A straight-through speed decision-making method for intelligent vehicles at intersections based on driving style game

Abstract

This invention is applicable to the field of intelligent vehicle driving behavior decision-making technology. It provides a method for intelligent vehicle through-the-road speed decision-making at intersections that integrates a driving style game. The method collects driving status information from surrounding vehicles and identifies the driving styles of the vehicle itself and surrounding vehicles. It then establishes a payoff matrix for a non-cooperative dynamic game involving vehicles traveling through intersections that incorporates driving styles. The decision-making sequence of participants in each round of the dynamic game is adjusted. The optimal acceleration strategy for each vehicle in that round is determined through a backward induction method. The method then updates vehicle status and environmental information and initiates the next round of the game, repeating all of the above steps until each intelligent vehicle safely passes through the intersection and exits the game. This method effectively eliminates intersection traffic conflicts and clarifies right-of-way while ensuring driving safety, enabling intelligent vehicles to safely and efficiently pass through intersections. It also improves the fluidity and humanlike nature of intelligent vehicle decision-making at intersections, making it more consistent with the operating habits of human drivers.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent vehicle driving behavior decision-making, and in particular relates to an intelligent vehicle intersection straight-through speed decision-making method integrating driving style game. Background Art

[0002] With the advancement of science and technology, the level of vehicle intelligence is constantly increasing, and the resulting traffic safety risks are becoming increasingly serious. Urban intersections are the most congested and complex traffic scenes, and are also the most prone to traffic accidents. According to statistics, globally, 30% to 35% of all traffic accidents occur at intersections. In China, the proportion of road traffic accidents at intersections is even higher than 50%. Unsignaled intersections, lacking the guidance of traffic lights, further complicate the passage of intelligent vehicles, resulting in numerous interferences and conflicts. Consequently, the accident rate is higher and the severity of accidents is more severe than at intersections with traffic lights.

[0003] Therefore, in scenarios involving straight-ahead driving at intersections without traffic lights, studying the interactive decision-making between autonomous vehicles is crucial in order to ensure that intelligent vehicles can safely and efficiently pass through intersections along a predetermined macroscopic path. Previous research on this issue has employed methods that have been somewhat unsatisfactory. For example, rule-based decision-making methods require modeling the research object and abstracting driving rules using linear representations. This process ignores a large amount of information, resulting in relatively complex models and numerous constraints. Decision-making methods based on utility theory only consider the maximum benefit to determine the appropriate traffic strategy, ignoring the impact of multi-vehicle interactions during the traffic process. Decision-making methods based on deep reinforcement learning, while highly adaptable, suffer from poor sampling efficiency, long training cycles, and difficulty designing reward functions.

[0004] Furthermore, previous research on autonomous driving decision-making has typically focused solely on indicators such as driving strategy safety, with few considering differences in driving style. However, a single driving style cannot meet the individual needs of different passengers within a smart car. Therefore, enabling smart cars to reflect driving behaviors characterized by driving styles during interactive decision-making at unsignaled intersections is crucial for enhancing the realism and diversity of simulated traffic environments. Summary of the Invention

[0005] The purpose of the present invention is to provide a smart car intersection straight-through speed decision-making method that integrates driving style game, aiming to solve the problems that rule-based decision-making methods have complex models and many constraints; decision-making methods based on utility theory ignore the impact of multi-vehicle interaction behavior; decision-making methods based on deep reinforcement learning have poor sampling efficiency, long training cycles, and difficult reward function design, and all fail to consider the impact of driving style differences.

[0006] The present invention is implemented as follows: a method for determining the straight-through speed of an intelligent vehicle at an intersection by integrating a driving style game, the method comprising the following steps:

[0007] S1. When each vehicle reaches the solid line before the intersection, it collects driving status information of surrounding vehicles and identifies the driving style of itself and surrounding vehicles;

[0008] S2. Analyze and record vehicle driving style types and establish a non-cooperative dynamic game payoff matrix for straight-going vehicles at intersections that integrates driving styles;

[0009] S3. Adjust the decision-making order of all participants in each round of the dynamic game in real time based on the possible collision time TTC;

[0010] S4. Solve the subgame-perfect Nash equilibrium solution of each round of the game through backward induction, and decide the optimal acceleration strategy for each car in this round;

[0011] S5. Update the vehicle status and environment information and start the next round of the game. Repeat all the above steps until all smart cars safely pass through the intersection and exit the game.

[0012] Preferably, in step S1, the driving style is identified by the rate of change of acceleration and is divided into aggressive, normal and conservative types;

[0013] The driving style recognition coefficient S is proposed vehicle Defined as:

[0014]

[0015]

[0016] Among them, J(t) is the rate of change of the acceleration of the smart car per unit time, S J is the standard deviation of the acceleration change rate during the driving style identification cycle, is the average value of the acceleration change rate during the identification period, v(t) is the speed of the intelligent vehicle;

[0017] The specific classification criteria are: Svehicle When ≥1, the vehicle's driving style is aggressive, 0.5<S vehicle When <1, the vehicle driving style is normal, S vehicle When ≤0.5, the vehicle's driving style is conservative.

[0018] Preferably, the vehicle non-cooperative dynamic game profit matrix established in step S2 includes: considering the safety profit P of the driving style safe , Driving efficiency benefit P considering driving style eff , considering the comfort benefit P of passenger experience com and the economic benefit P considering fuel consumption eco ;

[0019] The time it takes for the head of vehicle i to reach the conflict area is:

[0020]

[0021] in, is the speed of vehicle i at the end of the previous game phase or the speed of vehicle i at the beginning of the current game phase, is the acceleration strategy that vehicle i may adopt from the acceleration candidate set in the current stage, d i is the distance from the head of the current vehicle i to the near edge of the conflict area;

[0022] The time when the rear end of vehicle i just leaves the conflict area is:

[0023]

[0024] Among them, l i is the length of vehicle i, W is the lane width;

[0025] The absolute safety time difference between vehicles i and j is:

[0026] Δt ij =t i -τ j or Δt ij =t j -τ i ;

[0027] Preferably, the Δt ij When it is greater than 0, it is used to concentrate the acceleration candidates according to the safety threshold and exclude the acceleration strategy pairs with negative absolute safety time differences.

[0028] The safety benefit of considering driving style is defined as:

[0029] When vehicle i is an aggressive type:

[0030] When vehicle i is normal:

[0031] When vehicle i is conservative:

[0032] Where Δt max It is the upper limit of the absolute safety time difference.

[0033] The driving efficiency gain considering driving style is defined as:

[0034]

[0035] Among them, 0<θ<1, v max is the maximum speed limit at the intersection, is the speed corresponding to the acceleration strategy that vehicle i may take in the current stage, v ibest is the optimal speed corresponding to the driving style of vehicle i;

[0036] The speed gain increases rapidly at low speeds and saturates exponentially at high speeds.

[0037] The comfort benefit considering passenger experience is defined as:

[0038]

[0039] in, is the possible acceleration in the current stage of the game, is the optimal acceleration determined by the previous game, a max 、a min The maximum acceleration and deceleration are limited.

[0040] The fuel consumption rate is expressed as a function of speed:

[0041]

[0042]

[0043] in, is the average speed of the game vehicles in the current stage, is the speed of vehicle i at the beginning of the current stage game, T is the stage game period, a, e, f, g are constant coefficients;

[0044] The economic benefit considering fuel consumption is defined as:

[0045]

[0046] Among them, E min 、E max They represent the reference lower and upper limits of economic indicators respectively.

[0047] The total return matrix is ​​constructed from four return indicators and their corresponding weight coefficients:

[0048]

[0049] σ1+σ2+σ3+σ4=1

[0050]

[0051]

[0052] Among them, σ1, σ2, σ3, and σ4 represent the weight coefficients of the vehicle's safety requirements, efficiency requirements, comfort requirements, and economy requirements, respectively. The sum of the four weights is equal to 1. At the same time, constraints such as the maximum acceleration and deceleration and the maximum speed of the vehicle when passing through the intersection are set. max and a min They represent the maximum acceleration and maximum deceleration of the vehicle respectively. The minimum speed of the vehicle is 0, v max Indicates road speed limit conditions.

[0053] Preferably, in step S3, the present invention adjusts the decision-making sequence of all participants in each round of the dynamic game in real time according to the possible collision time TTC;

[0054] The possible collision time between two conflicting vehicles is expressed as the time difference between the two vehicles traveling at their current speed and acceleration and reaching the conflict area at the intersection. The specific formula is defined as follows:

[0055]

[0056] The speeds of vehicles i and j at the end of the previous game and the optimal accelerations determined by the previous game are respectively. The possible collision times of the four vehicles at the intersection are T AB ,T BC ,T CD ,T DA .

[0057] Determine the sequential rule as follows: First find the minimum value among the four possible collision times, that is, min{T AB ,T BC ,T CD ,T DA}, for example, if T AB The minimum, then the priority decision maker will consider between car A and car B; on this basis, further compare the T related to car A and car B. DA and T BC The size of T DA >TBC , then car B is the first decision maker, car A is the second decision maker, car C is the third decision maker, and car D is the third decision maker.

[0058] Preferably, the Nash equilibrium solution in step S4 is defined as:

[0059] In the game strategy G={A1,A2…A n ; U1, U2…U n}, there are n game participants, A1, A2…A n represents the set of behavioral strategies of participants 1, 2, ...N, U1, U2...U n For the behavioral benefits of each participant, any strategy of each game participant forms a strategy combination (a1*, a2*…a n *), if for any player i, strategy a i * are all given the strategies of other participants (a1*, a2*…a i-1 *,a i+1 *… a n-1 *a n *) The optimal response strategy for situation i is: U i (a1*,a2*…a i-1 *,a i+1 *…a n *)≥U i (a1*,a2*…a i-1 *,a k *,a i+1 *…a n-1 *a n *);

[0060] For any a k ∈A i If all of them hold true, then the strategy combination (a1*, a2*…a n *) is a Nash equilibrium solution of game G;

[0061] The present invention solves the subgame-perfect Nash equilibrium solution of each round of the stage game by using the backward induction method. Starting from the last stage or the last subgame at the end of the dynamic game tree, the method works backwards and forwards according to the size of the payoff, deleting the actions that are disadvantageous strategies in each optional strategy branch. The Nash equilibrium in each subgame is achieved step by step.

[0062] At the beginning of the next round of the game, the state parameter information of the own vehicle and surrounding vehicles and the road environment information are updated to make a new round of rolling game interactive decision-making, and steps one to four are repeated until all vehicles pass through the intersection safely and smoothly and exit the game, and travel at a constant speed at the time of exit.

[0063] The embodiment of the present invention provides a method for determining the straight-through speed of an intelligent vehicle at an intersection by integrating driving style game theory, which has the following beneficial effects:

[0064] 1. For complex traffic intersections without traffic lights, the system can fully consider the dynamics and uncertainty of other vehicles through game interaction without having to identify the driving intentions of other vehicles, thus enabling safe and effective speed decisions for intelligent vehicles as they travel along a predetermined path.

[0065] 2. Considering the need for personalized autonomous driving decision-making, driving style is integrated into the game-based decision-making algorithm. This allows drivers to make both rational optimal choices and intuitive judgments based on their own driving habits and styles. This effectively eliminates intersection conflicts and clarifies right-of-way while ensuring driving safety. It also meets the efficiency and comfort requirements of various vehicles, improving the fluidity and humanization of intelligent car decision-making at intersections.

[0066] 3. Considering the impact of the order of participants in a dynamic game on the solution, the order of participants' actions is adjusted based on the possible collision time (TTC). This is consistent with the order in which human drivers react to the urgency of a collision when encountering an intersection conflict. This makes the payoff value of each vehicle's decision in each round of the game more realistic and reliable, and more relevant to actual road traffic scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A flowchart of the straight-through speed decision-making process for an intelligent vehicle at an intersection, integrating driving style game theory, provided by an embodiment of the present invention;

[0068] Figure 2 A schematic diagram of the structural principle of the smart car's intersection straight-through speed decision-making integrated with the driving style game provided by an embodiment of the present invention;

[0069] Figure 3 A schematic diagram of a situation in which two vehicles pass through an intersection safely without collision, provided by an embodiment of the present invention;

[0070] Figure 4 A schematic diagram of another situation in which two vehicles pass through an intersection safely without collision, provided by an embodiment of the present invention;

[0071] Figure 5 A schematic diagram illustrating the principle of determining the decision sequence of each vehicle based on the possible collision time TTC provided by an embodiment of the present invention;

[0072] Figure 6 A schematic diagram of a dynamic game tree provided by an embodiment of the present invention;

[0073] Figure 7 A schematic diagram of the reverse induction method for solving the problem provided by an embodiment of the present invention;

[0074] Figure 8 A schematic diagram of the interaction and conflict areas between intelligent vehicles in the through lane of an intersection without signal lights provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0076] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0077] like Figure 1 As shown, an embodiment of the present invention provides a smart car intersection straight-through speed decision method integrating driving style game, which is used to make the optimal decision when a smart car goes straight at an intersection without traffic lights, while meeting safety and other indicators.

[0078] like Figure 2 As shown in the figure, the perception system obtains the status information of the vehicle and surrounding vehicles and identifies the corresponding driving style based on the obtained information. The obtained vehicle information is then used as the input of the non-cooperative dynamic game decision-making system; the optimal acceleration strategy of each vehicle is selected and output through model calculation and the principle of maximum benefit, and the relevant information of the vehicle and surrounding vehicles is updated at the next moment. The above operations are repeated in a cycle. When the vehicle has safely passed all conflict areas of the intersection, it automatically exits the game and resumes normal driving mode.

[0079] The specific steps include:

[0080] S1. When each vehicle reaches the solid line before the intersection, it collects driving status information of surrounding vehicles and identifies the driving style of itself and surrounding vehicles;

[0081] When a vehicle reaches the starting point of the solid line before the intersection, the game begins. The vehicle's perception system obtains driving status information from surrounding vehicles through the V2V system or onboard cameras, millimeter-wave radar, speed and distance sensors, and transmits its own driving status information to the decision-making system. This information includes horizontal and vertical position relative to the geodetic coordinate system, longitudinal speed, longitudinal acceleration, distance to the conflict zone, and surrounding driving environment.

[0082] In addition, the present invention uses the acquired longitudinal acceleration information to identify the driving style through the acceleration change rate, and divides the driving style types into aggressive, normal and conservative;

[0083] The driving style recognition coefficient S is proposedvehicle Defined as:

[0084]

[0085]

[0086] Among them, J(t) is the rate of change of the acceleration of the smart car per unit time, S J is the standard deviation of the acceleration change rate during the driving style identification period, J is the average value of the acceleration change rate during the identification period, and v(t) is the driving speed of the intelligent vehicle;

[0087] The specific classification criteria are: S vehicle When ≥1, the vehicle's driving style is aggressive, 0.5<S vehicle When <1, the vehicle driving style is normal, S vehicle When ≤0.5, the vehicle's driving style is conservative.

[0088] S2. Analyze and record vehicle driving style types and establish a non-cooperative dynamic game payoff matrix for straight-going vehicles at intersections that integrates driving styles;

[0089] The present invention combines intelligent vehicle driving styles with dynamic gaming to establish a dynamic gaming benefit matrix that integrates driving styles. This matrix includes: safety benefits that take into account driving styles; driving efficiency benefits that take into account driving styles; comfort benefits that take into account passenger experience; and economic benefits that take into account fuel consumption. The total benefit matrix is ​​constructed from these four benefit indicators and their corresponding weight coefficients.

[0090] The safety benefit of considering driving style represents a measure of the reliability of whether two vehicles in a potential collision risk are at risk or not, simulating the degree to which different driving styles prioritize right-of-way over absolute safety time difference. An aggressive driving style is more likely to seize the right-of-way in a potential conflict, while a conservative driving style is more likely to yield the right-of-way. Therefore, these two styles prioritize right-of-way over absolute safety time difference. A normal driving style is less concerned with right-of-way and therefore prioritizes the absolute safety time difference.

[0091] The road edges of the lanes where the two vehicles with potential conflict are located are extended to the intersection, and the square area formed by the intersection is the conflict area;

[0092] To ensure the safety of two vehicles passing through the intersection, the necessary and sufficient condition is that the tail of the first vehicle has completely passed the conflict area, and the head of the second vehicle has just reached the conflict area, that is, Figure 3 、 Figure 4 As shown, there are two situations where two cars pass through the intersection safely without collision;

[0093] The time it takes for the head of vehicle i to reach the conflict area is:

[0094]

[0095] in, is the speed of vehicle i at the end of the previous game phase or the speed of vehicle i at the beginning of the current game phase, is the acceleration strategy that vehicle i may adopt from the acceleration candidate set in the current stage, d i is the distance from the head of the current vehicle i to the near edge of the conflict area;

[0096] The time when the rear end of vehicle i just leaves the conflict area is:

[0097]

[0098] Among them, l i is the length of vehicle i, W is the lane width;

[0099] The absolute safety time difference between vehicles i and j is:

[0100] Δt ij =t i -τ j or Δt ij =t j -τ i ;

[0101] Preferably, the Δt ij When it is greater than 0, it is used to concentrate the acceleration candidates according to the safety threshold and exclude the acceleration strategy pairs with negative absolute safety time differences.

[0102] The safety benefit of considering driving style is defined as:

[0103] When vehicle i is an aggressive type:

[0104] When vehicle i is normal:

[0105] When vehicle i is conservative:

[0106] Where Δt max It is the upper limit of the absolute safety time difference.

[0107] The driving efficiency gain considering driving style is defined as:

[0108]

[0109] Among them, 0<θ<1, v max is the maximum speed limit at the intersection, is the speed corresponding to the acceleration strategy that vehicle i may take in the current stage, v ibest is the optimal speed corresponding to the driving style of vehicle i;

[0110] The speed gain increases rapidly at low speeds and saturates exponentially at high speeds.

[0111] The comfort benefit considering passenger experience is defined as:

[0112]

[0113] in, is the possible acceleration in the current stage of the game, is the optimal acceleration determined by the previous game, a max 、a min The maximum acceleration and deceleration are limited;

[0114] The economic benefits of considering fuel consumption represent the impact of the corresponding speed achieved by the acceleration strategy that may be adopted in the current stage of the game on the fuel consumption, which represents the benefits of energy saving from the perspective of economic cost;

[0115] Fuel consumption rate can be expressed as a function of speed:

[0116]

[0117]

[0118] in, is the average speed of the game vehicles in the current stage, is the speed of vehicle i at the beginning of the current game phase, T is the game phase period, a, e, f, g are constant coefficients, and we take a = –0.67944, e = 0.029665, f = –0.00028, g = 0.00000149;

[0119] The economic benefit considering fuel consumption is defined as:

[0120]

[0121] Among them, E min 、E max They represent the reference lower and upper limits of economic indicators respectively.

[0122] The total return matrix is ​​constructed from four return indicators and their corresponding weight coefficients:

[0123]

[0124] σ1+σ2+σ3+σ4=1

[0125]

[0126]

[0127] Among them, σ1, σ2, σ3, and σ4 represent the weight coefficients of the vehicle's safety requirements, efficiency requirements, comfort requirements, and economy requirements, respectively. The sum of the four weights is equal to 1. At the same time, constraints such as the maximum acceleration and deceleration and the maximum speed of the vehicle when passing through the intersection are set. max and a min They represent the maximum acceleration and maximum deceleration of the vehicle respectively. The minimum speed of the vehicle is 0, v max Indicates road speed limit conditions.

[0128] S3. Adjust the decision-making order of all participants in each round of the dynamic game in real time based on the possible collision time TTC;

[0129] In step S3, the present invention adjusts the decision sequence of all participants in each round of the dynamic game in real time according to the possible collision time TTC;

[0130] Since the order of participants in a dynamic game has a significant impact on the optimal solution, the present invention adjusts the decision-making order of all participants in each round of the dynamic game in real time based on the possible collision time TTC to ensure the orderly progress of the entire game decision-making;

[0131] The possible collision time between two conflicting vehicles is expressed as the time difference between the two vehicles traveling at their current speed and acceleration and reaching the conflict area at the intersection. The specific formula is defined as follows:

[0132]

[0133] They are the speed of vehicles i and j at the end of the previous game and the optimal acceleration determined by the previous game, the time when the head reaches the conflict area, and the possible collision time of the four vehicles at the intersection is T. AB ,T BC ,T CD ,T DA .

[0134] Determine the sequential rule as follows: First find the minimum value among the four possible collision times, that is, min{T AB ,T BC ,T CD ,T DA}, for example, if T ABThe minimum, then the priority decision maker will consider between car A and car B; on this basis, further compare the T related to car A and car B. DA and T BC The size of T DA >T BC , then car B is the first decision maker, car A is the second decision maker, car C is the third decision maker, and car D is the fourth decision maker. The overall decision order is: B→A→C→D; if T DA <T BC , then car A is the first decision maker, car B is the second decision maker, car D is the third decision maker, and car C is the fourth decision maker. The overall decision order is: A→B→D→C. Other different order situations can be determined by this rule.

[0135] like Figure 5 As shown in Figure 1, there are eight situations in which the decision order of each vehicle is determined by the possible collision time TTC, and all of them can be determined by this rule.

[0136] S4. Solve the subgame-perfect Nash equilibrium solution of each round of the game through backward induction, and decide the optimal acceleration strategy for each car in this round;

[0137] Establish the candidate set of acceleration for each car: set a max =4m / s, a min =-4m / s, collective accuracy Δa = 0.4;

[0138] like Figure 6 As shown in the figure, the payoff matrix is ​​used to calculate the payoff of each car corresponding to each acceleration strategy combination in the candidate set, and a dynamic game tree is generated according to the order of each car;

[0139] In the game strategy G={A1,A2…A n ; U1, U2…U n}, there are n game participants, A1, A2…A n represents the set of behavioral strategies of participants 1, 2, ...N, U1, U2...U n For the behavioral benefits of each participant, any strategy of each game participant forms a strategy combination (a1*, a2*…a n *), if for any player i, strategy a i * are all given the strategies of other participants (a1*, a2*…a i-1 *,a i+1 *… a n-1 *a n *) The optimal response strategy for situation i is: U i (a1*,a2*…a i-1 *,a i+1*…a n *)≥U i (a1*,a2*… a i-1 *,a k *,a i+1 *…a n-1 *a n *);

[0140] For any a k ∈A i If all of them hold true, then the strategy combination (a1*, a2*…a n *) is a Nash equilibrium solution of game G;

[0141] In a dynamic game, the first-acting participant will inevitably consider the strategy choices of the later-acting participants in the later stages when choosing their strategies in the early stages. Only the participants in the final stage can make their choices directly without being constrained by other participants. Once the choices of the participants in the later stages are determined, the strategies of the participants in the previous stage are also determined.

[0142] like Figure 7 As shown, the present invention solves the subgame-perfect Nash equilibrium solution of each round of the stage game by backward induction. Starting from the last stage or the last subgame at the end of the dynamic game tree, according to the size of the participants' benefits from the end to the beginning, in each branch, the principle of maximizing the own benefits is adopted, and the actions that are disadvantageous strategies in each optional strategy branch are deleted in reverse order. The Nash equilibrium in each subgame is achieved step by step.

[0143] The subgame-perfect Nash equilibrium solution removes Nash equilibria that contain unbelievable threats, so that the equilibrium strategy no longer contains unbelievable threats. This allows us to determine the optimal acceleration strategy for each vehicle in this round, taking into account the actual traffic environment.

[0144] S5. Update the vehicle status and environment information and start the next round of the game. Repeat all the above steps until all smart cars safely pass through the intersection and exit the game.

[0145] At the beginning of the next round of the game, the state parameter information of the own vehicle and surrounding vehicles and the road environment information are updated to make a new round of rolling game interactive decision-making, and steps one to four are repeated until all vehicles pass through the intersection safely and smoothly and exit the game, and travel at a constant speed at the time of exit.

[0146] Each smart car will pass through two conflict areas, so the condition for a car to exit the game is set to when the car has completely passed the second conflict area. That is, at this point, the car can be considered to have no impact on the driving safety of other smart cars;

[0147] The applicable scenarios of the embodiments of the present invention are as follows Figure 8 As shown, due to the restrictions of road traffic signs, the current intersection only allows vehicles to go straight and does not allow vehicles to turn left or right. When vehicles go straight at the intersection, they must stay in the middle lane;

[0148] There is a conflict area between the future driving trajectories of vehicles A, B, C, and D in the through lane of the intersection;

[0149] Each vehicle needs to pass through two conflict areas, that is, there are two vehicles with potential collision impact with it;

[0150] In particular, there is no conflict between car A and car C, and there is no conflict between car B and car D;

[0151] The parameter information vector that each vehicle needs to obtain is expressed as

[0152] This paper proposes a method for intelligent vehicle intersection speed decision-making that incorporates a driving style game. This method addresses the problem of intelligent vehicle navigation in complex, open intersections without signal lights. By factoring vehicle driving styles into the dynamic game process, a driving style-based game payoff matrix is ​​designed. The dynamic game is solved to determine the optimal behavior strategy for each vehicle, enabling the intelligent vehicle to navigate the intersection safely and smoothly while also conforming to the operating habits of human drivers.

[0153] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for intelligent vehicle intersection straight-through speed decision-making based on driving style game, characterized by: The intelligent vehicle intersection straight speed decision method integrating driving style game comprises the following steps: S1. When each vehicle reaches the solid line before the intersection, it collects driving status information of surrounding vehicles and identifies the driving style of itself and surrounding vehicles; S2. Analyze and record vehicle driving style types and establish a non-cooperative dynamic game payoff matrix for straight-going vehicles at intersections that integrates driving styles; S3. Adjust the decision-making order of all participants in each round of the dynamic game in real time based on the possible collision time TTC; S4. Solve the subgame-perfect Nash equilibrium solution of each round of the game through backward induction, and decide the optimal acceleration strategy for each car in this round; S5: Update vehicle status and environment information and start the next round of the game. Repeat all the above steps until all smart cars safely pass through the intersection and exit the game. The vehicle non-cooperative dynamic game profit matrix established in step S2 includes: the safety profit P considering the driving style safe , Driving efficiency benefit P considering driving style eff , considering the comfort benefit P of passenger experience com and the economic benefit P considering fuel consumption eco ; The time it takes for the head of vehicle i to reach the conflict area is: in, is the speed of vehicle i at the end of the previous game phase or the speed of vehicle i at the beginning of the current game phase, is the acceleration strategy that vehicle i may adopt from the acceleration candidate set in the current stage, d i is the distance from the head of the current vehicle i to the near edge of the conflict area; The time when the rear end of vehicle i just leaves the conflict area is: Among them, l i is the length of vehicle i, W is the lane width; The absolute safety time difference between vehicles i and j is: Δt ij =t i -t j orΔt ij =t j -t i ; The total return matrix is ​​constructed from four return indicators and their corresponding weight coefficients: P i n =σ1P safe +σ2P eff +σ3P com +σ4P eco σ1+σ2+σ3+σ4=1 Among them, σ1, σ2, σ3, and σ4 represent the weight coefficients of the vehicle's safety requirements, efficiency requirements, comfort requirements, and economy requirements, respectively. The sum of the four weights is equal to 1. At the same time, the maximum acceleration and deceleration and the maximum speed of the vehicle when passing through the intersection are set. max and a min They represent the maximum acceleration and maximum deceleration of the vehicle respectively. The minimum speed of the vehicle is 0, v max Indicates road speed limit conditions; The safety benefit of considering driving style is defined as: When vehicle i is an aggressive type: When vehicle i is normal: When vehicle i is conservative: Where Δt max It is the upper limit of the absolute safety time difference; The driving efficiency benefit considering driving style is defined as: Among them, 0<θ<1, v max is the maximum speed limit at the intersection, is the speed corresponding to the acceleration strategy that vehicle i may take in the current stage, vi best is the optimal speed corresponding to the driving style of vehicle i; The fuel consumption rate is expressed as a function of speed: in, is the average speed of the game vehicles in the current stage, is the speed of vehicle i at the beginning of the current stage game, T is the stage game period, a, e, f, g are constant coefficients; The economic benefit considering fuel consumption is defined as: Among them, E min 、E max Respectively represent the reference lower and upper bounds of the economic indicators; In step S3, the present invention adjusts the decision-making sequence of all participants in each round of the dynamic game in real time according to the possible collision time TTC; The possible collision time between two conflicting vehicles is expressed as the time difference between the two vehicles traveling at their current speed and acceleration and reaching the conflict area at the intersection. The specific formula is defined as follows: The speeds of vehicles i and j at the end of the previous game and the optimal accelerations determined by the previous game are respectively. The possible collision times of the four vehicles at the intersection are T AB ,T BC ,T CD ,T DA .

2. The intelligent vehicle intersection straight speed decision method integrating driving style game according to claim 1 is characterized in that: In step S1, the driving style is identified by the rate of change of acceleration and is classified into aggressive, normal, and conservative types; The driving style recognition coefficient S is proposed vehicle Defined as: Among them, J(t) is the rate of change of the acceleration of the smart car per unit time, S J is the standard deviation of the acceleration change rate during the driving style identification period, J is the average value of the acceleration change rate during the identification period, and v(t) is the driving speed of the intelligent vehicle; S vehicle When ≥1, the vehicle's driving style is aggressive, 0.5<S vehicle When <1, the vehicle driving style is normal, S vehicle When ≤0.5, the vehicle's driving style is conservative.

3. The intelligent vehicle intersection straight speed decision method integrating driving style game according to claim 1 is characterized in that: The Δt ij When it is greater than 0, it is used to concentrate the acceleration candidates according to the safety threshold and exclude the acceleration strategy pairs with negative absolute safety time differences.

4. The intelligent vehicle intersection straight speed decision method integrating driving style game according to claim 1 is characterized in that: The comfort benefit considering passenger experience is defined as: in, is the possible acceleration in the current stage of the game, is the optimal acceleration determined by the previous game, a max 、a min The maximum acceleration and deceleration are limited.

Images