . When approaching the conflict point in a short time, x i increases with the increase of vi. Combined with the weights k 1 and k 2 of the two parameters, the source equation for determining the cost of alliance members is as follows :
Figure RE-GDA0002364112880000041

所述步骤3-2中自动驾驶车辆行驶意图量化，是指利用随机函数实现自动驾驶车辆行驶意图的量化过程，首先利用参数a,b代表两自动驾驶车辆调整意图的大小，从而确定a,b为自动驾驶车辆A和自动驾驶车辆B的支付系数，若自动驾驶车辆希望进行行为调整则支付系数取值为1，若自动驾驶车辆不希望进行行为调整则支付系数的取值为大于1。The quantification of the driving intention of the autonomous driving vehicle in the step 3-2 refers to the quantification process of using a random function to realize the driving intention of the autonomous driving vehicle. First, the parameters a and b are used to represent the size of the two autonomous driving vehicles to adjust the intention, so as to determine a, b is the payment coefficient for autonomous vehicle A and autonomous vehicle B. If the autonomous vehicle wishes to perform behavior adjustment, the payment coefficient is 1; if the autonomous vehicle does not wish to perform behavior adjustment, the payment coefficient is greater than 1.

所述步骤3-3中计算个联盟虚拟成本总量的过程为，首先确定两自动驾驶车辆可选择调整行为的无需组合为4种，将j确定为各联盟对应序号，在此拟定j＝1 时，A、B两车辆均选择调整；j＝2时，车辆A选择调整，车辆B不进行调整；j＝3 时，车辆B选择调整，车辆A不进行调整；j＝4时，A、B两车辆均不选择调整，此时两车辆将会发生碰撞风险，于是虚拟支付成本将会无穷大，在实际运行过程中此方案(j＝4)将被直接排除。The process of calculating the total virtual cost of each alliance in the step 3-3 is as follows: first, it is determined that the optional adjustment behavior of the two autonomous vehicles does not need to be combined into 4 types, and j is determined as the corresponding serial number of each alliance, and j=1 is set here. When j = 2, vehicle A chooses to adjust and vehicle B does not adjust; when j = 3, vehicle B chooses to adjust, and vehicle A does not adjust; when j = 4, A, Both vehicles B do not choose to adjust. At this time, the two vehicles will have a collision risk, so the virtual payment cost will be infinite. In the actual operation process, this scheme (j=4) will be directly excluded.

于是根据步骤3-1、步骤3-2及步骤3-3中的参数量化与标定，确定各联盟支付成本总量的计算公式为

Figure RE-GDA0002364112880000042
Therefore, according to the parameter quantification and calibration in step 3-1, step 3-2 and step 3-3, the calculation formula of the total cost paid by each alliance is determined as:
Figure RE-GDA0002364112880000042

所述步骤3-4输出联盟虚拟成本总量最小值，确定虚拟成本总量最小值对应的联盟调整方案过程为，根据各联盟支付成本总量计算结果进行取值大小的比较，确定最小支付成本总量的计算公式为cmin(xA,xB)＝min(cj)，从而根据 cmin(xA,xB)＝min(cj)中的j值确定所对应的联盟调整方案。The step 3-4 outputs the minimum value of the total virtual cost of the alliance, and the process of determining the alliance adjustment scheme corresponding to the minimum value of the total virtual cost is as follows: according to the calculation result of the total payment cost of each alliance, the value is compared and the minimum payment cost is determined. The calculation formula of the total amount is c min (x A , x B )=min(c j ), so the corresponding alliance adjustment scheme is determined according to the j value in c min (x A , x B )=min(c j ) .

所述步骤3-7中计算虚拟成本总量最小值对应联盟调整方案中各联盟成员贡献值的过程为，首先确定非合作博弈条件下该方案中车辆A和B需支付的总成本，公式为The process of calculating the contribution value of each alliance member in the alliance adjustment scheme corresponding to the minimum virtual cost total amount in the steps 3-7 is as follows: First, determine the total cost to be paid by vehicles A and B in the scheme under non-cooperative game conditions, and the formula is:

Figure RE-GDA0002364112880000051
Figure RE-GDA0002364112880000051

然后计算合作博弈条件下A、B两车辆共同节省的支付成本Δc，计算公式为Then calculate the payment cost Δc saved by the two vehicles A and B under the cooperative game condition, and the calculation formula is as follows:

Figure RE-GDA0002364112880000052
Figure RE-GDA0002364112880000052

最后进行虚拟成本总量最小值对应联盟调整方案中各成员贡献值的计算，比较算法为

Figure RE-GDA0002364112880000053
Finally, calculate the contribution value of each member in the alliance adjustment plan corresponding to the minimum total virtual cost. The comparison algorithm is:
Figure RE-GDA0002364112880000053

所述步骤4中根据两自动驾驶车辆的避险解脱先后顺序确定两车辆是否需要进行调整，指根据步骤1中的时间需求强度及步骤2中的车辆类型优先级所确定的排序结果，确定避险解脱顺序优先的自动驾驶车辆保持原有行驶状态不变，避险解脱顺序靠后的自动驾驶车辆根据避险解脱顺序优先的自动驾驶车辆进行避险解脱自主调整。In the step 4, it is determined whether the two vehicles need to be adjusted according to the order of escape and avoidance of the two autonomous driving vehicles. The self-driving vehicle with priority in the order of escape from danger keeps the original driving state unchanged, and the self-driving vehicle with the priority in the order of escape from danger will automatically adjust the escape from danger according to the automatic driving vehicle with priority in the order of escape from danger.

所述步骤5中需进行行为调整的自动驾驶车辆确定本车辆的调节方式的过程为，利用概率削减法建立自动驾驶车意图概率更新模型。The process of determining the adjustment mode of the self-driving vehicle for the self-driving vehicle that needs to perform behavior adjustment in the step 5 is to use the probability reduction method to establish an intention probability update model of the self-driving vehicle.

避险解脱顺序靠后的自动驾驶车辆面对可选择行为调整时，以自动驾驶车辆意图量化值虚拟相等为初始状态，实时改变自动驾驶车辆选择可实时调整行为的概率，实现自动驾驶车辆意图决策调整行为的量化表征过程。假定可供选择的调整行为分别为速度调节(VC)和车道调节(HC),且最初发生的概率均相等并加和为1，满足公式PVC＝PHC,PVC+PHC＝1。当自动驾驶车辆处于潜在冲突环境条件下，上述两种调节方式中随机出现一种调节行为的概率削减为原来的

Figure RE-GDA0002364112880000054
为了满足概率值加和为1的初始条件，则另一调节方式的概率自然发生扩大变化。When the self-driving vehicle at the back of the escape sequence faces the optional behavior adjustment, it takes the virtual equal value of the self-driving vehicle's intention quantification value as the initial state, and changes the probability of real-time adjustment of the behavior of the self-driving vehicle selection in real time, so as to realize the decision-making of the self-driving vehicle's intention. Quantitative characterization of tuning behavior. Assuming that the optional adjustment behaviors are speed adjustment (VC) and lane adjustment (HC), and the probabilities of initial occurrence are equal and add up to 1, the formulas P VC =P HC , P VC +P HC =1 are satisfied. When the autonomous vehicle is in a potentially conflicting environment, the probability of randomly appearing one adjustment behavior in the above two adjustment methods is reduced to the original one.
Figure RE-GDA0002364112880000054
In order to satisfy the initial condition that the sum of the probability values is 1, the probability of another adjustment method naturally expands and changes.

概率削减过程为

Figure RE-GDA0002364112880000061
满足概率为1则有P'VC＝1-P'HC。或有概率削减过程为
Figure RE-GDA0002364112880000062
满足概率为1则有P”HC＝1-P”VC。The probability reduction process is
Figure RE-GDA0002364112880000061
If the probability is 1, then P' VC =1-P' HC . The contingent probability reduction process is
Figure RE-GDA0002364112880000062
If the probability is 1, then P” HC =1-P” VC .

当自动驾驶车辆调节意图概率关系满足P'VC＞P'HC或P”VC＞P”HC时，则对应车辆进行速度调节，否则进行车道调节。When the probability relationship of the adjustment intention of the automatic driving vehicle satisfies P' VC >P' HC or P" VC >P" HC , the corresponding vehicle is adjusted in speed, otherwise, the lane is adjusted.

本发明的有益效果在于：充分考虑自动驾驶车辆在实际运行过程的时间需求强度、车辆类型特性差异，实现以时间需求强度优先等级、车辆类型优先等级为基础的车辆冲突自主避险决策，结合现实条件，实用性好，能够满足自动驾驶车辆面对冲突自主决策避险的要求。The beneficial effects of the present invention are: fully considering the time demand intensity and vehicle type characteristic differences of the automatic driving vehicle in the actual operation process, and realizing the vehicle conflict autonomous risk avoidance decision based on the time demand intensity priority level and the vehicle type priority level, combined with reality conditions, good practicability, and can meet the requirements of autonomous vehicles for autonomous decision-making and risk avoidance in the face of conflicts.

附图说明Description of drawings

图1为本发明实施例提供的冲突避险自主决策方法的具体实施流程图；FIG. 1 is a specific implementation flowchart of a conflict risk avoidance autonomous decision-making method provided by an embodiment of the present invention;

图2为本发明实施例提供的基于合作博弈理论的自动驾驶车辆行驶意图交互冲突避险决策流程图；FIG. 2 is a flow chart of a risk-avoidance decision-making process based on cooperative game theory for autonomous driving vehicle driving intention interaction conflict provided by an embodiment of the present invention;

图3为本发明实施例仿真过程中各参数分析变化示意图；FIG. 3 is a schematic diagram of analysis and change of each parameter in the simulation process according to the embodiment of the present invention;

图4为本发明实施例仿真过程中合作博弈理论成本节省率对比图。FIG. 4 is a comparison diagram of the cost saving rate of cooperative game theory in the simulation process of the embodiment of the present invention.

具体实施方式Detailed ways

下面结合附图和具体实施例对本发明提供的适用于自动驾驶车辆的冲突避险自主决策方法进行详细描述。本实施例只作为本发明技术中一种情况的说明，并不能以本实施例为限制缩小本发明专利的保护范围。The method for autonomous decision-making for conflict avoidance suitable for autonomous vehicles provided by the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is only used as an illustration of a situation in the technology of the present invention, and the protection scope of the patent of the present invention cannot be narrowed by using this embodiment as a limitation.

以下将结合附图对本发明的实施例进行详细说明。The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

图1示出了本发明实施例所提供的一种自动驾驶车辆冲突避险自主决策方法流程图。如图1所示本发明实施例提供的自动驾驶车辆冲突避险自主决策方法包括以下步骤：FIG. 1 shows a flowchart of an autonomous decision-making method for conflict avoidance of an autonomous vehicle provided by an embodiment of the present invention. As shown in FIG. 1 , the autonomous decision-making method for conflict avoidance of an autonomous vehicle provided by an embodiment of the present invention includes the following steps:

首先已知自动驾驶车辆的行驶状态和行驶环境，并且各自动驾驶车辆的行驶状态、运输任务、车辆类型等均能够被其他车辆所获取，包括各自动驾驶车辆之间的实时距离、飞行速度及距离潜在冲突点的距离等参数均可悲获取，并且两自动驾驶车辆之间能够实现高质量信息交互。First, the driving status and driving environment of the autonomous vehicles are known, and the driving status, transportation tasks, and vehicle types of each autonomous vehicle can be obtained by other vehicles, including the real-time distance, flight speed and Parameters such as the distance to the potential conflict point can be obtained, and high-quality information exchange can be achieved between the two autonomous vehicles.

步骤1：基于车辆时间需求强度对两车辆避险解脱优先等级进行顺序排列。Step 1: Rank the priority levels of the two vehicles in order based on the time demand intensity of the vehicles.

本步骤中对两自动驾驶车辆依据时间需求强度对其进行冲突避险顺序排列，其依据是考虑整个传递和反应过程的最短时间间隔。在此根据已有自动驾驶车辆参数值的相关研究，以0.6s为参考，对自动驾驶车辆与实践需求强度的比较关系进行进一步描述。In this step, the two autonomous driving vehicles are arranged in a conflict avoidance order according to the time demand intensity, which is based on the consideration of the shortest time interval of the entire transmission and response process. Here, according to the relevant research on the parameter values of the existing autonomous driving vehicles, with 0.6s as the reference, the comparison relationship between the autonomous driving vehicles and the actual demand intensity is further described.

当

Figure RE-GDA0002364112880000071
时，定义两自动驾驶车辆时间需求强度相同；当
Figure RE-GDA0002364112880000072
时，即τA＞τB，说明编号为A的自动驾驶车辆时间需求强度大于编号为B的自动驾驶车辆时间需求强度，从而确定编号为A的自动驾驶车辆排序先于编号为B 的自动驾驶车辆；
Figure RE-GDA0002364112880000073
时，即τA＜τB时，说明编号为A的自动驾驶车辆时间需求强度小于编号为B的自动驾驶车辆时间需求强度，从而确定编号为B 的自动驾驶车辆排序先于编号为A的自动驾驶车辆。when
Figure RE-GDA0002364112880000071
, the time demand intensity of the two autonomous vehicles is defined to be the same; when
Figure RE-GDA0002364112880000072
When τ A > τ B , it means that the time demand intensity of the self-driving vehicle numbered A is greater than the time demand intensity of the self-driving vehicle numbered B, so it is determined that the automatic driving vehicle numbered A is ranked ahead of the automatic driving vehicle numbered B. vehicle;
Figure RE-GDA0002364112880000073
When τ A <τ B , it means that the time demand intensity of the autonomous vehicle numbered A is less than the time demand intensity of the autonomous driving vehicle numbered B, so it is determined that the automatic driving vehicle numbered B is ranked before the automatic driving vehicle numbered A. drive a vehicle.

两自动驾驶车辆的时间需求强度比较结果将会出现两自动驾驶车辆时间需求强度相同及不同两种结果。The comparison results of the time demand intensity of the two autonomous vehicles will show two results of the same and different time demand intensity of the two autonomous vehicles.

如果自动驾驶车辆时间需求强度不同，则根据时间需求强度关系确定自动驾驶车辆的冲突避险排列顺序，然后进入到步骤4，实现冲突避险决策顺序优先的自动驾驶车辆保持原有行驶状态不变，冲突避险决策顺序靠后的自动驾驶车辆进行行驶行为调整。If the time demand intensity of the self-driving vehicles is different, determine the conflict avoidance order of the self-driving vehicles according to the time demand intensity relationship, and then go to step 4, and the self-driving vehicles with the priority in the conflict risk avoidance decision-making order keep the original driving state unchanged. , the self-driving vehicle that is later in the conflict avoidance decision-making sequence adjusts its driving behavior.

如果两自动驾驶车辆时间需求强度相同，则进入到步骤2进行基于车辆类型的自动驾驶车辆冲突避险顺序排列。If the time demand intensity of the two autonomous driving vehicles is the same, then go to step 2 to arrange the collision avoidance sequence of the autonomous driving vehicles based on the vehicle type.

步骤2：基于车辆类型对两车辆避险解脱优先等级进行顺序排列。Step 2: Arrange the priority levels of the two vehicles to avoid danger based on the vehicle type.

本步骤中基于车辆类型对两车辆避险解脱优先等级进行顺序排列，对潜在冲突的两自动驾驶车辆进行车辆类型优先等级确定。本实施例根据长途巴士、救援车辆、普通社会车辆等常见的车辆类型及承载任务，确定各车辆类型的优先等级序列为救援车辆＞长途巴士＞通社会车辆。In this step, the priority levels of the two vehicles for avoiding danger are sequentially arranged based on the vehicle types, and the vehicle type priority levels are determined for the two potentially conflicting autonomous driving vehicles. In this embodiment, according to common vehicle types and carrying tasks such as long-distance buses, rescue vehicles, and ordinary social vehicles, the priority sequence of each vehicle type is determined as rescue vehicles>long-distance buses>social vehicles.

两自动驾驶车辆类型优先等级的比较结果将会有两种，一种是两车辆类型优先等级相同，一种是两车辆类型优先等级不同。There will be two results of the comparison of the priority levels of the two types of autonomous vehicles, one is that the priority levels of the two vehicle types are the same, and the other is that the priority levels of the two vehicle types are different.

如果两车辆类型优先等级不同，则根据车辆类型有限等级确定各车辆冲突避险排列顺序，并转入步骤4，实现冲突避险决策顺序优先的自动驾驶车辆保持原有行驶状态不变，冲突避险决策顺序靠后的自动驾驶车辆进行行驶行为调整。If the priority levels of the two vehicle types are different, determine the collision avoidance order of each vehicle according to the limited level of vehicle types, and go to step 4 to realize that the automatic driving vehicle with the priority in the conflict avoidance decision order keeps the original driving state unchanged, and the conflict avoidance The self-driving vehicle at the lower end of the risk decision-making sequence adjusts its driving behavior.

如果两车辆类型优先等级相同，则进入到步骤3进行两车辆行驶意图相互。If the priority levels of the two vehicle types are the same, go to step 3 to carry out the mutual driving intentions of the two vehicles.

步骤3：车辆行驶意图交互。Step 3: Interaction of vehicle driving intent.

本步骤中车辆行驶意图交互过程的实现，主要以合作博弈理论为基础，将两车辆行驶意图支付成本量化结果输入到联盟支付成本总函数中，计算各联盟支付成本中的最小值，根据支付成本最小值确定所对应调整联盟并确定该联盟所对应车辆的调整方案。The realization of the interaction process of vehicle driving intention in this step is mainly based on the cooperative game theory. The quantification result of the payment cost of the driving intention of the two vehicles is input into the total payment cost function of the alliance, and the minimum value of the payment cost of each alliance is calculated. According to the payment cost The minimum value determines the corresponding adjustment alliance and determines the adjustment scheme of the vehicle corresponding to the alliance.

图2示出了本发明实施例所提供的自动驾驶车辆冲突避险自主决策方法中基于合作博弈理论的车辆行驶意图交互冲突避险决策流程图。如图2所示本发明实施例提供的车辆行驶意图交互冲突避险决策包括以下步骤：FIG. 2 shows a flow chart of the decision-making process of vehicle driving intention interaction conflict risk avoidance based on cooperative game theory in the autonomous vehicle conflict risk avoidance decision-making method provided by the embodiment of the present invention. As shown in FIG. 2 , the vehicle driving intention interaction conflict risk avoidance decision provided by the embodiment of the present invention includes the following steps:

步骤3-1：输入各联盟成员支付成本。Step 3-1: Enter the cost to be paid by each alliance member.

本步骤中输入各联盟成员支付成本，其中支付成本来源方程为

Figure RE-GDA0002364112880000091
In this step, input the payment cost of each alliance member, where the payment cost source equation is
Figure RE-GDA0002364112880000091

步骤3-2：车辆行驶意图量化。Step 3-2: Quantification of vehicle driving intention.

本步骤中利用随机函数对车辆行驶意图进行量化，确定自动驾驶车辆A和自动驾驶车辆B的支付系数，若车辆在行驶过程中希望进行行为调整则次支付系数确定为1，若车辆不希望进行行为调整则次支付系数确定为一个大于1的数。In this step, a random function is used to quantify the driving intention of the vehicle, and the payment coefficient of the automatic driving vehicle A and the automatic driving vehicle B is determined. Behavior adjustment then the subpayment factor is determined as a number greater than 1.

步骤3-3：计算各联盟虚拟成本总量。Step 3-3: Calculate the total virtual cost of each alliance.

本步骤中计算各联盟虚拟成本总量，根据步骤3-1和步骤3-2中的联盟成员支付成本及车辆行驶意图量化结果，确定各联盟虚拟成本总量计算公式为In this step, the total virtual cost of each alliance is calculated, and according to the payment cost of the alliance members in steps 3-1 and 3-2 and the quantification result of the vehicle driving intention, the calculation formula of the total virtual cost of each alliance is determined as:

Figure RE-GDA0002364112880000092
Figure RE-GDA0002364112880000092

步骤3-4：确定虚拟成本总量最小值对应的联盟调整方案。Step 3-4: Determine the alliance adjustment scheme corresponding to the minimum value of the total virtual cost.

本步骤首先要对各成本支付总量进行比较，确定输出虚拟成本总量的最小值，计算公式为cmin(xA,xB)＝min(cj)。在此确定联盟虚拟成本最小值的基础上进一步确定该支付成本所对应的联盟调整方案。In this step, the total amount of each cost payment is firstly compared, and the minimum value of the total output virtual cost is determined. The calculation formula is c min (x A , x B )=min(c j ). On the basis of determining the minimum value of the alliance virtual cost, the alliance adjustment scheme corresponding to the payment cost is further determined.

步骤3-5：确定虚拟成本总量最小值对应联盟调整方案中是否需进行两车辆均做调整。Step 3-5: Determine whether both vehicles need to be adjusted in the alliance adjustment plan corresponding to the minimum total virtual cost.

在本步骤中，确定虚拟支付成本总量最小值对应的联盟调整方案，其中调整方案主要包括，两车辆同时进行调整及只有其中一车辆进行调整两类结果。In this step, the alliance adjustment scheme corresponding to the minimum value of the total virtual payment cost is determined, wherein the adjustment scheme mainly includes two types of results: two vehicles are adjusted at the same time and only one vehicle is adjusted.

如果对应联盟调整方案为只有一辆车进行调整，则转入到步骤3-6。如果对应联盟调整方案为两车辆均需要调整，则转入到步骤3-7。If only one vehicle is to be adjusted in the corresponding alliance adjustment plan, go to step 3-6. If the corresponding alliance adjustment plan is that both vehicles need to be adjusted, go to step 3-7.

步骤3-6：确定不需要调整的自动驾驶车辆冲突避险顺序优先，确定需要调整的自动驾驶车辆冲突避险顺序靠后。Step 3-6: It is determined that the automatic driving vehicle conflict avoidance order that does not need to be adjusted is prioritized, and the conflict risk avoidance order of the automatic driving vehicle that needs to be adjusted is determined to be later.

在本步骤中确定两车辆冲突避险顺序，主要根据步骤3-5中联盟调整方案确定两车辆是否调整的方案来确定。其中不需要进行行驶行为调整的车辆冲突避险顺序靠前，需要调整的车辆冲突避险顺序靠后，冲突避险顺序确定后转入到步骤3-9。In this step, the collision avoidance sequence of the two vehicles is determined, which is mainly determined according to the scheme of determining whether the two vehicles are adjusted in the alliance adjustment scheme in steps 3-5. Among them, the vehicles that do not need to adjust the driving behavior are in the front of the collision avoidance order, and the vehicles that need to be adjusted are in the back.

步骤3-7计算虚拟成本总量最小值对应调整方案中联盟成员贡献值。Steps 3-7 calculate the contribution value of alliance members in the adjustment plan corresponding to the minimum total virtual cost.

本步骤中计算调整方案中各联盟成员的贡献值，首先要计算非合作博弈条件下最小支付成本对应方案中车辆A和车辆B需支付的总成本，其计算公式为

Figure RE-GDA0002364112880000101
然后计算合作博弈条件下A、B两车辆共同节省的支付成本Δc，计算公式为
Figure RE-GDA0002364112880000102
最后计算各联盟成员在合作博弈过程中的贡献值，计算公式为
Figure RE-GDA0002364112880000103
In this step, the contribution value of each alliance member in the adjustment scheme is calculated. First, the total cost to be paid by vehicle A and vehicle B in the corresponding scheme of minimum payment cost under non-cooperative game conditions should be calculated. The calculation formula is:
Figure RE-GDA0002364112880000101
Then calculate the payment cost Δc saved by the two vehicles A and B under the cooperative game condition, and the calculation formula is as follows:
Figure RE-GDA0002364112880000102
Finally, the contribution value of each alliance member in the cooperative game process is calculated, and the calculation formula is:
Figure RE-GDA0002364112880000103

步骤3-8：根据各联盟成员贡献值确定自动驾驶车辆冲突避险排列顺序。Step 3-8: According to the contribution value of each alliance member, determine the order of collision avoidance of autonomous driving vehicles.

本步骤中确定自动驾驶车辆冲突避险排列顺序首先要进行各联盟成员贡献值大小的比较，计算公式为

Figure RE-GDA0002364112880000104
In this step, to determine the order of collision avoidance of autonomous vehicles, the first step is to compare the contribution value of each alliance member. The calculation formula is:
Figure RE-GDA0002364112880000104

根据比较结果，确定取值叫嚣着所对应的的自动驾驶车辆冲突避险排序优先，取值较大者所对应的自动驾驶车辆冲突避险排序靠后，冲突避险排列顺序确定后转入步骤3-9。According to the comparison result, it is determined that the collision avoidance order of the autopilot vehicle corresponding to the value clamoring is the priority, the collision avoidance order of the autopilot vehicle corresponding to the larger value is lower, and the conflict avoidance order is determined and then goes to the step 3-9.

步骤3-9：转入到两车辆冲突解脱逻辑步骤5。Step 3-9: Go to step 5 of the two-vehicle conflict resolution logic.

本步骤为将已经确定的自动驾驶车辆冲突避险排列顺序转入到步骤5周，而后进行下一步步骤。This step is to transfer the determined sequence of collision avoidance of the automatic driving vehicle to step 5, and then proceed to the next step.

步骤4：确定各车辆是否需要进行调整。Step 4: Determine if each vehicle needs to be adjusted.

根据上述各步骤进展结果，确定两车辆最终是否需进行调整，然后转入到步骤5。According to the progress results of the above steps, it is determined whether the two vehicles finally need to be adjusted, and then go to step 5.

步骤5：需进行行为调节的自动驾驶车辆根据行驶意图确定本车辆调节方式。Step 5: The self-driving vehicle that needs to perform behavior adjustment determines the adjustment method of the vehicle according to the driving intention.

在本步骤中，根据已知速度调节(VC)和车道调节(HC)最初发生的概率均相等并加和为1，满足公式PVC＝PHC,PVC+PHC＝1。在进行车辆意图确定本车辆行驶行为调整过程中将会出现

Figure RE-GDA0002364112880000111
P'VC＝1-P'HC。或出现
Figure RE-GDA0002364112880000112
P”HC＝1-P”VC。因此当自动驾驶车辆调节意图概率关系满足P'VC＞P'HC或 P”VC＞P”HC时，则对应车辆进行速度调节，否则进行车道调节。In this step, according to the known probabilities of speed regulation (VC) and lane regulation (HC), the initial occurrence probability is equal and the sum is 1, which satisfies the formula P VC =P HC , P VC +P HC =1. During the process of vehicle intent to determine the driving behavior of the vehicle, there will be
Figure RE-GDA0002364112880000111
P' VC =1-P' HC . or appear
Figure RE-GDA0002364112880000112
P" HC =1-P" VC . Therefore, when the automatic driving vehicle adjustment intention probability relationship satisfies P' VC >P' HC or P" VC >P" HC , the corresponding vehicle is adjusted in speed, otherwise, the lane is adjusted.

步骤6：自动驾驶车辆行驶行为调整。Step 6: Adjust the driving behavior of the autonomous vehicle.

本步骤的进行，指自动驾驶车辆按车辆行驶意图展开行驶行为调节，顺利避开潜在冲突点，保证车辆的安全运行。The execution of this step means that the autonomous driving vehicle adjusts the driving behavior according to the driving intention of the vehicle, smoothly avoids potential conflict points, and ensures the safe operation of the vehicle.

本发明实施例提供的车辆冲突避险决策中只是利用了两自动驾驶车辆可能存在的几种决策组合。The vehicle conflict risk avoidance decision provided by the embodiment of the present invention only utilizes several possible decision combinations of two autonomous driving vehicles.

本发明实施例以仿真实验中的车辆冲突避险决策的初始冲突状态为开始。仿真实验设定为两自动驾驶车辆的时间需求强度及车辆类型相同，并分别设定 vA60Km/h、vB60Km/h，vA60Km/h、vB40Km/h，以及lA15m、lB15m，lA15m、lB20m 等几种仿真前提。基本场景为高速公路匝道合流区附近，主车道自动驾驶车辆A 与匝道中即将进入主车道的自动驾驶车辆B之间存在一定的潜在冲突。The embodiment of the present invention starts with the initial conflict state of the vehicle conflict risk avoidance decision in the simulation experiment. The simulation experiment is set as the time demand intensity and vehicle type of the two autonomous vehicles are the same, and respectively set v A 60Km/h, v B 60Km/h, v A 60Km/h, v B 40Km/h, and l A 15m , l B 15m, l A 15m, l B 20m and other simulation conditions. The basic scenario is near the merging area of the expressway ramp, and there is a certain potential conflict between the autonomous vehicle A in the main lane and the autonomous vehicle B about to enter the main lane on the ramp.

本发明实施例在仿真实验中的车辆冲突避险决策可能存在的几种决策方式。即为匝道中即将进入主道的自动驾驶车辆B与主道自动驾驶车辆A为了避开潜在冲突在进行了一系列冲突解脱合作博弈后分别做出了冲突避让自主决策。There are several possible decision modes for vehicle conflict risk avoidance decision in the simulation experiment of the embodiment of the present invention. That is, the autonomous vehicle B and the autonomous vehicle A on the main road, which are about to enter the main road on the ramp, respectively make conflict avoidance autonomous decisions after a series of conflict resolution and cooperation games in order to avoid potential conflicts.

可能存在A进行变道调整，为B让出汇入空间，B按原有行驶状态行进，逐渐并入主道。可能存在A进行加速调整，为B让出汇入空间，B按原有行驶状态行进，逐渐并入主道。可能存在B进行减速调整，为自身汇出主道留出空间后逐渐并入主道。可能存在A进行变道调整，B进行减速调整，同时为B让出汇入空间后使其逐渐并入主道等多种调整方式There may be a lane change adjustment by A to make room for B to enter and exit, and B travels in the original driving state and gradually merges into the main road. There may be an acceleration adjustment for A to make room for B to enter and exit, and B to travel according to the original driving state and gradually merge into the main road. There may be a deceleration adjustment for B to make room for itself to exit the main road and gradually merge into the main road. There may be a variety of adjustment methods such as A for lane change adjustment, B for deceleration adjustment, and at the same time for B to gradually merge into the main road after leaving the entry space.

图3示出了本发明实施例在仿真实验中车辆冲突自主避险决策支付成本随参数变化而产生的变化趋势，也表明了自动驾驶车辆行驶调整意图与支付成本之间的高度相关性，当车辆之间行驶速度及距离潜在冲突点长度在一定范围内时，两车辆合作博弈效果越好则系统所需支付的成本也就越低。Fig. 3 shows the variation trend of the vehicle conflict autonomous risk avoidance decision payment cost with the parameter change in the simulation experiment of the embodiment of the present invention, and also shows the high correlation between the driving adjustment intention of the autonomous vehicle and the payment cost. When When the driving speed between vehicles and the length of the distance to the potential conflict point are within a certain range, the better the cooperative game effect between the two vehicles is, the lower the cost to be paid by the system.

图4示出了本发明实施例仿真实验中合作博弈理论成本节省率的对比分析图，图中表明在仿真实验的各种情形下，合作博弈理论的运用确实降低了系统的支付成本，并且很好地实现了冲突解脱避让决策。FIG. 4 shows a comparative analysis diagram of the cost saving rate of cooperative game theory in the simulation experiment of the embodiment of the present invention. Good realization of conflict resolution and avoidance decision.

Claims (10)

1. An autonomous decision-making method for collision avoidance in a highway confluence area is characterized by comprising the following steps:
step 1: comparing the danger avoiding and releasing sequence of the two vehicles based on the time demand intensity of the vehicles, if the time demand intensity of the two automatic driving vehicles is the same, turning to the step 2, if the time demand intensity of the two automatic driving vehicles is different, determining the danger avoiding and releasing sequence of the two vehicles, and then turning to the step 4;
step 2: comparing the danger avoiding and releasing priority levels of the two vehicles based on the vehicle types, if the vehicle types of the two automatic driving vehicles are consistent, turning to a step 3, if the vehicle types are different, determining the sequence of danger avoiding and releasing of the two vehicles according to the vehicle type priority levels, and then turning to a step 4;
and step 3: the two automatic driving vehicles carry out driving intention interaction, a danger avoiding and releasing scheme of the two automatic driving vehicles is determined according to a vehicle driving intention interaction result, and then the step 5 is carried out;
and 4, step 4: determining whether the two vehicles need to be adjusted according to the danger avoiding and releasing sequence of the two automatic driving vehicles, and turning to the step 5 after determining whether each vehicle needs to be adjusted;
and 5: determining the adjusting mode of the vehicle by the automatic driving vehicle needing behavior adjustment, and turning to step 6;
step 6: and adjusting the running behavior of the automatic driving vehicle.
2. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 1, characterized in that: comparing the risk avoidance and release priority levels of the two vehicles based on the time demand intensity of the vehicles, representing the time demand intensity of the ith automatic driving vehicle as tau, and
Figure RE-FDA0002364112870000011
where i-A, B, a stands for autonomous vehicle a, B stands for autonomous vehicle B, l stands for length of autonomous vehicle i from potential conflict point, v stands for current shape of autonomous vehicle iPotential velocity; the time demand intensity priority comparison is performed for two potential conflicting autonomous vehicles according to τ.
3. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 1, characterized in that: comparing the risk avoidance and release priority levels of the two vehicles based on the vehicle types, and developing the vehicle types on the basis that the time demand intensity of the two vehicles is the same in step 1, wherein the vehicle types are mainly divided from the transportation function and the operation task of the vehicle types, and the vehicle type priority levels of the two potentially conflicting automatic driving vehicles are determined.
4. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 1, characterized in that: the interaction method for the driving intentions of the two automatic driving vehicles is developed on the basis that the time requirement intensity in the step 1 is the same and the vehicle type priority level in the step 2 is the same, and comprises the following steps of:
step 3-1: inputting payment cost of each alliance;
step 3-2: quantifying the adjustment intention of the automatic driving vehicle by using a random function, and inputting the adjustment intention of the automatic driving vehicle and the payment cost of each alliance into the calculation of a total virtual cost function;
step 3-3: calculating the total virtual cost of each alliance, and comparing the total virtual cost of each alliance;
step 3-4: outputting the minimum value of the virtual total cost of the alliance and determining an alliance adjusting scheme corresponding to the minimum value of the virtual total cost;
step 3-5: judging whether the adjustment scheme corresponding to the minimum value of the total virtual cost needs to be adjusted by two vehicles at the same time, if not, turning to the step 3-6, and if so, turning to the step 3-7;
step 3-6: determining that the conflict resolution sequence is arranged in front of the vehicles which do not need to be adjusted, determining that the conflict resolution sequence is arranged in back of the vehicles which need to be adjusted, and switching to the step 3-9 after the conflict resolution sequence is determined;
step 3-7: calculating contribution values of all the coalition members in the coalition adjustment scheme corresponding to the minimum value of the total virtual cost;
step 3-8: comparing the contribution values of all the coalition members in the adjustment scheme, determining that the conflict resolution sequence of the automatic driving vehicles corresponding to the smaller contribution value is arranged in front, determining that the conflict resolution sequence of the automatic driving vehicles corresponding to the larger contribution value is arranged in back, and switching to the step 3-9 after the conflict resolution sequence is determined;
step 3-9: and (4) switching to a two-vehicle conflict resolution logic step 4.
5. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 4, characterized in that: the method for inputting the payment cost of each alliance in the step 3-1 is specifically that the payment cost x of each alliance memberiWherein i is A, B, a represents the automatic driving vehicle a, B represents the automatic driving vehicle B, and the length l of the automatic driving vehicle from the potential conflict point is considered according to the position relation and the driving speed between the two automatic driving vehiclesiThe farther away the autonomous vehicle is, the safer it is, xiWith liIs increased and decreased; speed v of traveliThe larger the autonomous vehicle will be closer to the conflict point in a short time, xiWith viIs increased by increasing the action weight k of the two parameters1,k2Determining the origin equation of the payment cost of the coalition members as
Figure RE-FDA0002364112870000031
6. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 4, characterized in that: in the step 3-2, the method for quantifying the driving intention of the automatic driving vehicle specifically comprises the step of utilizing a random function to realize the quantifying process of the driving intention of the automatic driving vehicle, firstly utilizing parameters a and B to represent the adjustment intention of the two automatic driving vehicles, thereby determining that a and B are the payment coefficients of the automatic driving vehicle A and the automatic driving vehicle B, if the automatic driving vehicle wants to perform behavior adjustment, the value of the payment coefficient is 1, and if the automatic driving vehicle does not want to perform behavior adjustment, the value of the payment coefficient is more than 1.
7. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 4, characterized in that: step 3-3, a method for calculating the total virtual cost of each alliance is specifically that 4 crossed combinations of selectable adjustment behaviors of two automatic driving vehicles are determined, j is determined as a corresponding serial number of each alliance, and when j is set to be 1, A, B two vehicles are selected to be adjusted; when j is 2, the vehicle A selects adjustment, and the vehicle B does not perform adjustment; when j is 3, the vehicle B selects adjustment, and the vehicle A does not perform adjustment; when j is 4, A, B neither vehicle selects adjustment, and at this time, both vehicles will have collision risk, so the virtual payment cost will be infinite, and this scheme (j is 4) will be directly excluded in the actual operation process.
8. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 4, characterized in that: outputting the minimum value of the virtual total cost of the alliances and determining the alliance adjustment scheme corresponding to the minimum value of the virtual total cost in the step 3-4, wherein the calculation formula of the total amount of the alliance payment cost is
Figure RE-FDA0002364112870000032
Comparing the values according to the calculated value of the total payment cost of each alliance, and determining the calculation formula of the minimum total payment cost as cmin(xA,xB)＝min(cj) And further determining the alliance adjustment scheme corresponding to the minimum payment cost total amount.
9. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 4, characterized in that: and 3-7, calculating the contribution value of each coalition member in the coalition adjustment scheme corresponding to the minimum value of the total virtual cost, specifically, determining the total cost to be paid by the vehicles A and B in the scheme under the non-cooperative game condition by the formula
Figure RE-FDA0002364112870000041
And then calculating A, B payment cost deltac saved by two vehicles under the cooperative game condition, wherein the calculation formula is
Figure RE-FDA0002364112870000042
Finally, the contribution value of each member in the coalition adjustment scheme corresponding to the minimum value of the total virtual cost is calculated, and the comparison algorithm is
Figure RE-FDA0002364112870000043
10. The highway confluence area collision risk avoidance autonomous decision-making method according to claim 1, characterized in that: a method for determining the regulation mode of the vehicle by the automatically driven vehicle needing behavior regulation specifically includes utilizing probability subtraction method to establish an intention probability updating model of the automatically driven vehicle, supposing that the optional regulation behaviors are speed regulation (VC) and lane regulation (HC), the initial probabilities are equal and are added to be 1, and satisfying formula PVC＝PHC,PVC+PHC1. When the automatic driving vehicle is in the condition of potential conflict environment, the probability of one regulation behavior randomly appearing in the two regulation modes is reduced to the original one
Figure RE-FDA0002364112870000044
In order to meet the initial condition that the sum of the probability values is 1, the probability of the other adjusting mode naturally expands and changes;
wherein the probability reduction process is
Figure RE-FDA0002364112870000045
P 'if the probability of satisfaction is 1'VC＝1-P'HC(ii) a Or with probability the reduction process is
Figure RE-FDA0002364112870000046
If the probability of 1 is satisfied, P is "HC＝1-P”VC；
When the automatic driving vehicle adjusts the intention probability relation to satisfy P'VC＞P'HCOr P'VC＞P”HCAnd if not, adjusting the speed of the corresponding vehicle, otherwise, adjusting the lane.
CN201911115216.1A 2019-11-14 2019-11-14 An autonomous decision-making method for conflict risk avoidance in expressway merging areas Pending CN110867097A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911115216.1A CN110867097A (en) 2019-11-14 2019-11-14 An autonomous decision-making method for conflict risk avoidance in expressway merging areas

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911115216.1A CN110867097A (en) 2019-11-14 2019-11-14 An autonomous decision-making method for conflict risk avoidance in expressway merging areas

Publications (1)

Publication Number Publication Date
CN110867097A true CN110867097A (en) 2020-03-06

Family

ID=69654031

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911115216.1A Pending CN110867097A (en) 2019-11-14 2019-11-14 An autonomous decision-making method for conflict risk avoidance in expressway merging areas

Country Status (1)

Country Link
CN (1) CN110867097A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111583640A (en) * 2020-04-30 2020-08-25 重庆大学 On-ramp cooperative merging control method based on vehicle-road coordination
CN111627205A (en) * 2020-04-28 2020-09-04 南京理工大学 Automatic driving vehicle and non-motor vehicle right of way distribution method based on collision risk
US11644835B2 (en) * 2020-07-29 2023-05-09 Toyota Research Institute, Inc. Game-theoretic planning for risk-aware interactive agents

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101573994A (en) * 2006-10-23 2009-11-04 泰克迪亚科技公司 A roadside network unit using local peer groups as network groups
US20130265175A1 (en) * 2012-04-04 2013-10-10 Mando Corporation Parking control apparatus and method for providing an alarm thereof
CN103700251A (en) * 2013-11-27 2014-04-02 东南大学 Variable speed limiting and ramp control coordination and optimization control method on expressway
CN110009940A (en) * 2019-05-13 2019-07-12 南京航空航天大学 A general aviation mixed operation flight conflict resolution method and system

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101573994A (en) * 2006-10-23 2009-11-04 泰克迪亚科技公司 A roadside network unit using local peer groups as network groups
US20130265175A1 (en) * 2012-04-04 2013-10-10 Mando Corporation Parking control apparatus and method for providing an alarm thereof
CN103700251A (en) * 2013-11-27 2014-04-02 东南大学 Variable speed limiting and ramp control coordination and optimization control method on expressway
CN110009940A (en) * 2019-05-13 2019-07-12 南京航空航天大学 A general aviation mixed operation flight conflict resolution method and system

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
MIN YANG ET AL.: ""Exploring the impact or residential relocation on modal shift in commute trips Evidence from a quasi-longitudinal analysis"", 《TRANSPORT POLICY》 *
宋威龙: ""城区动态环境下智能车辆行为决策研究"", 《博士学位论文全文数据库工程科技Ⅱ辑》 *
江欣国 等: ""基于元胞自动机的高速公路临时瓶颈交通流仿真"", 《西南交通大学学报》 *
盘意伟: "" 快速路入口匝道合流区的瓶颈形成机理与特性研究"", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111627205A (en) * 2020-04-28 2020-09-04 南京理工大学 Automatic driving vehicle and non-motor vehicle right of way distribution method based on collision risk
CN111627205B (en) * 2020-04-28 2022-06-07 南京理工大学 Method for assigning right of way between autonomous vehicles and non-motor vehicles
CN111583640A (en) * 2020-04-30 2020-08-25 重庆大学 On-ramp cooperative merging control method based on vehicle-road coordination
CN111583640B (en) * 2020-04-30 2022-04-22 重庆大学 On-ramp cooperative merging control method based on vehicle-road coordination
US11644835B2 (en) * 2020-07-29 2023-05-09 Toyota Research Institute, Inc. Game-theoretic planning for risk-aware interactive agents

Similar Documents

Publication Publication Date Title
CN110750877B (en) Vehicle following behavior prediction method under Apollo platform
Zhang et al. A game theoretic model predictive controller with aggressiveness estimation for mandatory lane change
Talebpour et al. Influence of connected and autonomous vehicles on traffic flow stability and throughput
CN106874597B (en) highway overtaking behavior decision method applied to automatic driving vehicle
CN114265398B (en) Automatic driving vehicle track planning method and device, electronic equipment and storage medium
Wang et al. Velocity control in car-following behavior with autonomous vehicles using reinforcement learning
CN111222630A (en) An autonomous driving rule learning method based on deep reinforcement learning
CN110867097A (en) An autonomous decision-making method for conflict risk avoidance in expressway merging areas
Wei et al. Game theoretic merging behavior control for autonomous vehicle at highway on-ramp
Nakka et al. A multi-agent deep reinforcement learning coordination framework for connected and automated vehicles at merging roadways
CN114987538A (en) Collaborative lane changing method considering multi-objective optimization in internet automatic driving environment
Li et al. Research on car-following control and energy management strategy of hybrid electric vehicles in connected scene
CN108711285B (en) Hybrid traffic simulation method based on road intersection
CN114802306A (en) Intelligent vehicle integrated decision-making system based on man-machine co-driving concept
CN117671978A (en) Vehicle formation and signal lamp cooperative control method and system based on multi-agent reinforcement learning
CN114475607B (en) Method and device for changing lanes for automatic driving vehicle, vehicle and storage medium
Wang et al. Resolution strategies for cooperative vehicle fleets for reducing rear-end collision risks near recurrent freeway bottlenecks
CN114492043A (en) A Personalized Driver-Following Modeling Approach Considering Perceptually Restricted Properties
Hu et al. Multi-objective optimization for autonomous driving strategy based on Deep Q Network
CN110853335B (en) An autonomous decision-making method for collision avoidance and risk avoidance for cooperative fleets in highway bottleneck sections
CN110390398A (en) On-line study method
WO2022044210A1 (en) Driving assistance device, learning device, driving assistance method, driving assistance program, learned model generation method, and learned model generation program
Liu et al. Eco-driving strategy for connected automated vehicles in mixed traffic flow
Elallid et al. Vehicles control: Collision avoidance using federated deep reinforcement learning
Wen et al. Trafficmcts: A closed-loop traffic flow generation framework with group-based monte carlo tree search

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200306