CN115352449A - Intelligent vehicle intersection straight-going speed decision method fusing driving style game - Google Patents
Intelligent vehicle intersection straight-going speed decision method fusing driving style game Download PDFInfo
- Publication number
- CN115352449A CN115352449A CN202210969316.6A CN202210969316A CN115352449A CN 115352449 A CN115352449 A CN 115352449A CN 202210969316 A CN202210969316 A CN 202210969316A CN 115352449 A CN115352449 A CN 115352449A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- game
- driving style
- speed
- intersection
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 230000001133 acceleration Effects 0.000 claims abstract description 58
- 239000011159 matrix material Substances 0.000 claims abstract description 13
- 230000006698 induction Effects 0.000 claims abstract description 7
- 230000007613 environmental effect Effects 0.000 claims abstract description 4
- 230000008901 benefit Effects 0.000 claims description 11
- 239000000446 fuel Substances 0.000 claims description 11
- 230000008859 change Effects 0.000 claims description 10
- 230000004044 response Effects 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 230000003993 interaction Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 206010039203 Road traffic accident Diseases 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000010485 coping Effects 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- 238000005096 rolling process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 208000001613 Gambling Diseases 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/18—Propelling the vehicle
- B60W30/18009—Propelling the vehicle related to particular drive situations
- B60W30/18159—Traversing an intersection
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/08—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to drivers or passengers
- B60W40/09—Driving style or behaviour
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2554/00—Input parameters relating to objects
- B60W2554/40—Dynamic objects, e.g. animals, windblown objects
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W2554/00—Input parameters relating to objects
- B60W2554/40—Dynamic objects, e.g. animals, windblown objects
- B60W2554/404—Characteristics
-
- 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
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention is suitable for the technical field of intelligent automobile driving behavior decision making, and provides an intelligent automobile intersection straight-ahead speed decision making method integrating driving style games, wherein the method is used for acquiring driving state information of surrounding vehicles and identifying the driving styles of the vehicles and the surrounding vehicles; establishing a driving style fused intersection straight-going vehicle non-cooperative dynamic game income matrix; adjusting the decision sequence of participants of each round of stage game in the dynamic game; deciding an optimal acceleration strategy to be adopted by each vehicle in the current round through a reverse induction method; and updating the vehicle state and the environmental information, starting the next round of stage game, and repeatedly executing all the steps until the intelligent vehicles safely pass through the intersection to exit the game. The method effectively eliminates the intersection traffic conflict and makes the right of the road clear on the premise of ensuring the driving safety, so that the intelligent vehicle can safely and efficiently pass through the intersection, the smoothness and the anthropomorphic degree of the intelligent vehicle at the intersection are improved, and the method is more in line with the operation habits of human drivers.
Description
Technical Field
The invention belongs to the technical field of intelligent vehicle driving behavior decision making, and particularly relates to an intelligent vehicle intersection straight-going speed decision making method integrating driving style games.
Background
With the progress of the scientific and technical level, the intelligent degree of automobiles is also continuously improved, and the potential safety hazard of traffic caused by the intelligent degree of automobiles is also more serious. The urban structured road intersection is the most congested and complex scene in road traffic, and is also the scene in which traffic accidents are most likely to happen. It is counted that the average traffic accidents occurring at the intersections all over the world are 30% to 35% of the total number of accidents. In China, the proportion of road traffic accidents occurring at intersections is more than 50%. And the intersection without the signal lamp is lack of the signal lamp to guide, the passing problem of the intelligent vehicle is more complex, a large amount of interference and conflict exist among the intersections, the accident rate is higher than that of the intersection with the signal lamp, and the accident degree is more serious.
Therefore, under the situation of straight-ahead driving at the intersection without the signal lamp, the research on the interactive decision among the automatic driving vehicles is very important to ensure that the intelligent vehicle safely and efficiently passes through the intersection along a set macroscopic path. In the past, the research on the problem has been carried out in a more or less unsatisfactory way. For example, a rule-based decision method needs to model a research object, and abstract a driving rule by a linear representation method, so that more information is ignored in the process, the model is relatively complex, and related constraint conditions are more; the decision method based on the utility theory only considers selecting the maximum profit to decide which traffic strategy to adopt and neglects the influence caused by multi-vehicle interaction in the traffic process; although the decision method based on deep reinforcement learning has strong adaptability, the sampling efficiency is poor, the training period is long, and the reward function design is difficult.
In addition, in the past, only indexes such as driving strategy safety and the like are usually paid attention to in research on automatic driving decisions, and differences of driving styles are rarely considered in research. For different passengers in the intelligent vehicle, a single driving style cannot meet individual requirements of all the passengers. Therefore, the intelligent automobile embodies the driving behavior with the driving style characteristic in the traffic interaction decision of the intersection without the signal lamp, and has important significance for the reality and the diversity of the simulated traffic environment.
Disclosure of Invention
The invention aims to provide a driving style game-fused intelligent vehicle intersection straight-going speed decision method, aiming at solving the problems that a rule-based decision method is complex in model and more in constraint conditions; the decision method based on the utility theory neglects the influence caused by the multi-vehicle interaction behavior; the decision-making method based on deep reinforcement learning is poor in sampling efficiency, long in training period and difficult in reward function design, and the problem of influence of driving style difference is not considered.
The invention is realized in such a way, and provides an intelligent vehicle intersection straight speed decision method integrating a driving style game, which comprises the following steps:
s1, when each vehicle reaches a solid line in front of an intersection, acquiring running state information of surrounding vehicles and identifying the driving styles of the vehicle and the surrounding vehicles;
s2, analyzing and recording the driving style types of the vehicles, and establishing a driving style fused non-cooperative dynamic game income matrix of the vehicles running straight at the intersection;
s3, adjusting the decision sequence of all participants of each round of stage game in the dynamic game in real time according to the TTC (time to live);
s4, solving a sub-game refined Nash equilibrium solution of each round of stage games through a reverse induction method, and deciding an optimal acceleration strategy to be adopted by each vehicle in the current round;
and S5, updating the vehicle state and the environmental information, starting the next round of stage game, and repeatedly executing all the steps until the intelligent vehicles safely pass through the intersection to quit the game.
Preferably, in the step S1, the driving style is identified according to the change rate of the acceleration, and the driving style is divided into an aggressive type, a normal type and a conservative type;
proposing a driving style recognition coefficient S vehicle Is defined as follows:
wherein J (t) is the change rate of the acceleration of the intelligent vehicle in unit time, S J The standard deviation of the acceleration change rate in the driving style recognition period,identifying the average value of the acceleration change rate in the period, wherein v (t) is the running speed of the intelligent vehicle;
the specific division standard is as follows: s vehicle The driving style of the vehicle is aggressive when the driving style is more than or equal to 1, and S is more than 0.5 vehicle When the vehicle driving style is less than 1, S vehicle The driving style of the vehicle is conservative when the driving style is less than or equal to 0.5.
Preferably, the vehicle non-cooperative dynamic game income matrix established in the step S2 includes: safety gain P considering driving style safe Driving efficiency gain P considering driving style eff Comfort benefit P considering passenger experience com And economic profit P taking into account fuel consumption eco ;
The time for the head of the vehicle i to reach the collision area is as follows:
wherein,the speed of vehicle i at the end of the previous stage game or the speed of vehicle i at the beginning of the current stage game,acceleration strategy that may be taken from the acceleration candidate set for the gaming vehicle i at the current stage, d i The distance from the head of the current vehicle i to the near end edge of the collision region;
the time for the tail of the vehicle i to just drive away from the collision area is as follows:
wherein l i Is the length of vehicle i, W is the lane width;
the absolute safe time difference between the vehicles i and j is:
Δt ij =t i -τ j or Δ t ij =t j -τ i ;
Preferably, the Δ t ij And if the absolute safety time difference is greater than 0, the acceleration strategy pair is used for centralizing the acceleration candidates according to the safety threshold and enabling the absolute safety time difference to be negative.
The safety gains considering driving style are defined as:
wherein, Δ t max Is the upper limit value of the absolute safety time difference.
The driving efficiency yield considering the driving style is defined as:
wherein 0 < theta<1,v max For the maximum speed that is limited at the intersection,speed, v, corresponding to the acceleration strategy that the gaming vehicle i may adopt at the current stage ibest The optimal speed corresponding to the driving style of the vehicle i is obtained;
the vehicle speed gain increases rapidly at lower vehicle speeds and saturates exponentially at high speeds.
The comfort benefit considering the passenger experience is defined as:
wherein,for the accelerations that may be taken in the current stage of the game,optimum acceleration determined for the game of the previous stage, a max 、a min Defined maximum acceleration and maximum deceleration.
The fuel consumption rate is represented by a speed function:
wherein,for the average speed of the gaming vehicle at the current stage,for the current phaseThe speed of a vehicle i when playing is started, T is a stage game playing period, and a, e, f and g are constant coefficients;
the economic benefit considering fuel consumption is defined as:
wherein, E min 、E max A reference lower and upper bound representing an economic indicator, respectively.
The total revenue matrix is constructed from four revenue indexes and their corresponding weight coefficients:
σ 1 +σ 2 +σ 3 +σ 4 =1
wherein σ 1 、σ 2 、σ 3 、σ 4 Weight coefficients respectively representing safety requirements, efficiency requirements, comfort requirements and economic requirements of vehicles, wherein the sum of the four weights is equal to 1, and constraints such as maximum acceleration and deceleration and maximum speed of vehicles when crossing passes are set, a max And a min Respectively representing the maximum acceleration and the maximum deceleration of the vehicle, the minimum speed of the vehicle is 0 max Indicating the road speed limit condition.
Preferably, in the step S3, the decision sequence of all participants in each round of the stage game in the dynamic game is adjusted in real time according to the possible time to collision TTC;
the possible collision time of two vehicles with a collision is expressed as: the time difference between the two vehicles which travel at the speed and the acceleration at the current moment to reach the intersection conflict area is defined by the following specific formula:
the optimal acceleration determined by the game of the previous stage and the speed of the vehicles at the end of the game of the previous stage is respectively taken as the speed of the vehicles at the end of the game of the previous stage i and j, and the possible collision time of the four vehicles at the intersection is T AB ,T BC ,T CD ,T DA 。
Determining the sequential rule as: first find the minimum of the four possible collision times, min { T } AB ,T BC ,T CD ,T DA E.g. if T AB Minimum, the priority decision maker will consider between the A car and the B car; based on this, further comparisons of T associated with cars A and B DA And T BC Of size, if T DA >T BC If the vehicle B is the first decision maker, the vehicle A is the second decision maker, the vehicle C is the third decision maker, and D
Preferably, the nash equalization solution in step S4 is defined as:
in game strategy G = { A = } 1 ,A 2 …A n ;U 1 ,U 2 …U n In the game, there are n game participants, A 1 ,A 2 …A n Representing the participant 1,2, \8230N, a set of behavior strategies, U 1 ,U 2 …U n For the behavior profit of each participant, any one strategy of each game participant forms a strategy combination (a) 1 *,a 2 *…a n * ) If for any gambling party i, strategy a i * Are all given other participant policies (a) 1 *,a 2 *…a i-1 *,a i+1 *… a n-1 *a n * ) The optimal coping strategy for the case i is as follows: u shape i (a 1 *,a 2 *…a i-1 *,a i+1 *…a n *)≥U i (a 1 *,a 2 *…a i-1 *,a k *,a i+1 *…a n-1 *a n *);
For any of a k ∈A i All the above are true, the policy combination is called (a) 1 *,a 2 *…a n * ) A Nash equilibrium solution for game G;
according to the method, a reverse induction method is used for solving a refined Nash equilibrium solution of each round of sub-game of each stage game, the last stage or the last sub-game at the tail end of a dynamic game tree is started, and actions serving as disadvantaged strategies in each optional strategy branch are deleted forward in a reverse-pushing mode according to the income value; realizing Nash balance in each sub game layer by layer;
and updating the state parameter information of the vehicle and the surrounding vehicles and the road environment information when the next round of game is started so as to perform a new round of rolling game interaction decision, and repeatedly executing the steps I to the step IV until all the vehicles safely and smoothly pass through the intersection to exit the game and run at the constant speed at the exit moment.
The intelligent vehicle intersection straight-going speed decision method fusing the driving style game has the following beneficial effects:
1. aiming at a signal lamp-free road intersection with complex traffic conditions, the dynamic property and uncertainty of other vehicles can be fully considered through game interaction on the premise of not identifying the driving intentions of other vehicles, and the safe and effective speed decision of the intelligent vehicle during driving along the established path is realized.
2. The requirements of the automatic driving decision-making individuation are considered, the driving style is integrated into the game decision-making algorithm, so that the driving income is not only reasonably and optimally selected, but also is sensitively judged according to the driving habit style of the driver, the intersection traffic conflict is effectively eliminated and the right of way is defined on the premise of ensuring the driving safety, the requirements of various vehicles on high efficiency, comfort and the like can be met, and the smoothness and anthropomorphic degree of intelligent vehicle decision-making at the intersection are improved.
3. The influence of the game sequence of the dynamic game participants on the solution is considered, the action sequence of the participants is adjusted according to the possible collision time TTC, and the sequence of the response of human drivers to the collision urgency degree when encountering intersection collision is met, so that the profit value of the decision made by each vehicle in each round of game is more real and reliable, and the actual road traffic scene is more matched.
Drawings
Fig. 1 is a flow chart of a decision on a straight-moving speed at an intersection of an intelligent vehicle with a game of a driving style integrated according to an embodiment of the present invention;
fig. 2 is a structural schematic diagram of the decision on the straight-ahead speed of the intersection of the intelligent vehicle fusing the driving style game provided by the embodiment of the invention;
FIG. 3 is a schematic diagram of a situation where two vehicles pass through an intersection safely without collision according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of another situation in which two vehicles may pass through an intersection safely without collision according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a decision sequence for determining each vehicle by the time to collision TTC according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a dynamic gaming tree provided by an embodiment of the present invention;
FIG. 7 is a schematic diagram of a solution of inverse induction method according to an embodiment of the present invention;
fig. 8 is a schematic diagram of interaction and conflict areas among intelligent vehicles in a straight lane of a signal-free road intersection according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
Specific implementations of the present invention are described in detail below with reference to specific embodiments.
As shown in fig. 1, an embodiment of the invention provides a driving style game-fused intelligent vehicle intersection straight-going speed decision method, which is used for performing optimal decision making on the premise that indexes such as safety are met when an intelligent vehicle runs straight at a signal-lamp-free intersection.
As shown in fig. 2, the sensing system identifies the corresponding driving style according to the acquired information by acquiring the state information of the vehicle and the surrounding vehicles, and then takes the acquired vehicle information as the input of the non-cooperative dynamic game decision-making system; and selecting and outputting the optimal acceleration strategy of each vehicle through model calculation and a maximum profit principle, updating the relevant information of the vehicle and the vehicles at the next moment, repeatedly and circularly executing the operations, and automatically quitting the game and recovering the normal driving mode when the vehicle safely passes through all conflict areas of the intersection.
The method specifically comprises the following working steps:
s1, when each vehicle reaches a solid line in front of an intersection, acquiring running state information of surrounding vehicles and identifying the driving styles of the vehicle and the surrounding vehicles;
the method comprises the steps that when a vehicle reaches the starting point of a solid line in front of an intersection, the vehicle starts to enter a game, a vehicle sensing system can acquire running state information of surrounding vehicles according to a V2V system or a vehicle-mounted camera, a millimeter wave radar, a speed and distance measuring sensor and the like, and simultaneously transmit the running state information of the vehicle to a decision system, wherein the parameter information comprises a transverse position and a longitudinal position relative to a geodetic coordinate system, a vehicle longitudinal speed, a vehicle longitudinal acceleration, a distance from a collision area, a surrounding driving environment and the like;
in addition, for the acquired longitudinal acceleration information, driving styles are identified through the acceleration change rate, and the driving style types are divided into an aggressive type, a normal type and a conservative type;
proposing a driving style recognition coefficient S vehicle Is defined as follows:
wherein J (t) is the change rate of the acceleration of the intelligent vehicle in unit time, S J Identifying the standard deviation of the acceleration rate in the period for the driving style, J is the average value of the acceleration rate in the period, and v (t) is the running speed of the intelligent vehicle;
the specific division standard is as follows: s vehicle The driving style of the vehicle is aggressive when the driving style is more than or equal to 1, and S is more than 0.5 vehicle When the driving style of the vehicle is less than 1, S vehicle The driving style of the vehicle is conservative when the driving style is less than or equal to 0.5.
S2, analyzing and recording the driving style types of the vehicles, and establishing a driving style fused intersection direct-driving vehicle non-cooperative dynamic game income matrix;
the invention combines the driving style of the intelligent vehicle with the dynamic game, establishes a dynamic game income matrix fusing the driving style, and comprises the following steps: considering safety gains of the driving style, driving efficiency gains of the driving style, comfort gains of passengers and economic gains of fuel consumption, wherein the total gain matrix is constructed by four gain indexes and corresponding weight coefficients thereof;
the safety benefit considering driving style represents: the method is used for measuring the reliability of whether two vehicles possibly having collision risk or not have collision risk, and simulating the lateral weight between different driving styles and the difference between the road weight and the absolute safety time. The aggressive style is more likely to acquire the right of way when facing potential conflict, and the conservative style is more likely to give out the right of way, so that the priority of the two styles for treating the right of way is higher than the absolute safety time difference; the normal style does not look at the weight of the road, so only the magnitude of the absolute safety time difference is noted.
Respectively extending the road side lines of lanes where two vehicles with potential conflicts are located to an intersection, wherein a square area formed by intersection is a conflict area;
in order to ensure the safety of two vehicles passing through the intersection, the sufficient condition is that the tail of the vehicle firstly passes through the conflict area completely, and then the head of the vehicle just reaches the conflict area, namely as shown in fig. 3 and 4, two conditions that the two vehicles pass through the intersection safely without collision are shown;
the time for the head of the vehicle i to reach the collision area is as follows:
wherein,the speed of vehicle i at the end of the previous stage game or the speed of vehicle i at the start of the current stage game,acceleration strategy that the gaming vehicle i may take from the acceleration candidate set for the current stage, d i The distance from the head of the current vehicle i to the near end edge of the conflict area is calculated;
the time for the tail of the vehicle i to just drive away from the collision area is as follows:
wherein l i Is the length of vehicle i, and W is the lane width;
the absolute safe time difference between vehicles i and j is:
Δt ij =t i -τ j or Δ t ij =t j -τ i ;
Preferably, the Δ t ij And if the absolute safety time difference is greater than 0, the acceleration strategy pair is used for centralizing the acceleration candidates according to the safety threshold and enabling the absolute safety time difference to be negative.
The safety gains considering driving style are defined as:
wherein, Δ t max Is the upper limit value of the absolute safety time difference.
The driving efficiency gain considering the driving style is defined as:
wherein 0 < theta < 1,v max The maximum speed that is limited at the intersection,speed, v, corresponding to the acceleration strategy that the gaming vehicle i may adopt at the current stage ibest The optimal speed corresponding to the driving style of the vehicle i is obtained;
the vehicle speed gain increases rapidly at lower vehicle speeds and saturates exponentially at high speeds.
The comfort benefit considering the passenger experience is defined as:
wherein,for accelerations that may be taken in the current stage game,optimum acceleration determined for the game of the previous stage, a max 、a min As defined maximum acceleration and maximum deceleration;
the economic benefit considering fuel consumption represents: the influence of the corresponding speed which is possibly adopted by the acceleration strategy in the current stage game on the fuel consumption is shown as the benefit brought by energy conservation from the economic cost perspective;
specific fuel consumption may be expressed as a function of speed:
wherein,for the average speed of the gaming vehicle at the current stage,the speed of a vehicle i when the current stage game starts, T is the stage game period, a, e, f and g are constant coefficients, and a = -0.67944, e = -0.029665, f = -0.00028 and g = -0.00000149 are taken;
the economic gain in consideration of fuel consumption is defined as:
wherein E is min 、E max A reference lower and upper bound representing an economic indicator, respectively.
The total revenue matrix is constructed from four revenue indexes and their corresponding weight coefficients:
σ 1 +σ 2 +σ 3 +σ 4 =1
wherein σ 1 、σ 2 、σ 3 、σ 4 Weight coefficients respectively representing safety requirements, efficiency requirements, comfort requirements and economic requirements of the vehicles, wherein the sum of the four weights is equal to 1, and meanwhile, the maximum acceleration and deceleration, the maximum speed and other constraints of the vehicles when the vehicles pass through the intersection are set, a max And a min Respectively representing the maximum acceleration and the maximum deceleration of the vehicle, the minimum speed of the vehicle is 0 max Indicating the road speed limit condition.
S3, adjusting the decision sequence of all participants of each round of stage game in the dynamic game in real time according to the TTC (time to live);
in the step S3, the decision sequence of all participants of each stage of game in the dynamic game is adjusted in real time according to the TTC (time to live) of the possible collision time;
because the sequence of the dynamic game participants has great influence on the optimal solution, the invention adjusts the decision sequence of all participants of each phase of game in the dynamic game in real time according to the TTC (time to collision), thereby ensuring the ordered progress of the whole game decision;
the possible collision time of two vehicles with a collision is expressed as: the time difference that two vehicles travel at the speed and the acceleration of the current moment to reach the intersection conflict area is defined as follows:
for one-stage game of i and j of vehicles respectivelyThe speed at the end and the optimal acceleration determined by the game in the last stage are used for driving, the time when the head reaches a collision area and the possible collision time of four vehicles at the intersection are sequentially T AB ,T BC ,T CD ,T DA 。
Determining the sequential rule as: first find the minimum of the four possible collision times, min { T } AB ,T BC ,T CD ,T DA E.g. if T AB Minimum, the priority decision maker will consider between vehicle A and vehicle B; further comparing T related to the A vehicle and the B vehicle DA And T BC Size of (c), if T DA >T BC If the vehicle B is the first decision maker, the vehicle a is the second decision maker, the vehicle C is the third decision maker, the vehicle D is the fourth decision maker, and the overall decision sequence is as follows: b → A → C → D; if T DA <T BC If the vehicle a is the first decision maker, the vehicle B is the second decision maker, the vehicle D is the third decision maker, the vehicle C is the fourth decision maker, and the overall decision sequence is as follows: a → B → D → C, other different order cases can be determined by this rule.
As shown in fig. 5, the decision sequence of each vehicle is determined by the possible time to collision TTC, which is 8 cases, and all the cases can be determined by the rule;
s4, solving a sub-game refined Nash equilibrium solution of each round of stage game through a reverse induction method, and deciding an optimal acceleration strategy to be adopted by each car in the current round;
establishing an acceleration candidate set of each vehicle: setting a max =4m/s,a min = 4m/s, aggregate precision Δ a =0.4;
as shown in fig. 6, the earnings of the vehicles corresponding to each group of acceleration strategy combinations in the candidate set are calculated through the earning matrix, and a dynamic game tree is generated according to the sequence of the vehicles;
in game strategy G = { A = } 1 ,A 2 …A n ;U 1 ,U 2 …U n In the } there are n game participants, A 1 ,A 2 …A n Representing the participant 1,2, \8230N, a set of behavior strategies, U 1 ,U 2 …U n For the behavioral benefit of each participant, each beatAny one strategy of the chess participants forms a strategy combination (a) 1 *,a 2 *…a n * ) If for any gambler i, policy a i * Are all given other participant policies (a) 1 *,a 2 *…a i-1 *,a i+1 *… a n-1 *a n * ) The optimal coping strategy of the case i is as follows: u shape i (a 1 *,a 2 *…a i-1 *,a i+1 *…a n *)≥U i (a 1 *,a 2 *… a i-1 *,a k *,a i+1 *…a n-1 *a n *);
For any of a k ∈A i All the above are true, the policy combination is called (a) 1 *,a 2 *…a n * ) A Nash equilibrium solution for game G;
the method comprises the following steps that firstly, a participant acting firstly in the dynamic game must consider the strategy selection of a later action participant in the later stage when the strategy is selected in the earlier stage, only the participant in the last stage can directly make a selection without the restriction of other participants, and after the selection of the later stage participant is determined, the strategy of the previous stage participant is also determined;
as shown in fig. 7, the method solves the refined nash equilibrium solution of the sub-game of each round of stage game by the inverse induction method, starts from the last stage or the last sub-game at the tail end of the dynamic game tree, and sequentially pushes backwards and deletes the action as the disadvantaged strategy in each selectable strategy branch in each branch according to the principle of selecting the maximum income of the participants from back to front; realizing Nash equilibrium in each sub game layer by layer;
the sub-game refining Nash equilibrium solution eliminates Nash equilibrium containing the incredible threats in the Nash equilibrium, so that the equilibrium strategy does not contain the incredible threats any more, and the optimal acceleration strategy to be adopted by each vehicle in the round is decided under the condition of considering the actual traffic environment;
and S5, updating the vehicle state and the environmental information, starting a next round of stage game, and repeatedly executing all the steps until the intelligent vehicles safely pass through the intersection to quit the game.
And updating the state parameter information of the vehicle and the surrounding vehicles and the road environment information when the next round of game starts to perform a new round of rolling game interaction decision, and repeatedly executing the first step to the fourth step until all the vehicles safely and smoothly pass through the intersection to exit the game and run at the constant speed of the exit moment.
Each intelligent vehicle passes through the two conflict areas, so that the condition that the vehicle exits the game is set to exit the game when the vehicle completely passes through the second conflict area, namely the vehicle can be considered to have no influence on the driving safety of other intelligent vehicles;
the applicable scene in the embodiment of the invention is shown in fig. 8 and limited by road traffic marks, the current intersection only allows the vehicle to go straight and does not allow the vehicle to turn left and right, and the middle lane is kept to run when the vehicle is running straight at the intersection;
conflict areas exist among the future driving tracks of four vehicles A, B, C and D in the straight lane of the intersection;
each vehicle needs to pass through two collision areas, namely, two vehicles have potential collision influence with the vehicle;
specially, the vehicle A and the vehicle C have no conflict relationship, and the vehicle B and the vehicle D have no conflict relationship;
The invention provides a driving style game-fused intelligent vehicle intersection straight-going speed decision method, which is used for solving the passing problem of an intelligent vehicle in a complex no-signal-lamp intersection open scene. The driving style of the vehicle is considered in the dynamic game process, the game income matrix design based on the driving style is carried out, and the dynamic game is solved to obtain the optimal behavior strategy of each vehicle, so that the intelligent vehicle can safely and smoothly pass through a road junction, and the operation habit of a human driver is better met.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (9)
1. A driving style game fused intelligent vehicle intersection straight-going speed decision method is characterized by comprising the following steps:
s1, when each vehicle reaches a solid line in front of an intersection, acquiring running state information of surrounding vehicles and identifying the vehicle and the driving style of the surrounding vehicles;
s2, analyzing and recording the driving style types of the vehicles, and establishing a driving style fused intersection straight-going vehicle non-cooperative dynamic game profit matrix;
s3, adjusting the decision sequence of all participants of each round of stage game in the dynamic game in real time according to the TTC (time to live);
s4, solving a sub-game refined Nash equilibrium solution of each round of stage games through a reverse induction method, and deciding an optimal acceleration strategy to be adopted by each vehicle in the current round;
and S5, updating the vehicle state and the environmental information, starting the next round of stage game, and repeatedly executing all the steps until each intelligent vehicle safely passes through the intersection to quit the game.
2. The intelligent vehicle intersection straight-going speed decision method fused with the driving style game is characterized in that in the step S1, the driving style is identified through the change rate of the acceleration, and the driving style is divided into an aggressive type, a normal type and a conservative type;
proposing a driving style identification coefficient S vehicle Is defined as follows:
wherein J (t) is the change rate of the acceleration of the intelligent vehicle in unit time, S J The standard deviation of the acceleration change rate in the driving style recognition period,identifying the average value of the acceleration change rate in the period, wherein v (t) is the running speed of the intelligent vehicle;
S vehicle the driving style of the vehicle is aggressive when the driving style is more than or equal to 1, and S is more than 0.5 vehicle When the driving style of the vehicle is less than 1, S vehicle The driving style of the vehicle is conservative when the driving style is less than or equal to 0.5.
3. The intelligent vehicle intersection straight-going speed decision method fusing driving style game according to claim 1, characterized in that the vehicle non-cooperative dynamic game income matrix established in the step S2 comprises: safety gain P considering driving style safe Driving efficiency gain P considering driving style eff Comfort benefit P considering passenger experience com And economic profit P taking into account fuel consumption eco ;
The time for the head of the vehicle i to reach the collision area is as follows:
wherein,the speed of vehicle i at the end of the previous stage game or the speed of vehicle i at the start of the current stage game,possible slave acceleration of the gaming vehicle i for the current stageAcceleration strategy taken in the selection, d i The distance from the head of the current vehicle i to the near end edge of the conflict area is calculated;
the time for the tail of the vehicle i to just drive away from the collision area is as follows:
wherein l i Is the length of vehicle i, and W is the lane width;
the absolute safe time difference between vehicles i and j is:
Δt ij =t i -τ j or Δ t ij =t j -τ i ;
The total revenue matrix is constructed from four revenue indexes and their corresponding weight coefficients:
P i n =σ 1 P safe +σ 2 P eff +σ 3 P com +σ 4 P eco
σ 1 +σ 2 +σ 3 +σ 4 =1
wherein σ 1 、σ 2 、σ 3 、σ 4 Respectively representing the weight coefficients of the vehicle for the safety requirement, the efficiency requirement, the comfort requirement and the economic requirement, wherein the sum of the four weights is equal to 1, and meanwhile, the constraints of the maximum acceleration and deceleration, the maximum speed and the like of the vehicle during crossing traffic are set, a max And a min Respectively representing the maximum acceleration and the maximum deceleration of the vehicle, the minimum speed of the vehicle being 0,v max Indicating the road speed limit condition.
4. The intelligent vehicle intersection straight-going speed decision-making method fused with driving style game as claimed in claim 3, wherein the delta t is ij And if the absolute safety time difference is greater than 0, the acceleration candidate is concentrated according to the safety threshold, and the acceleration strategy pair with the negative absolute safety time difference is eliminated.
5. The decision-making method for the intersection straight-going speed of the intelligent vehicle fusing the driving style game as claimed in claim 3, wherein the safety yield considering the driving style is defined as:
wherein, Δ t max Is the upper limit value of the absolute safety time difference.
6. The decision-making method for the intersection straight-going speed of the intelligent vehicle fusing the driving style game as claimed in claim 3, wherein the driving efficiency and yield considering the driving style are defined as follows:
7. The decision-making method for the intersection straight-going speed of the intelligent vehicle fusing the driving style game as claimed in claim 6, wherein the comfort benefit considering the passenger experience is defined as:
8. The decision method for the intersection straight-going speed of the intelligent vehicle fusing the driving style game as claimed in claim 6, wherein the fuel consumption rate is expressed by a speed function:
wherein,for the average speed of the gaming vehicle at the current stage,the speed of a vehicle i when the current stage game starts, T is the period of the stage game, and a, e, f and g are constant coefficients;
the economic gain considering fuel consumption is defined as:
wherein E is min 、E max A reference lower and upper bound representing an economic indicator, respectively.
9. The intelligent vehicle intersection straight-going speed decision method fused with the driving style game is characterized in that in the step S3, the decision sequence of all participants of each stage of the dynamic game is adjusted in real time according to the possible collision time TTC;
the possible collision time of two vehicles with a collision is expressed as: the time difference between the two vehicles running at the speed and the acceleration of the current moment to reach the intersection conflict area is defined as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210969316.6A CN115352449A (en) | 2022-08-12 | 2022-08-12 | Intelligent vehicle intersection straight-going speed decision method fusing driving style game |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210969316.6A CN115352449A (en) | 2022-08-12 | 2022-08-12 | Intelligent vehicle intersection straight-going speed decision method fusing driving style game |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115352449A true CN115352449A (en) | 2022-11-18 |
Family
ID=84033042
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210969316.6A Pending CN115352449A (en) | 2022-08-12 | 2022-08-12 | Intelligent vehicle intersection straight-going speed decision method fusing driving style game |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115352449A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116168550A (en) * | 2022-12-30 | 2023-05-26 | 福州大学 | Traffic coordination method for intelligent network-connected vehicles at signalless intersections |
CN117973660A (en) * | 2024-03-29 | 2024-05-03 | 华东交通大学 | Multi-vehicle dynamic path selection method and system |
-
2022
- 2022-08-12 CN CN202210969316.6A patent/CN115352449A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116168550A (en) * | 2022-12-30 | 2023-05-26 | 福州大学 | Traffic coordination method for intelligent network-connected vehicles at signalless intersections |
CN117973660A (en) * | 2024-03-29 | 2024-05-03 | 华东交通大学 | Multi-vehicle dynamic path selection method and system |
CN117973660B (en) * | 2024-03-29 | 2024-06-11 | 华东交通大学 | Multi-vehicle dynamic path selection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108595823B (en) | Autonomous main vehicle lane changing strategy calculation method combining driving style and game theory | |
CN115352449A (en) | Intelligent vehicle intersection straight-going speed decision method fusing driving style game | |
CN106114507B (en) | Local path planning method and device for intelligent vehicle | |
CN113291308B (en) | Vehicle self-learning lane-changing decision-making system and method considering driving behavior characteristics | |
CN110304074B (en) | Hybrid driving method based on layered state machine | |
CN113276884B (en) | Intelligent vehicle interactive decision passing method and system with variable game mode | |
JP5375805B2 (en) | Driving support system and driving support management center | |
CN113253739B (en) | Driving behavior decision method for expressway | |
CN113044064A (en) | Vehicle self-adaptive automatic driving decision method and system based on meta reinforcement learning | |
CN113312752B (en) | Traffic simulation method and device for main road priority control intersection | |
CN110363986A (en) | A kind of centralized merging area car speed optimization method based on the game of vehicle vehicle and driving potential field power | |
CN114852105A (en) | Method and system for planning track change of automatic driving vehicle | |
CN113222678A (en) | Point reward method for encouraging vehicles to change lanes cooperatively in internet environment | |
CN116740945A (en) | Method and system for multi-vehicle collaborative grouping intersection of expressway confluence region in mixed running environment | |
CN114802306A (en) | Intelligent vehicle integrated decision-making system based on man-machine co-driving concept | |
CN117227755A (en) | Automatic driving decision method and system based on reinforcement learning under complex traffic scene | |
CN112542061B (en) | Lane borrowing and overtaking control method, device and system based on Internet of vehicles and storage medium | |
WO2023004698A1 (en) | Method for intelligent driving decision-making, vehicle movement control method, apparatus, and vehicle | |
CN113743767B (en) | Vehicle dispatching method, system, computer and medium based on time and safety | |
CN112590791B (en) | Intelligent vehicle lane change gap selection method and device based on game theory | |
CN113837211B (en) | Driving decision method and device | |
CN114492157A (en) | Automatic driving test scene generation method based on personalized driver model | |
CN117275240B (en) | Traffic signal reinforcement learning control method and device considering multiple types of driving styles | |
CN117207961A (en) | Automatic driving lane keeping method based on Swim-TD3 | |
CN118172953B (en) | Dynamic control method and system for traffic in expressway junction area under intelligent networking background |
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 |