CN110992695B - Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution - Google Patents
Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution Download PDFInfo
- Publication number
- CN110992695B CN110992695B CN201911280040.5A CN201911280040A CN110992695B CN 110992695 B CN110992695 B CN 110992695B CN 201911280040 A CN201911280040 A CN 201911280040A CN 110992695 B CN110992695 B CN 110992695B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- conflict
- intersection
- vehicles
- decision
- 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.)
- Active
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000003993 interaction Effects 0.000 claims abstract description 30
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 230000001133 acceleration Effects 0.000 claims abstract description 19
- 230000009471 action Effects 0.000 claims abstract description 9
- 230000002068 genetic effect Effects 0.000 claims abstract description 8
- 238000004458 analytical method Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 13
- 206010039203 Road traffic accident Diseases 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 230000008901 benefit Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 5
- 230000009916 joint effect Effects 0.000 claims 2
- 230000003542 behavioural effect Effects 0.000 abstract 1
- 230000002452 interceptive effect Effects 0.000 description 37
- 230000006399 behavior Effects 0.000 description 29
- 238000005516 engineering process Methods 0.000 description 8
- 238000011160 research Methods 0.000 description 8
- 238000011161 development Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 6
- 238000012795 verification Methods 0.000 description 5
- 230000002860 competitive effect Effects 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 230000006855 networking Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241000408659 Darpa Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/16—Anti-collision systems
- G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
本发明公开了一种基于冲突消解的智能车辆城市交叉口通行决策多目标优化模型，以交叉口目标车辆的穿越意愿高低及冲突车辆的的驾驶类型，应用模糊逻辑方法建立车辆穿越交互关系判定模型，基于此进行交叉口冲突消解的决策模型建立，以加速度作为决策变量控制车辆，将交叉口冲突消解问题转化为一个带约束的多目标优化问题，通过引入可变合作系数，完成交叉口不同交互方式下的无人驾驶车辆动作生成结果的优化；采用改进型非劣遗传算法NSGA‑II求解Pareto最优解集，然后采用熵权TOPSIS对Pareto最优解集进行多属性决策分析，根据决策者的偏好挑选出最优解。将冲突车辆的行驶时间、行驶速度得到优化，提高行为决策水平。
The invention discloses a multi-objective optimization model for intelligent vehicle city intersection passing decision based on conflict resolution. According to the crossing willingness of the intersection target vehicle and the driving type of the conflicting vehicle, a fuzzy logic method is applied to establish a vehicle crossing interaction relationship determination model. , based on this, a decision-making model for intersection conflict resolution is established. The vehicle is controlled by acceleration as a decision variable, and the intersection conflict resolution problem is transformed into a multi-objective optimization problem with constraints. By introducing a variable cooperation coefficient, the different interactions at the intersection are completed. The optimization of the results of the unmanned vehicle action generation under the mode of operation; using the improved non-inferior genetic algorithm NSGA-II to solve the Pareto optimal solution set, and then using the entropy weight TOPSIS to perform multi-attribute decision analysis on the Pareto optimal solution set, according to the decision maker. preference to select the optimal solution. The driving time and driving speed of conflict vehicles are optimized to improve the level of behavioral decision-making.
Description
Technical Field
The invention belongs to the technical field of intelligent traffic systems and intelligent vehicle research, relates to driving interactive behavior classification, and relates to a method for establishing a vehicle lane change interactive relationship judgment model and a vehicle crossing interactive relationship judgment model by applying a fuzzy logic method, verifying the effectiveness of the crossing interactive relationship judgment model through China intelligent automobile races and real road real vehicle data, helping an unmanned vehicle to understand and judge the behavior of vehicles with people around, and having important significance for improving the autonomous driving level of the unmanned vehicle.
Background
The development of unmanned technology has become a global consensus, and has great potential in solving traffic safety and treating traffic congestion. The department of transportation in the united states of america in 2014 proposes "ITS strategic plan 2015-. The european union committee proposed the "GEAR 2030 strategy" in 2015, focusing on the advancement and cooperation in the fields of highly automated and networked driving. In 2014, the Japan pavilion combines a plurality of government departments, Toyota and other major automobile enterprises, and the aim of realizing the marketization of the completely unmanned system in 2020 is provided, so that Europe, America and Japan respectively promote the development of unmanned and intelligent network-linked automobile technologies from national strategic planning. In China, policies related to car networking and unmanned driving are brought out one after another, intelligent networked cars are clearly proposed in ' China manufacturing 2025 ' in 2015 to be key development contents, a ' energy-saving and new energy car technology route map ' published in ' 2016 to be published by the Chinese car engineering to clearly define short-term, medium-term and long-term targets of development of a Chinese intelligent networked car technology route, a ' car networking development innovation action plan ' in 2017 Ministry of industry and belief is brought out to accelerate research and development and standard formulation of car networking technologies, and development of intelligent car industry has already risen to the national level. The intelligent networked automobile is further developed and applied in the directions of networked cooperative sensing, networked cooperative decision and control, conditional automatic driving, full automatic driving and the like.
Although the research on the unmanned vehicle related technology has made great progress, enabling low-speed driving in a limited area of an urban road, and autonomous driving in a simple environment of an expressway, the application of unmanned in the automotive field is also becoming problematic, so that people are still in worry and apprehension about the unmanned technology. According to statistics of road safety records of manned and unmanned vehicles by the university of michigan in traffic research, research shows that unmanned vehicles are more prone to accidents than manned vehicles, the accident ratio of the unmanned vehicles to the manned vehicles is 9.1:1.9, and the rear-end collision accident ratio of the unmanned vehicles is 50% higher than that of the manned vehicles. For example, in 2016, a traffic accident occurred when Tesla Model S had initiated autonomous driving in California, USA, which was the earliest reported traffic accident due to an error in the autonomous driving program. In 2018, in 3 months, a Uber unmanned test vehicle crashes into a pedestrian in arizona to cause death, which is the death accident caused by the pedestrian being crashed together with the automatic driving vehicle, and after the accident, the Uber suspends the automatic driving test work. Therefore, the automatic driving function is not mature, and the decision control system still has potential safety hazards.
In addition, in the united states DARPA urban challenge race, "chinese intelligent vehicle future challenge race (IVFC)", "chinese intelligent vehicle race (CIVC)", "world intelligent driving challenge race (WIDC)", when facing the interference of the manned vehicles, the racing vehicles mostly adopt the conservative driving behaviors of deceleration or waiting to avoid conflict, the influence of the dynamic interactive behaviors of other vehicles is not considered, the difference influence of the interactive behaviors of different types of drivers is ignored, and the intersection passing potential of the unmanned vehicles is greatly limited.
At present, the intelligent degree of an automobile has realized full automatic driving under a limited condition, the automation level basically reaches the level of L3, but the unmanned level at the level of L4 and L5 has a certain distance, so that the unmanned vehicle needs to realize real road traveling in the future and faces a mixed traffic environment coexisting with a manned vehicle, and the unmanned vehicle needs to have an effective interaction mechanism, realize cooperative driving of the manned vehicle and the unmanned vehicle in the mixed traveling environment, and fully exert the advantages of accurate control, internet communication and the like.
Aiming at the problems, when a person and an unmanned automobile share the road resources, the road resources can be effectively utilized only by realizing the interactive cooperation of the person and the unmanned automobile, and traffic accidents are prevented. The research combines the national key research and development plan subject 'quantitative evaluation technology research of environmental adaptability of the automatic driving electric automobile', and establishes an intersection crossing interactive relation judgment model by taking a vehicle-vehicle interactive relation under the condition that people and unmanned vehicles are mixed to run as a key link of safe interactive driving, and the intersection crossing interactive relation judgment model is expressed by vehicle cooperation competition degree to help the unmanned vehicles to understand and judge the behavior of the people and vehicles around.
Disclosure of Invention
The invention aims to provide a conflict resolution-based multi-objective optimization method for traffic decision of urban intersections of vehicles. The vehicle crossing interactive relation judgment model is based on a fuzzy logic method, comprehensively considers the crossing desire of two interactive parties and the types of drivers of other vehicles, and the cooperation competition degree between vehicles is subjected to fuzzy reasoning, and a mapping relation between a fuzzy reasoning output value and an interactive relation prediction result is established so as to analyze vehicle-vehicle interactive behaviors under mixed conditions, help unmanned vehicles to understand and judge the behaviors of people in the periphery, and lay a foundation for realizing multi-vehicle cooperative driving.
The invention aims to provide a conflict resolution-based multi-objective optimization method for a traffic decision of an urban intersection of a vehicle, which converts the conflict resolution problem of the intersection into a constrained multi-objective optimization problem (MOP). The model is supported by a virtual simulation experiment, a decision rule for analyzing and extracting mutual avoidance of manned vehicles and unmanned vehicles at the intersection is taken as a research entry point, the maximization of the overall driving benefit of the system is taken as a target, the acceleration is taken as a decision variable, a multi-objective optimization model for conflict resolution of vehicles at the intersection is established, and a solid foundation is provided for a decision system of intelligent driving vehicles.
The invention provides a conflict resolution-based multi-objective optimization method for traffic decision of urban intersections of vehicles, which comprises the following steps:
the vehicle crossing interaction model is characterized in that a fuzzy logic method is applied to establish a vehicle crossing interaction relationship judgment model according to crossing intention of crossing target vehicles and driving types of conflicting vehicles, the degree of cooperation competition between two interactive parties and the types of drivers of other vehicles is considered, and the mapping relationship between a fuzzy reasoning output value and an interaction relationship prediction result is established to represent the cooperation and competition relationship of other vehicles in the interaction process; by analyzing the vehicle-vehicle interaction behavior under the mixed-driving condition, the unmanned vehicle is helped to understand and judge the behavior of the vehicles with people around;
based on the establishment of a decision model for intersection conflict resolution, the vehicle is controlled by taking acceleration as a decision variable, the intersection conflict resolution problem is converted into a multi-objective optimization problem with constraint, the maximization of the overall driving income of the system is taken as a target, the acceleration is taken as the decision variable, the multi-objective optimization model for intersection vehicle conflict resolution is established, and the optimization of the unmanned vehicle action generation result under different interaction modes of the intersection is completed by introducing a variable cooperation coefficient;
and solving the Pareto optimal solution set by adopting an improved non-inferiority genetic algorithm NSGA-II, then carrying out multi-attribute decision analysis on the Pareto optimal solution set by adopting entropy weight TOPSIS, and selecting the optimal solution according to the preference of a decision maker.
Furthermore, the level of the vehicle crossing interactive relation model and the crossing target vehicle crossing willingness is related to the following two factors:
(1) pressure of conflict P
Due to reasons such as traffic regulations, traffic conflicts can be generated at intersections, overlapping areas can be generated on spatial tracks, and traffic accidents are easy to cause; therefore, the vehicles with conflicts need to adjust the time of passing through the overlapping area, so as to realize conflict resolution; as the driver approaches the conflict point, the sensed conflict pressure increases, and the desire to cross the intersection as soon as possible also increases with increasing pressure, increasing the probability of crossing.
(2) Severity of conflict Tc
Whether the target vehicle passes through or not needs to consider the possibility of collision with the collision vehicle, and T is usedcEvaluating the degree of danger of vehicle collision at the intersection; assuming there are conflicting points or planes between traffic participants, the time difference across the conflicting points or planes is Tc；
In the formula LCSV(t) is the distance (m) of the target vehicle CSV from the conflict point; l isCFV(t) is the distance (m) of the collision vehicle CFV from the conflict point; v. ofCSV(t) is the speed (m/s) of the target vehicle CSV; v. ofCFV(t) is the speed of the conflicting vehicle CFV (m/s);
the greater the collision pressure P, the greater the collision severity TcThe smaller, the more likely the driver is to choose to cross over to get better driving conditions; conversely, the smaller the collision pressure P, the greater the degree of collision TcThe larger the more likely the driver is to cross.
Further, the driving style of the conflicting vehicle:
for vehicles passing through the intersection, the willingness of the vehicles to select to accept or reject the crossing request is different for different driver types, and the fuzzy logic construction method of the conflicting vehicle driving types is similar to the construction method of other vehicle driver types in the lane changing.
Still further, mathematical models of multi-objective optimization problems are often represented in the form:
wherein the independent variable X ═ X1，x2，...，xn)TIs n-dimensional Euclidean space EnThe vector of (1); f (X) is the objective function hi(X) 0 and gj(X) ≧ 0 is a constraint.
For solving the maximum value problem of the revenue function, it can be converted into a minimum value problem of the negative function thereof, so that the mathematical description of the optimization problem is:
s.t.amin≤ai≤amax，i＝1，2
vmin≤vi≤vmax，i＝1，2
in the formula, amin，amaxFor vehicle acceleration limitation, vmin，vmaxIs a vehicle speed limit; p is a cooperative coefficient and takes a value of [ 0-1%]. When the vehicles are in complete competition relationship, p is 0; when the vehicles are in a complete cooperative relationship, p is 1.
For any group (a)1，a2) The speed change scheme of (2) needs to satisfy the following safety constraint conditions:
in the formula
L is the vehicle length and w is the vehicle width.
When the speed of a vehicle to be crossed is adjusted, the calculation formula of the motion state of the vehicle to be crossed is as follows:
in the formula, x1(t)，y1(t) is the position coordinate at the moment t when the vehicle has speed regulation behavior, (x)1(0)，y1(0) Is an initial position, v1(0) At an initial speed, a1(0) In order to be able to accelerate the vehicle,
is the heading angle.
For another vehicle to be crossed, when the speed is adjusted, the calculation formula of the motion state is as follows:
in the formula, x2(t)，y2(t) is the position coordinate at the moment t when the vehicle has speed regulation behavior, (x)2(0)，y2(0) Is an initial position, v2(0) At an initial speed, a2(0) In order to be able to accelerate the vehicle,
is the heading angle.
2) Due to the influence of the intersection lanes and the like, the speeds of the vehicles MV and UV cannot exceed the maximum limit speed at any moment, and the allowed minimum speed is zero (i.e., parking), and the speed constraint conditions are as follows:
0＜vi(t)≤vmax
wherein v ismaxDepending on the maximum speed limit of the road and the intersection.
3) The acceleration of the vehicle in actual traffic is limited by the vehicle performance, and cannot be too large or too small, then for the acceleration constraint:
amin＜ai(t)≤amax
wherein, amin、amaxThe magnitude of (d) depends on vehicle performance and driving comfort requirements;
the improved non-inferiority genetic algorithm (NSGA-II) proposed by Srinivas scholars is adopted, so that the algorithm calculation complexity is reduced; and an elite strategy is adopted to ensure the diversity of the population and improve the convergence and the operation precision of the algorithm.
The invention has the beneficial effects that:
(1) establishing a vehicle lane change interactive relation judgment model and an intersection crossing interactive relation judgment model based on a fuzzy logic method to analyze vehicle-vehicle interactive behaviors under mixed-driving conditions, and helping an unmanned vehicle to understand and judge the behavior of a peripheral vehicle; and several groups of road section and intersection driving data under the mixed running condition in the intelligent automobile competition of China are selected as input parameters of the model, the effectiveness of the vehicle lane changing interactive relationship and intersection crossing interactive relationship judgment model is verified, and the result shows that the accuracy of the fuzzy reasoning method for predicting the cooperative competition degree in the vehicle lane changing and intersection crossing process exceeds 87.3% and 91.6%, the more extreme the vehicle cooperation and competition degree is, the higher the accuracy is, the unmanned vehicle has the capability of understanding human behaviors, and the foundation is laid for realizing multi-vehicle cooperative driving.
(2) According to the characteristic that unmanned driving and manned driving are mixed at the intersection, a multi-objective optimization model for intersection conflict resolution is provided, and when the model is adopted for cooperative driving, the efficiency of conflict vehicles passing through the intersection is effectively improved, so that the running time and the running speed of the conflict vehicles are optimized, and the behavior decision level of the intelligent driving vehicle is improved.
Drawings
FIG. 1 is a diagram of vehicle-vehicle interaction under mixed pedestrian and unmanned vehicle conditions;
FIG. 2 is a schematic view of vehicle crossing interaction relationship determination;
FIG. 3 is a diagram of a predicted result of interaction relationship between crossing vehicles.
FIG. 4 is a schematic diagram of a no-signal intersection conflict;
FIG. 5 is a diagram of the idea of the NSGA-II algorithm;
FIG. 6 is a flow chart of the NSGA-II algorithm;
FIG. 7 is a flow chart of a multi-objective optimization algorithm for intersection conflict resolution;
FIG. 8 is a V model design diagram of a vehicle interactive avoidance system;
FIG. 9 is an example model diagram of intersection vehicle conflict resolution;
fig. 10 is a graph of simulation results (p ═ 1) of group a experimental vehicle MV and UV collision resolution algorithms;
fig. 11 is a graph of simulation results (p is 0) of the MV and UV collision resolution algorithm of the group B experimental vehicles.
Detailed Description
The invention adopts a fuzzy logic method to establish a vehicle crossing interactive relationship judgment model, the model comprehensively considers the crossing intention of two interactive parties and the types of drivers of other vehicles, the cooperation competition degree between the vehicles is subjected to fuzzy reasoning, and the mapping relationship between the output value of the fuzzy reasoning and the interactive relationship prediction result is established to represent the cooperation and competition relationship of other vehicles in the interactive process. By analyzing the vehicle-vehicle interaction behavior under the mixed-driving condition, the unmanned vehicle is helped to understand and judge the behavior of the people around the unmanned vehicle.
The behavior reaction of the opposite side when the two sides conflict for the 1 st time is taken as a judgment criterion, and the Crossing behavior is divided into Cooperative Crossing (COC) and Competitive Crossing (CMC).
Cooperative ride through behavior: that is, vehicle CSV sends a crossing request, vehicle CFV decelerates to avoid and allows it to cross, and then vehicle CFV completes the crossing.
Competitive crossing behavior: the vehicle CSV issues a cross-over request, which is either held at speed or accelerated to reject, forcing the vehicle CSV to make a compromise until the next gap allows it to cross over again.
According to the interaction relation, in order to complete crossing safely, drivers of two parties have a choice of cooperation and competition, and finally, the drivers give a gift to the crossing result of the other party to safely pass through, and complete competition or compromise between the drivers can cause traffic accidents or reduce traffic efficiency, so that the cooperative driving of vehicles at the crossing can ensure safety and improve traffic efficiency.
First, vehicle crossing interactive relation model
And similarly, a vehicle crossing interactive relation judgment model is established according to the crossing intention of the target vehicle at the intersection and the driving types of the conflicting vehicles, and the cooperation degree between the vehicles is deduced by using a fuzzy logic method.
1. Level of target vehicle crossing will
Whether a vehicle can pass through an intersection is mainly related to the collision pressure and the collision severity sensed by the vehicle, namely the intersection collision pressure P and the collision severity TcAnd as the input quantity of the model, the output variable is the crossing willingness degree. The level of the crossing target vehicle's willingness to cross is generally related to two factors:
(1) pressure of conflict P
Generally, due to traffic regulations and other reasons, traffic conflicts may occur at intersections, for example, straight-going and left-turning vehicles may form a cross conflict, and a coincidence area is generated on a spatial trajectory, which is very likely to cause traffic accidents. Therefore, the vehicles that generate the conflict need to adjust the time for passing through the overlapping area, so as to resolve the conflict. As the driver approaches the conflict point, the sensed conflict pressure increases, and the desire to cross the intersection as soon as possible also increases with increasing pressure, increasing the probability of crossing. To better describe the magnitude of the collision pressure, assuming that a vehicle enters the intersection effective communication range at 150m, the collision pressure is set to 0, the pressure to the collision point is set to 1, and the collision pressure expression is:
p-conflict pressure;
Li(t) — the distance (m) of the ith vehicle from the conflict point.
Constructing a membership function of the conflict pressure, setting the domain scope of the conflict pressure P as {0.1, 0.3, 0.5, 0.7, 0.9}, and setting the fuzzy set as { Very Small (VS), small (S), medium (M), large (L) and Very Large (VL) }.
(2) Severity of conflict Tc
Target vehicleWhether to cross or not needs to consider the possibility of collision with the collision vehicle, and T is usedcThe degree of risk of vehicle collision at the intersection is evaluated. Assuming there are conflicting points or planes between traffic participants, the time difference across the conflicting points or planes is Tc。
In the formula LCSV(t) -distance (m) of the target vehicle CSV from the conflict point; l isCFV(t) is the distance (m) of the collision vehicle CFV from the conflict point; v. ofCSV(t) -speed of target vehicle CSV (m/s); v. ofCFV(t) -speed of conflicting vehicle CFV (m/s).
Combining the measured data with the existing research to determine the conflict severity TcThe domain scope of (c) is set as {0, 3, 5, 7, 10}, and the fuzzy set is { Very Large (VL), large (L), medium (M), small (S), Very Small (VS) }.
(3) Fuzzy control rule
From the above analysis, it can be seen that the greater the collision pressure P, the greater the collision severity TcThe smaller, the more likely the driver is to choose to cross over to get better driving conditions; conversely, the smaller the collision pressure P, the greater the degree of collision TcThe larger the vehicle crossing will be, the less likely the driver will cross, and thus the vehicle crossing will fuzzy rule is determined as shown in table 3.5.
TABLE 3.5 pass-through willingness fuzzy logic rule Table
2. Driving type of conflicting vehicle
For vehicles passing through the intersection, the willingness of the vehicles to select to accept or reject the crossing request is different for different driver types, and the fuzzy logic construction method of the conflicting vehicle driving types is similar to the construction method of other vehicle driver types in the lane changing. Fuzzy rules for determining conflicting vehicle driving types based on expert experience and observation are shown in table 3.6.
TABLE 3.6 fuzzy logic rules Table
3. Fuzzy inference
And (4) taking the traversing intention of the target vehicle and the types of drivers of other vehicles obtained in the two steps as input variables, outputting the input variables as the cooperation degree of the vehicle traversing process, and constructing a vehicle traversing interactive relation judgment model. The vehicle cooperation degree values are expressed by language variables of high (H), medium (M) and low (L), and correspond to 3 results of the interaction relationship among other vehicles, namely cooperation relationship, ambiguity and competitive relationship, respectively, and the vehicle passing through the fuzzy logic rule of the interaction relationship is shown in a table 3.7.
TABLE 3.7 fuzzy logic rule Table for vehicle crossing interaction relationships
Similarly, the fuzzy implication relations of the partial fuzzy reasoning adopt the Mamdani rule, the ambiguity resolution adopts the gravity center method, the corresponding relation between the vehicle cooperation competition degree and the traversing intention and the driving type is obtained by calculation,
and performing defuzzification processing on the fuzzy inference output value, and expressing the output value of the vehicle cooperation degree by using p. And p takes a value between 0 and 1, the lower the quantized value of p is, the more intense the competition among the vehicles is, and conversely, the higher the quantized value of p is, the higher the cooperation degree among the vehicles is. Next, a mapping relationship between the fuzzy inference output value and the interactive relationship prediction result is established, as shown in the following formula, where 3 represents a cooperative relationship, 2 represents an ambiguous state, and 1 represents a competitive relationship. Through the prediction of the vehicle interaction relation, the vehicle-vehicle interaction behavior under the mixed-driving condition is analyzed, and the unmanned vehicle is helped to understand and judge the behavior of the people around the unmanned vehicle.
The vehicle crossing interaction model is characterized in that a vehicle crossing interaction relation judgment model is established according to the crossing intention of target vehicles at the crossing and the driving types of conflicting vehicles, and the cooperation degree between the vehicles is deduced by using a fuzzy logic method.
The method comprises the steps of firstly constructing a vehicle intersection crossing interactive relation judgment model, establishing a decision model for intersection conflict resolution based on the model, controlling vehicles by taking acceleration as a decision variable because relevant states and actions of the vehicles in an intersection passing process are continuous values, converting an intersection conflict resolution problem into a constrained multi-objective optimization problem (MOP), maximizing the whole driving income of a system as a target, establishing the intersection vehicle conflict resolution multi-objective optimization model by taking the acceleration as the decision variable, and finishing optimization of unmanned vehicle action generation results under different interaction modes of the intersection by introducing variable cooperation coefficients. The Pareto optimal solution set is solved by adopting an improved non-inferiority genetic algorithm (NSGA-II), then multi-attribute decision analysis is carried out on the Pareto optimal solution set by adopting an entropy weight TOPSIS, the optimal solution is selected according to the preference of a decision maker, the benefits of safety and comfort are considered, the global optimality of an optimization result is ensured, and experimental verification is carried out on a virtual simulation platform.
The vehicle intersection crossing interactive relation judgment model is established based on crossing intention of crossing target vehicles and driving types of conflicting vehicles, and the cooperation degree between the vehicles is inferred by using a fuzzy logic method.
The multi-objective optimization model defines a model objective function, determines constraint conditions, and finally utilizes an improved non-inferiority genetic algorithm (NSGA-II) to solve, so that the algorithm computation complexity is reduced, population diversity is ensured by adopting an elite strategy, and the algorithm convergence and the operation precision are improved.
Mathematical models of multi-objective optimization problems are often represented in the form:
wherein the independent variable X ═ X1，x2，...，xn)TIs n-dimensional Euclidean space EnThe vector of (1); f (X) is the objective function hi(X) 0 and gj(X) ≧ 0 is a constraint.
For solving the maximum value problem of the revenue function, it can be converted into a minimum value problem of the negative function thereof, so that the mathematical description of the optimization problem is:
s.t.amin≤ai≤amax，i＝1，2
vmin≤vi≤vmax，i＝1，2
in the formula, amin、amaxFor vehicle acceleration limitation, vmin、vmaxIs a vehicle speed limit; p is a cooperative coefficient and takes a value of [ 0-1%]. When the vehicles are in complete competition relationship, p is 0; when the vehicles are in a complete cooperative relationship, p is 1.
For any group (a)1，a2) The speed change scheme of (2) needs to satisfy the following safety constraint conditions:
in the formula
L is the vehicle length and w is the vehicle width.
When the speed of a vehicle to be crossed is adjusted, the calculation formula of the motion state of the vehicle to be crossed is as follows:
in the formula, x1(t)y1(t) when there is a speed-governing action in the vehiclePosition coordinate at time t, (x)1(0)，y1(0) Is an initial position, v1(0) At an initial speed, a1(0) In order to be able to accelerate the vehicle,
is the heading angle.
For another vehicle to be crossed, when the speed is adjusted, the calculation formula of the motion state is as follows:
in the formula, x2(t)，y2(t) is the position coordinate at the moment t when the vehicle has speed regulation behavior, (x)2(0)，y2(0) Is an initial position, v2(0) At an initial speed, a2(0) In order to be able to accelerate the vehicle,
is the heading angle.
2) Due to the influence of the intersection lanes and the like, the speeds of the vehicles MV and UV cannot exceed the maximum limit speed at any moment, and the allowed minimum speed is zero (i.e., parking), and the speed constraint conditions are as follows:
0＜vi(t)≤vmax
wherein v ismaxDepending on the maximum speed limit of the road and the intersection.
3) The acceleration of the vehicle in actual traffic is limited by the vehicle performance, and cannot be too large or too small, then for the acceleration constraint:
amin＜ai(t)≤amax
wherein, amin、amaxThe magnitude of (d) depends on vehicle performance and driving comfort requirements;
the improved non-inferiority genetic algorithm (NSGA-II) proposed by Srinivas scholars is adopted, so that the algorithm calculation complexity is reduced; and an elite strategy is adopted to ensure the diversity of the population and improve the convergence and the operation precision of the algorithm.
And (3) simulation experiment verification, namely, a vehicle interaction avoidance decision simulation verification platform under the intersection mixed condition is built by using the Pre Scan software and the Matlab/SimuLink software, the effectiveness of a multi-objective optimization model is verified, a multi-objective optimization conflict resolution algorithm well simulates cooperation and competition modes among vehicles, the robustness of a decision system is improved, and the hybrid driving scene has stronger adaptability.
And the vehicle crossing interactive relation judgment model is verified, effective vehicle motion parameters are obtained by combining video processing of the intelligent automobile competition and GPS data, and intersection driving data under intersection mixed conditions are selected as input parameters of the model and are used for verifying the effectiveness of the intersection vehicle crossing interactive relation judgment model.
Examples
As shown in fig. 1, the driving states of the unmanned vehicle and the manned vehicle are acquired based on the environment sensing module, a vehicle interaction relation determination model is established by using fuzzy reasoning, and the unmanned vehicle decision is made according to the prediction result obtained by the fuzzy reasoning.
As shown in fig. 2, a vehicle crossing interactive relationship determination model is established according to the crossing intention of the target vehicle at the intersection and the driving types of the conflicting vehicles, and the cooperation degree between the vehicles is inferred by using a fuzzy logic method.
As shown in fig. 3, 120 sets of intersection crossing data samples are selected to classify the degree of vehicle cooperation and competition, where 3 is a cooperation relationship indicating the intention of the other vehicle to yield, 2 is an ambiguity indicating an uncertain state of the other vehicle, and 1 is a competition relationship indicating the intention of the other vehicle to not yield. From the figure, it is understood that 38 samples of cooperation relationship, 35 samples of undefined state and 37 samples of competition relationship are correctly identified by the model, and the total number of correct identifications is 110, and the detection accuracy is 91.6%. The accuracy of the fuzzy inference method for predicting the cooperation degree of the vehicle exceeds 90%, which shows that the unmanned vehicle has the capability of understanding human behaviors and lays a foundation for the cooperation of the unmanned vehicle and the manned vehicle.
As shown in FIG. 4, taking a two-lane signalless intersection as an example, an unmanned vehicle may go straight and turn right, straight and left to turn a manned vehicleForming confluence conflict and cross conflict, and the track coincidence area is C in spaceABCDTherefore, traffic accidents are easily caused.
As shown in fig. 5, people and unmanned vehicles having conflicts at the intersection are regarded as a multi-vehicle cooperation system, a conflict vehicle set is obtained through conflict analysis in an internet of vehicles environment, diversity simulation of types of drivers of other vehicles is added, profits of all actions to be selected are calculated, actions corresponding to the maximum profits are output, and cooperation optimization of conflict resolution of people and unmanned vehicles at the intersection is achieved.
As shown in fig. 6, a vehicle interactive avoidance decision simulation verification platform under the intersection mixed-row condition is set up, and the platform relates to four parts of concept design, model design, code generation and system verification in a V cycle.
As shown in fig. 7, for two vehicles having a potential conflict at an intersection, it is first determined whether a conflict exists, and if no conflict exists, a game is not needed, and the vehicle can normally travel. Once the conflict area is entered, the game is started and conflict resolution is achieved through the game.
As shown in fig. 8, the types of drivers are divided into an impulse type (a), a normal type (N), and a conservative type (C), the driving style of the unmanned vehicle is set to 2 types of conservative type and normal conservative type, taking the example that the driving acceleration degree of two vehicles is a conservative-conservative type game, the vehicle MV and UV always perform 4 strategies of acceleration-acceleration, acceleration-deceleration, deceleration-acceleration, and deceleration-deceleration in a game time period, and the variation trend of utility function values of the unmanned vehicle and the manned vehicle is obtained.
As shown in fig. 9, in the whole game process, the unmanned vehicle UV adjusts its behavior strategy according to the behavior of the opposite driver MV, and the behavior decision is greatly different from that in the conflict game process of different types of drivers.
As shown in fig. 10 and 11, the vehicles MV and UV in group a pass through the conflict point at 11s and 14s, respectively, and the vehicles MV and UV in group B pass through the conflict point at 13s and 18s, respectively. Compared with the group B, the group A has shorter average passing time, so that the delay of the intersection is reduced, and the passing efficiency of the intersection is improved.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (4)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911280040.5A CN110992695B (en) | 2019-12-13 | 2019-12-13 | Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911280040.5A CN110992695B (en) | 2019-12-13 | 2019-12-13 | Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110992695A CN110992695A (en) | 2020-04-10 |
| CN110992695B true CN110992695B (en) | 2021-01-29 |
Family
ID=70093303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911280040.5A Active CN110992695B (en) | 2019-12-13 | 2019-12-13 | Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110992695B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111915888B (en) * | 2020-07-14 | 2021-09-03 | 同济大学 | Method for calculating complexity of traffic participants in automatic driving test scene |
| CN112907986B (en) * | 2021-01-12 | 2022-10-21 | 浙江大学 | A dynamic time window intersection scheduling method based on digital twin scene and edge cloud |
| CN113055474B (en) * | 2021-03-12 | 2022-11-11 | 成都格林希尔德交通科技有限公司 | Micro road right transaction system |
| CN113312732B (en) * | 2021-04-28 | 2022-11-15 | 东南大学 | Non-signal control intersection simulation control method and device combining decision advance and dynamic adjustment |
| CN113120003B (en) * | 2021-05-18 | 2022-06-03 | 同济大学 | Unmanned vehicle motion behavior decision method |
| CN113536538B (en) * | 2021-06-11 | 2022-12-13 | 东南大学 | Pedestrian and vehicle interaction simulation method and system based on traffic conflict |
| CN113239244B (en) * | 2021-07-12 | 2021-10-01 | 中国人民解放军国防科技大学 | Multi-uncertain preference acquisition method, device and electronic device based on reliability structure |
| CN113658432B (en) * | 2021-08-17 | 2022-10-14 | 上海交通大学 | Traffic optimization method at intersection without traffic lights based on vehicle traffic priority game |
| CN114387781B (en) * | 2021-12-30 | 2024-04-09 | 北京建筑大学 | Vehicle guidance control method |
| CN114580890B (en) * | 2022-03-02 | 2024-07-26 | 博雷顿科技股份公司 | Collaborative decision analysis neural network algorithm system design method, collaborative decision analysis neural network algorithm system design system and vehicle |
| CN114648877B (en) * | 2022-05-13 | 2022-08-09 | 北京理工大学前沿技术研究院 | Method, system, equipment and storage medium for vehicle passing at signalless intersection |
| CN116153137B (en) * | 2022-12-28 | 2024-05-31 | 清华大学 | Multi-lane ramp collaborative import optimization scheduling method based on vehicle conflict relation graph |
| CN118609420A (en) * | 2024-06-17 | 2024-09-06 | 招商局重庆交通科研设计院有限公司 | A dynamic construction method of intersection topology network for connected collaborative driving |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102350990A (en) * | 2011-06-29 | 2012-02-15 | 北京理工大学 | Comparison model for obstacle avoidance behaviors of vehicle under manned and unmanned conditions |
| KR101469757B1 (en) * | 2013-04-23 | 2014-12-05 | 국방과학연구소 | Method for calculating communication coverage using 3D Object Intersection and Storage medium thereof |
| CN110298122A (en) * | 2019-07-03 | 2019-10-01 | 北京理工大学 | Automatic driving vehicle urban intersection left-hand rotation decision-making technique based on conflict resolution |
| CN110473404A (en) * | 2019-09-19 | 2019-11-19 | 西南交通大学 | A kind of city decussation mouth mixed traffic flow collaboration optimization bottom control method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10319235B2 (en) * | 2017-03-10 | 2019-06-11 | Rovi Guides, Inc. | Systems and methods for resolving conflicts between paths of driverless vehicles based on time remaining in media assets being consumed in the driverless vehicles |
-
2019
- 2019-12-13 CN CN201911280040.5A patent/CN110992695B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102350990A (en) * | 2011-06-29 | 2012-02-15 | 北京理工大学 | Comparison model for obstacle avoidance behaviors of vehicle under manned and unmanned conditions |
| KR101469757B1 (en) * | 2013-04-23 | 2014-12-05 | 국방과학연구소 | Method for calculating communication coverage using 3D Object Intersection and Storage medium thereof |
| CN110298122A (en) * | 2019-07-03 | 2019-10-01 | 北京理工大学 | Automatic driving vehicle urban intersection left-hand rotation decision-making technique based on conflict resolution |
| CN110473404A (en) * | 2019-09-19 | 2019-11-19 | 西南交通大学 | A kind of city decussation mouth mixed traffic flow collaboration optimization bottom control method |
Non-Patent Citations (5)
| Title |
|---|
| Decision making for autonomous driving considering interaction and uncertain prediction of surrounding vehicles;Constantin Hubmann 等;《2017 IEEE Intelligent Vehicles Symposium (IV)》;20170630;第1671-1678页 * |
| 基于NGSIM数据的车辆变道行为博弈模型;张颖达;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20170115(第1期);第36-37页 * |
| 有人与无人驾驶车辆交叉口驾驶博弈模型;成英 等;《北京理工大学学报》;20190930;第39卷(第9期);第938-943页 * |
| 求解约束多目标区间优化的交互多属性决策NSGA_II算法;陈志旺 等;《控制与决策》;20150531;第30卷(第5期);第865-870页 * |
| 道路通行行为研究进展及其展望;石建军 等;《交通信息与安全》;20150430;第33卷(第2期);第1-9页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110992695A (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110992695B (en) | Vehicle urban intersection traffic decision multi-objective optimization method based on conflict resolution | |
| CN111081065B (en) | Intelligent vehicle cooperative lane change decision-making model under mixed traffic conditions | |
| Bai et al. | Hybrid reinforcement learning-based eco-driving strategy for connected and automated vehicles at signalized intersections | |
| Huang et al. | Personalized trajectory planning and control of lane-change maneuvers for autonomous driving | |
| Wang et al. | Risk assessment and mitigation in local path planning for autonomous vehicles with LSTM based predictive model | |
| Ye et al. | Prediction-based eco-approach and departure at signalized intersections with speed forecasting on preceding vehicles | |
| WO2022052406A1 (en) | Automatic driving training method, apparatus and device, and medium | |
| Jayawardana et al. | Learning eco-driving strategies at signalized intersections | |
| CN112233413B (en) | A multi-lane spatiotemporal trajectory optimization method for intelligent networked vehicles | |
| CN111199284A (en) | Vehicle-vehicle interaction model under the condition of mixed driving with manned and unmanned vehicles | |
| CN108106623A (en) | A kind of unmanned vehicle paths planning method based on flow field | |
| Zhang et al. | Structured road-oriented motion planning and tracking framework for active collision avoidance of autonomous vehicles | |
| Chen et al. | A conflict decision model based on game theory for intelligent vehicles at urban unsignalized intersections | |
| CN117227755A (en) | Automatic driving decision method and system based on reinforcement learning under complex traffic scene | |
| CN116540747A (en) | Mobile robot motion planning and obstacle avoidance decision method with priority of avoidance and stop | |
| Ma et al. | Overtaking Path Planning for CAV based on Improved Artificial Potential Field | |
| CN115257820A (en) | A forward collision avoidance driving decision-making method for operating vehicles for open interference scenarios | |
| CN112721948A (en) | Method for realizing lane change scheduling of automatic driving automobile based on prediction and search framework | |
| Zhang et al. | Simulation research on driving behaviour of autonomous vehicles on expressway ramp under the background of vehicle-road coordination | |
| Li et al. | A decision-making approach for complex unsignalized intersection by deep reinforcement learning | |
| CN114779764B (en) | Vehicle reinforcement learning motion planning method based on driving risk analysis | |
| Tan et al. | RCP‐RF: A comprehensive road‐car‐pedestrian risk management framework based on driving risk potential field | |
| CN116909131A (en) | Vehicle formation track planning modeling method for signalless intersection | |
| Guo et al. | Research on integrated decision control algorithm for autonomous vehicles under multi-task hybrid constraints in intelligent transportation scenarios | |
| Wang et al. | Performance analysis of road intersections based on game theory and dynamic level-k model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |