A hierarchical game-based intelligent vehicle brain-like decision-making method for ramp merging scenarios

Abstract

The application discloses a kind of ramp convergence scene intelligent car brain-like decision method based on hierarchical game, including the following steps: information sharing is carried out between intelligent car and main road vehicle;Intelligent car and main road vehicle interactive decision implementation logic is constructed, and intelligent car in ramp is divided to hierarchical game top space, and main road vehicle is divided to hierarchical game bottom space;Internal driving incentive model is constructed;Hierarchical game optimization model is constructed based on the internal driving incentive model;Under the behavior coordination constraint condition of construction, the hierarchical game optimization model is solved, and the stable equilibrium strategy that intelligent car and main road vehicle should implement in current stage game is obtained;Intelligent car and main road vehicle execute respective stable equilibrium strategy;The above steps are repeated until intelligent car is safely converged into main road from ramp.The application helps to improve the safety level of intelligent car autonomous decision in ramp convergence scene and the brain-like humanization level of intelligent car driving behavior.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent vehicles, in particular to a ramp merging scene intelligent vehicle brain-like decision-making method based on hierarchical game. BACKGROUND

[0002] At present, intelligent vehicles have become the strategic direction of global automobile industry development. Developing intelligent vehicles is conducive to breaking through key technical bottlenecks, improving industrial foundation capabilities, enhancing leadership in a new round of technological revolution and industrial change, and is also conducive to accelerating the transformation and upgrading of the automobile industry, cultivating new development advantages of the industry, and forming new economic growth poles. The "Intelligent Vehicle Innovation and Development Strategy" lists "intelligent decision-making control" as a key basic technology that needs to be broken through in building a collaborative and open intelligent vehicle technology innovation system. Decision-making is the key to improving the intelligent level of intelligent vehicles and accurately and smoothly completing various driving tasks. Brain-like thinking logic is a necessary prerequisite for intelligent vehicles to truly and naturally integrate into complex traffic ecology. Only when the brain-like thinking logic of the driving behavior of intelligent vehicles is similar to that of human drivers and can be understood and accepted by other traffic participants, can intelligent vehicles overcome the obstacle of social acceptance of intelligent vehicles pointed out by the global well-known Boston Consulting Group at the World Economic Forum.

[0003] Ramp merging is a typical dangerous scene. The merging behavior of intelligent vehicles on ramps has strong interactivity and is strongly dependent on the traffic situation.

[0004] Therefore, how to provide a ramp merging scene intelligent vehicle brain-like decision-making method, which can enable intelligent vehicles to make decisions in a way that conforms to the brain-like thinking logic of human drivers, effectively resolve potential traffic conflicts in continuous interaction with main road vehicles, and accurately and smoothly merge into the main road, has become a problem that needs to be solved by those skilled in the art. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a ramp merging scene intelligent vehicle brain-like decision-making method based on hierarchical game.

[0006] In order to achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0007] A ramp merging scene intelligent vehicle brain-like decision-making method based on hierarchical game, comprising the following steps:

[0008] S1: Intelligent vehicles and main road vehicles share information;

[0009] Constructing an interactive decision-making implementation logic of intelligent vehicles and main road vehicles, dividing intelligent vehicles in the ramp into a top layer space of hierarchical game, and dividing main road vehicles into a bottom layer space of hierarchical game;

[0010] S2: constructing an intrinsic driving incentive model; constructing a top-level space intelligent vehicle game decision objective function and a bottom-level space main road vehicle game decision objective function based on the intrinsic driving incentive model;

[0011] S3: constructing a hierarchical game optimization model based on the top-level space intelligent vehicle game decision objective function and the bottom-level space main road vehicle game decision objective function;

[0012] S4: solving the hierarchical game optimization model under the constructed behavior coordination constraint condition to obtain a stable equilibrium strategy that should be implemented by the intelligent vehicle and the main road vehicle in the current stage game;

[0013] S5: the intelligent vehicle and the main road vehicle execute the respective stable equilibrium strategies;

[0014] S6: repeating S1-S5 until the intelligent vehicle safely merges into the main road from the ramp.

[0015] Preferably, the intrinsic driving incentive model includes a first-type intrinsic driving incentive model, a second-type intrinsic driving incentive model and a third-type intrinsic driving incentive model;

[0016] The first-type intrinsic driving incentive model includes an intelligent vehicle driving intrinsic driving incentive model and a main road vehicle driving intrinsic driving incentive model;

[0017] The second-type intrinsic driving incentive model includes an intelligent vehicle ontology intrinsic driving incentive model and a main road vehicle ontology intrinsic driving incentive model;

[0018] The third-type intrinsic driving incentive model includes an intelligent vehicle passenger intrinsic driving incentive model and a main road vehicle passenger intrinsic driving incentive model.

[0019] Preferably, the intelligent vehicle driving intrinsic driving incentive model is:

[0020]

[0021]

[0022] wherein, W is a kinematics equation, OR is a ramp; is an effective distance of the intelligent vehicle in the τth round of game; is a component along an X direction; is a component along a Y direction; is a component along an X direction; is a component along a Y direction; is a component along an X direction; is a component along a Y direction; is a component along an X direction; is a component along a Y direction; is an arbitrary strategy taken from the set of game decision strategies of the intelligent vehicle in the τth round of the game; is the instantaneous speed of the intelligent vehicle in the τth round of the game; is the spatial position coordinate of the intelligent vehicle at the beginning of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the intelligent vehicle at the beginning of the τth round of the game; is the component along the Y direction; is the decision period of the intelligent vehicle in one round of the game; τ+1 and t τ are the decision time stamp of the (τ+1)th round of the game and the decision time stamp of the τth round of the game, respectively; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the component along the Y direction; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the intelligent vehicle at the completion of the τth round of the game; is the component along the Y direction.

[0023] The preferred model of the intrinsic driving incentive of the host vehicle driver is :

[0024]

[0025] wherein W is a kinematic equation, MR is the host road; is the effective distance of the host vehicle in the τth round of the game; is the spatial position coordinate of the host vehicle at the beginning of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the host vehicle at the beginning of the τth round of the game; is the component along the Y direction; is the spatial position coordinate of the host vehicle at the beginning of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the host vehicle at the beginning of the τth round of the game; is the component along the Y direction; is an arbitrary strategy taken from the set of game decision strategies of the host vehicle in the τth round of the game; is the instantaneous speed of the host vehicle in the τth round of the game; is the spatial position coordinate of the host vehicle at the completion of the τth round of the game; is the component along the X direction; is the spatial position coordinate of the host vehicle at the completion of the τth round of the game; is the component along the Y direction; is the decision period of the host vehicle in one round of the game; τ+1 and t τ are the decision time stamp of the (τ+1)th round of the game and the decision time stamp of the τth round of the game, respectively; The spatial coordinates of the main road vehicle when the game is completed in the τth round. for The component along the X direction; for The component along the Y direction; The spatial coordinates of the main road vehicle at the start of the τth round of the game; for The component along the X direction; for The component along the Y direction.

[0026] Preferably, the intelligent vehicle's intrinsic driving force model for:

[0027]

[0028] Where W represents the kinematic equation. for The component along the X direction; for The component along the Y direction; Let τ be any strategy adopted from the decision strategy set of the intelligent vehicle game in the τth round of the game. for The component along the X direction; for The component along the Y direction; Let be the transient speed of the intelligent vehicle in the τth round of the game; For intelligent vehicles, this refers to the decision-making cycle of a stage of game theory. and Let be the spatial coordinates of the intelligent vehicle at the start of the τth round of the game. and Let r be the spatial coordinates of the intelligent vehicle at the end of the τth round of the game; OR r is the radius of the circumcircle of the intelligent vehicle's exterior. MR The radius of the circumcircle of the main road vehicle's shape.

[0029] Preferably, the intrinsic driving motive model of the main road vehicle body for:

[0030]

[0031] Where W represents the kinematic equation. for The component along the X direction; for The component along the Y direction; Let τ be any strategy adopted from the decision strategy set of the main road vehicles in the τth round of the game. for The component along the X direction; for The component along the Y direction; Let be the transient speed of the vehicles on the main road during the τth round of the game; The decision-making cycle of vehicles on the main road in a phase of game; and The spatial coordinates of the main road vehicle at the start of the τth round of the game; and r represents the spatial coordinates of the main road vehicle at the end of the τth round of the game; OR r is the radius of the circumcircle of the intelligent vehicle's exterior. MR The radius of the circumcircle of the main road vehicle's shape.

[0032] Preferably, the intelligent vehicle driver and passenger intrinsic motivation model for:

[0033]

[0034] in, Let be any strategy adopted from the decision strategy set of the intelligent vehicle game in the τth round of the game. This represents the stable equilibrium strategy reached by the intelligent vehicle in the τ-1 round of the game. Let t be the stable equilibrium strategy reached by the intelligent vehicle in the (τ-2)th round of the game. τ-2 t τ-1 and t τ The timestamps are, in order, the decision timestamps for the (τ-2)th round of the stage game, the (τ-1)th round of the stage game, and the decision timestamps for the τth round of the stage game; Denotes any strategy adopted in the τth round of the game. The resulting degree of urgency; This represents the stable equilibrium strategy reached in the (τ-1)th round of the phase game. The resulting urgency.

[0035] Preferably, the intrinsic driving motivation model for drivers and passengers of main road vehicles for:

[0036]

[0037] in, Let t be any strategy adopted from the decision strategy set of the main road vehicles in the τth round of the game. The stable equilibrium strategy reached by the main road vehicles in the τ-1 round of the game. the stable equilibrium strategy reached by the main road vehicle in the (t-2)-th round of game, t τ-2 τ-1 τ are the decision time stamps of the (t-2)-th round of game, the (t-1)-th round of game and the t-th round of game respectively; represents the jerk caused by any strategy adopted in the t-th round of game; represents the jerk caused by the stable equilibrium strategy reached in the (t-1)-th round of game.

[0038] Preferably, the top-level space intelligent vehicle game decision objective function and the bottom-level space main road vehicle game decision objective function are:

[0039]

[0040] wherein, is the top-level space intelligent vehicle game decision objective function, is the bottom-level space main road vehicle game decision objective function, π OR ,σ OR ,μ OR are characteristic factors of the intelligent vehicle individualized embedding module, ε MR ,δ MR ,γ MR are characteristic factors of the main road vehicle individualized embedding module.

[0041] Preferably, the hierarchical game optimization model is:

[0042]

[0043] wherein, the behavior coordination constraint conditions include top-level space intelligent vehicle behavior coordination constraint conditions and bottom-level space main road vehicle behavior coordination constraint conditions g and r are constraint condition numbers, g = 1, 2, 3…G, r = 1, 2, 3…R; G and R represent the upper limit of the constraint condition number;

[0044]

[0045] wherein, and are maximum acceleration constraints of the intelligent vehicle and the main road vehicle respectively, and are minimum acceleration constraints of the intelligent vehicle and the main road vehicle respectively, and are road speed limit constraints of the ramp and the main road respectively, and​​ minimum speed constraints of ramp and main road respectively, maximum jerk constraints of intelligent vehicle and main road vehicle respectively.

[0046] Through the above technical solutions, compared with the prior art, the application provides a ramp merging scene intelligent vehicle brain-like decision method based on hierarchical game, which can obtain the following beneficial technical effects:

[0047] 1: It helps to improve the safety level of intelligent vehicle autonomous decision and the brain-like humanization level of intelligent vehicle driving behavior, so that the difficult ramp merging scene can be effectively dealt with;

[0048] 2: It helps to protect the life and property safety of intelligent vehicle passengers;

[0049] 3: It helps to promote the development of traffic power and intelligent transportation strategy, and lays a foundation for large-scale industrialization of intelligent vehicles. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced below. Obviously, the drawings in the following description are only embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0051] Figure 1 The overall flowchart of the intelligent vehicle brain-like decision method based on hierarchical game provided by the application in a ramp merging scene;

[0052] Figure 2 The schematic diagram of a typical ramp merging scene to which the application is applied;

[0053] Figure 3 The schematic diagram of hierarchical game rolling stage by stage provided by the application;

[0054] Figure 4 The schematic diagram of the intelligent vehicle and main road vehicle interactive decision implementation logic constructed by the application;

[0055] Figure 5 The flowchart of the intelligent vehicle brain-like decision method based on hierarchical game provided by the application in a ramp merging scene;

[0056] Figure 6 The intelligent vehicle brain-like decision result diagram of the application in a certain specific application. DETAILED DESCRIPTION

[0057] ​With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below, obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work belong to the protection scope of the present application.

[0058] Human driving is a closed-loop process of continuous optimization, real-time rolling, continuous adjustment and timely feedback. The present application takes human driving wisdom as the natural blueprint, regards intelligent cars and main road vehicles as intelligent agents with autonomous and flexible functions, fully considers the sequential hierarchical nature of multi-vehicle interactive decision-making, deeply excavates the internal driving incentives in the human driving process, makes up for the shortcomings of the traditional one-time decision-making idea that cannot effectively deal with complex random uncertain factors in traffic and ignores the continuous interaction of decision-making at a deeper level, and is beneficial to fundamentally improve the autonomous decision-making performance of intelligent cars.

[0059] The collaborative solution of intelligent cars in ramp merging scenarios with main road vehicles is a very challenging task, which significantly increases the dynamics, continuity and interaction of decision-making. Excessive aggressive and irrational blind merging behavior can significantly affect the normal driving of main road vehicles, forcing them to slow down to ensure driving safety, further causing a series of chain reactions, and even causing traffic accidents in severe cases. However, overly cautious and conservative merging behavior can cause intelligent cars in the ramp to be unable to complete the task of merging into the main road for a long time, or even produce multiple repeated game exploration behaviors at the end of the ramp, greatly reducing the traffic efficiency of the ramp. The present application effectively solves the problems faced by intelligent cars in ramp merging scenarios by designing a decision-making method similar to the brain thinking logic of human drivers.

[0060] As shown in Figures 1-5 The embodiment of the present application discloses a hierarchical game-based intelligent car brain-like decision-making method for ramp merging scenarios, comprising the following steps:

[0061] S1: Intelligent cars and main road vehicles share information;

[0062] Constructing an interactive decision-making implementation logic for intelligent cars and main road vehicles, dividing intelligent cars in the ramp into a top-level space of hierarchical game, and dividing main road vehicles into a bottom-level space of hierarchical game;

[0063] S2: Constructing an internal driving incentive model; based on the internal driving incentive model, constructing a top-level space intelligent car game decision-making objective function and a bottom-level space main road vehicle game decision-making objective function;

[0064] S3: Based on the top-level space intelligent car game decision-making objective function and the bottom-level space main road vehicle game decision-making objective function, constructing a hierarchical game optimization model;

[0065] S4: solving the hierarchical game optimization model under the constructed behavioral coordination constraint condition to obtain a stable equilibrium strategy implemented by the intelligent vehicle and the main road vehicle in the current stage game;

[0066] S5: the intelligent vehicle and the main road vehicle execute the respective stable equilibrium strategies;

[0067] S6: repeating S1-S5 until the intelligent vehicle safely merges into the main road from the ramp.

[0068] It can be understood that the above-mentioned constructed intelligent vehicle and main road vehicle interactive decision implementation logic fully considers the initiative and inducement of the intelligent vehicle in the ramp, which is the premise of the game interaction with the main road vehicle. The intelligent vehicle in the ramp conveys its driving intention to the main road vehicle through appropriate driving signals or driving behaviors (driving signals include turn signal, etc., and driving behaviors include appropriate lateral movement behavior, etc.), and the main road vehicle makes reasonable and appropriate response according to the preceding behavior of the intelligent vehicle in the ramp. The application excavates the essence of the sequential nature of the decision-making behavior of the intelligent vehicle in the ramp and the main road vehicle. The intelligent vehicle in the ramp is divided into a hierarchical game top space with sequential action initiative, and the main road vehicle is divided into a hierarchical game bottom space with sequential action follow-up. The information flow, behavior group and situation field dynamically circulate between the top space and the bottom space, forming a tightly coupled collaborative decision-making whole.

[0069] It can be understood that the intelligent vehicle in the ramp and the main road vehicle share information through communication devices and environmental perception sensors. The information shared by the intelligent vehicle and the main road vehicle includes motion speed, acceleration, driving direction, driving intention, style type and positioning coordinates, etc.

[0070] The communication devices and environmental perception sensors include vehicle-to-vehicle communication equipment, vehicle-to-road communication equipment, perception cameras, laser radars and millimeter wave radars, etc.

[0071] In an embodiment, the internal driving incentive model includes a first type of internal driving incentive model, a second type of internal driving incentive model and a third type of internal driving incentive model.

[0072] The first type of internal driving incentive model includes an intelligent vehicle driving internal driving incentive model and a main road vehicle driving internal driving incentive model.

[0073] The second type of internal driving incentive model includes an intelligent vehicle ontology internal driving incentive model and a main road vehicle ontology internal driving incentive model.

[0074] The third type of intrinsic driving incentive model includes an intelligent automobile driver intrinsic driving incentive model and a main road vehicle driver intrinsic driving incentive model.

[0075] It should be noted that:

[0076] The first type of intrinsic driving incentive model faces the overall demand of the ramp merging scene, aims to improve the traffic timeliness of the ramp merging scene from the perspective of the overall demand of the ramp merging scene, make the intelligent automobile in the ramp and the main road vehicle efficiently resolve the right-of-way conflict between them through hierarchical game interaction, and make the intelligent automobile in the ramp merge into the main road as soon as possible, and realize the goal of efficient traffic dredging of the ramp merging scene. The present application firstly aims at the single element stage game process, takes the relationship between the effective distance traveled by adopting any strategy from the respective game decision strategy set and the ideal distance traveled by maintaining the initial speed at the beginning of the hierarchical game of this stage as the core of establishing the intelligent automobile driving intrinsic driving incentive model and the main road vehicle driving intrinsic driving incentive model induced by the overall demand of the ramp merging scene.

[0077] The second type of intrinsic driving incentive model faces the intrinsic demand of the intelligent automobile and the intrinsic demand of the main road vehicle, aims to simulate the thinking cognition of the driver in the driving process when he is in the ramp and the main road branch respectively, and measure the spatial safety level of the intelligent automobile in the ramp and the main road vehicle in the merging interaction process. The present application focuses on the future spatial position of the intelligent automobile in the ramp and the main road vehicle at the completion of the hierarchical game of this stage, and takes the exponential function of the distance between the future spatial positions of the intelligent automobile and the main road vehicle as the core of establishing the intrinsic driving incentive model of the intelligent automobile and the intrinsic driving incentive model of the main road vehicle, which is jointly determined by the arbitrary strategy adopted by the intelligent automobile and the main road vehicle from the respective game decision strategy set and the current traffic situation, and considers the information such as the size of the intelligent automobile and the main road vehicle in modeling.

[0078] The Type III intrinsic driving incentive model addresses the human needs of drivers and passengers, involving both intelligent vehicle drivers and passengers and those in main road vehicles. It aims to improve the travel experience of both groups by penalizing frequent changes in acceleration and braking modes in intelligent vehicles and main road vehicles, thus avoiding negative driving experiences and increasing the acceptance and satisfaction of autonomous decision-making in intelligent vehicles. This invention proposes a jump amplitude (i.e., jerkiness) index to measure acceleration changes, serving as the basis for establishing the intrinsic driving incentive models for both intelligent vehicle and main road vehicle drivers and passengers. The jump amplitude index comprehensively considers the changes in intelligent vehicle decision-making strategies brought about by the current and previous stages of the game in a time-series manner. This is related to any strategy that intelligent vehicles and main road vehicles may adopt from their respective game decision-making strategy sets in the current stage of the game, as well as the stable equilibrium strategy pair reached by both parties after a complete round of hierarchical game interaction in the previous stage.

[0079] In one embodiment, the intelligent vehicle driving intrinsic driving incentive model for:

[0080]

[0081] Where W is the kinematic equation and OR is the ramp; The effective path for intelligent vehicles in the τth round of the game; for The component along the X direction; for The component along the Y direction; for The component along the X direction; for The component along the Y direction; For the decision strategy set of the intelligent vehicle game in the τth round stage game Any strategy adopted in China; Let be the transient speed of the intelligent vehicle in the τth round of the game; for The component along the X direction; for The component along the Y direction; For intelligent vehicles, this represents the decision-making cycle of a phased game; t τ+1 and t τ These are the decision timestamps for the (τ+1)th round of the stage game and the decision timestamp for the τth round of the stage game, respectively. Let be the spatial coordinates of the intelligent vehicle when the game in the τth round is completed; for The component along the X direction; for component along the Y direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction.

[0082] In an embodiment, the main road vehicle driving intrinsic driving incentive model is:

[0083]

[0084]

[0085] wherein W is a kinematic equation, MR is a main road; is the effective distance of the main road vehicle in the τth round of game; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction. is any strategy taken from the main road vehicle game decision strategy set in the τth round of game; is the instantaneous speed of the main road vehicle in the τth round of game; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction; is the decision period of the main road vehicle in a round of game; t τ+1 and t τ are the decision time stamp of the τ+1th round of game and the decision time stamp of the τth round of game, respectively; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the X direction; is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; component along the Y direction.

[0086] In an embodiment, the intelligent vehicle body intrinsic driving incentive model For:

[0087]

[0088] where W represents the kinematic equation, For the component along the X direction; For the component along the Y direction; For any strategy taken by the intelligent vehicle from the set of game decision strategies in the τth round of game; For the component along the X direction; For the component along the Y direction; For the instantaneous speed of the intelligent vehicle in the τth round of game; For the decision period of the intelligent vehicle in one round of game; For and For the spatial position coordinates of the intelligent vehicle at the beginning of the τth round of game; and For the spatial position coordinates of the intelligent vehicle at the completion of the τth round of game; OR For the circumscribed circle radius of the shape of the intelligent vehicle; MR For the circumscribed circle radius of the shape of the main road vehicle.

[0089] In an embodiment, the main road vehicle body is driven by a driving incentive model For:

[0090]

[0091] where W represents the kinematic equation, For the component along the X direction; For the component along the Y direction; For any strategy taken by the main road vehicle from the set of game decision strategies in the τth round of game; For the component along the X direction; For the component along the Y direction; For the instantaneous speed of the main road vehicle in the τth round of game; For the decision period of the main road vehicle in one round of game; For and For the spatial position coordinates of the main road vehicle at the beginning of the τth round of game; and For the spatial position coordinates of the main road vehicle at the completion of the τth round of game;OR is the circumscribed circle radius of the intelligent vehicle shape; r MR is the circumscribed circle radius of the host vehicle shape.

[0092] In an embodiment, the intelligent vehicle driver intrinsic motivation model is:

[0093]

[0094] wherein, is any strategy taken from the intelligent vehicle game decision strategy set in the tth round of game, is the stable equilibrium strategy reached by the intelligent vehicle in the (t-1)th round of game, is the stable equilibrium strategy reached by the intelligent vehicle in the (t-2)th round of game, t τ-2 , t τ-1 and t τ are the decision time stamp of the (t-2)th round of game, the decision time stamp of the (t-1)th round of game and the decision time stamp of the tth round of game, respectively; represents the jerk caused by any strategy taken in the tth round of game . represents the jerk caused by the stable equilibrium strategy reached in the (t-1)th round of game .

[0095] In an embodiment, the host vehicle driver intrinsic motivation model is:

[0096]

[0097] wherein, is any strategy taken from the host vehicle game decision strategy set in the tth round of game, is the stable equilibrium strategy reached by the host vehicle in the (t-1)th round of game, is the stable equilibrium strategy reached by the host vehicle in the (t-2)th round of game, t τ-2 , t τ-1 and t τ are the decision time stamp of the (t-2)th round of game, the decision time stamp of the (t-1)th round of game and the decision time stamp of the tth round of game, respectively; represents the jerk caused by any strategy taken in the tth round of game . represents the jerk caused by the stable equilibrium strategy reached in the (t-1)th round of game .

[0098] In an embodiment, the top-level space intelligent vehicle game decision objective function and the bottom-level space main road vehicle game decision objective function are:

[0099]

[0100] wherein, is the top-level space intelligent vehicle game decision objective function, is the bottom-level space main road vehicle game decision objective function, π OR ,σ OR ,μ OR is a characteristic factor of the intelligent vehicle individualized embedded module, ε MR ,δ MR ,γ MR is a characteristic factor of the main road vehicle individualized embedded module.

[0101] As can be seen from the above, the top-level space intelligent vehicle game decision objective function is composed of an intelligent vehicle driving intrinsic driving incentive model, an intelligent vehicle ontology intrinsic driving incentive model, an intelligent vehicle driver and passenger intrinsic driving incentive model, and an intelligent vehicle individualized embedded module. The bottom-level space main road vehicle game decision objective function is composed of a main road vehicle driving intrinsic driving incentive model, a main road vehicle ontology intrinsic driving incentive model, a main road vehicle driver and passenger intrinsic driving incentive model, and a main road vehicle individualized embedded module.

[0102] In an embodiment, the hierarchical game optimization model is:

[0103]

[0104] wherein, the behavior coordination constraint conditions include top-level space intelligent vehicle behavior coordination constraint conditions and bottom-level space main road vehicle behavior coordination constraint conditions g and r are constraint condition numbers, g = 1, 2, 3…G, r = 1, 2, 3…R; G and R represent the upper limit of the number of constraint conditions;

[0105] The behavior coordination constraint conditions provide rationality guarantee for the selection of intelligent vehicle and main road vehicle decision strategies, so that they can be naturally integrated with the road environment.

[0106]

[0107] wherein, and are maximum acceleration constraints of the intelligent vehicle and the main road vehicle, respectively, and are minimum acceleration constraints of the intelligent vehicle and the main road vehicle, respectively, and road speed limit constraints for ramp and main road respectively, and minimum speed constraints for ramp and main road respectively, and maximum jerk constraints for intelligent vehicle and main road vehicle respectively.

[0108] As can be seen from the above: the hierarchical game optimization model is composed of a top-level space intelligent vehicle game decision objective function and a bottom-level space main road vehicle game decision objective function nested; wherein the top-level space intelligent vehicle game decision objective function is related to the decision strategy and driving state of the intelligent vehicle itself, and also related to the decision strategy and driving state of the bottom-level space main road vehicle; similarly, the bottom-level space main road vehicle game decision objective function is related to the decision strategy and driving state of itself, and also related to the decision strategy and driving state of the top-level space intelligent vehicle, showing a complex coupling and nesting feature. The present application takes the top-level space intelligent vehicle game decision objective function as the outer model in the hierarchical game optimization model, and takes the bottom-level space main road vehicle game decision objective function as the kernel model in the hierarchical game optimization model, to form a double-level space collaborative game task.

[0109] In an embodiment, the chassis system of the intelligent vehicle and the main road vehicle executes the respective stable equilibrium strategy through controlling the throttle subsystem, the brake subsystem and the steering subsystem.

[0110] In summary, the intelligent vehicle in the ramp of the present application resolves the ramp merging conflict through the hierarchical game of rolling evolution with the main road vehicle in stages, and realizes the goal of safely and smoothly merging into the main road.

[0111] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same and similar parts between the embodiments can be referred to each other. For the device disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0112] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A hierarchical game-based ramp merging scene intelligent vehicle brain-like decision-making method, characterized in that, The method comprises the following steps: S1: the intelligent vehicle and the main road vehicle share information; S2: an internal driving incentive model is constructed; S3: a hierarchical game optimization model is constructed based on the top-level space intelligent vehicle game decision target function and the bottom-level space main road vehicle game decision target function; S4: the hierarchical game optimization model is solved under the constructed behavior coordination constraint condition, to obtain a stable equilibrium strategy that the intelligent vehicle and the main road vehicle should implement in the current stage game; S5: the intelligent vehicle and the main road vehicle execute the respective stable equilibrium strategies; S6: S1-S5 are repeated until the intelligent vehicle safely merges into the main road from the ramp. The top-level space intelligent vehicle game decision target function and the bottom-level space main road vehicle game decision target function are: The hierarchical game optimization model is: The intelligent automobile driving internal driving incentive model Is: Wherein, W is kinematics equation, OR is ramp; is the effective distance of the intelligent vehicle in the τth round of stage game; is the component along the X direction; is the component along the Y direction; is the component along the X direction; is the component along the Y direction; is any strategy taken by the intelligent vehicle from the strategy set of the intelligent vehicle in the τth round of stage game; is the instantaneous speed of the intelligent vehicle in the τth round of stage game; is the component along the X direction; is the component along the Y direction; is the decision period of the intelligent vehicle in a stage game; τ+1 and t τ are the decision time stamp of the (τ+1)th round of stage game and the decision time stamp of the τth round of stage game respectively; is the spatial position coordinate of the intelligent vehicle when the τth round of stage game is completed; is the component along the X direction; is the component along the Y direction; is the spatial position coordinate of the intelligent vehicle when the τth round of stage game is started; is the component along the X direction; is the component along the Y direction; The main road vehicle driving intrinsic motivation model Is: wherein W is a kinematic equation, MR is a main road; is an effective distance of the main road vehicle in the τth round of stage game; is a component along the X direction; is a component along the Y direction; is a component along the X direction; is a component along the Y direction; is an arbitrary strategy of the main road vehicle from a set of game decision strategies in the τth round of stage game; is a transient speed of the main road vehicle in the τth round of stage game; is a component along the X direction; is a component along the Y direction; is a decision period of the main road vehicle in a stage game; t τ+1 and t τ are a decision time stamp of the τ+1th round of stage game and a decision time stamp of the τth round of stage game, respectively; is a spatial position coordinate of the main road vehicle when the τth round of stage game is completed; is a component along the X direction; is a component along the Y direction; is a spatial position coordinate of the main road vehicle when the τth round of stage game is started; is a component along the X direction; is a component along the Y direction; The intelligent automobile body intrinsic driving incentive model Is: wherein W represents a kinematic equation, is a component along the X direction; is a component along the Y direction; is an arbitrary strategy taken from the set of intelligent vehicle game decision strategies in the τth round of game; is a component along the X direction; is a component along the Y direction; is the instantaneous speed of the intelligent vehicle in the τth round of game; is the decision period of the intelligent vehicle in one stage of game; and are the spatial position coordinates of the intelligent vehicle at the beginning of the τth round of game; and are the spatial position coordinates of the intelligent vehicle at the completion of the τth round of game; OR is the circumscribed circle radius of the intelligent vehicle shape; MR is the circumscribed circle radius of the main road vehicle shape; The main road vehicle body intrinsic driving incentive model Is: where W represents the kinematic equation, is the component along the X direction; is the component along the Y direction; is an arbitrary strategy taken from the set of game decision strategies of the main road vehicle in the τth round of game; is the component along the X direction; is the component along the Y direction; is the instantaneous speed of the main road vehicle in the τth round of game; is the decision period of the main road vehicle in one round of game; and is the spatial position coordinate of the main road vehicle at the beginning of the τth round of game; and is the spatial position coordinate of the main road vehicle at the completion of the τth round of game; OR is the circumscribed circle radius of the intelligent vehicle shape; MR is the circumscribed circle radius of the main road vehicle shape; The intelligent automobile driver / passenger internal driving incentive model Is: wherein, is an arbitrary strategy taken by the intelligent vehicle from the set of game decision strategies in the τth round of game play, is the stable equilibrium strategy reached by the intelligent vehicle in the (τ-1)th round of game play, is the stable equilibrium strategy reached by the intelligent vehicle in the (τ-2)th round of game play, t τ-2 , t τ-1 and t τ are the decision time stamps of the (τ-2)th round of game play, the (τ-1)th round of game play and the τth round of game play, respectively; denotes the jerk caused by an arbitrary strategy taken by the intelligent vehicle in the τth round of game play ; and denotes the jerk caused by the stable equilibrium strategy reached in the (τ-1)th round of game play . The main road vehicle driver internal driving incentive model Is: wherein, is an arbitrary strategy taken by the host vehicle from the set of game decision strategies in the τth round of game, is the stable equilibrium strategy reached by the host vehicle in the (τ-1)th round of game, is the stable equilibrium strategy reached by the host vehicle in the (τ-2)th round of game, t τ-2 , t τ-1 and t τ are the decision time stamps of the (τ-2)th round of game, the (τ-1)th round of game and the τth round of game, respectively; denotes the jerk caused by an arbitrary strategy taken by the host vehicle in the τth round of game ; denotes the jerk caused by the stable equilibrium strategy reached by the host vehicle in the (τ-1)th round of game ; ​ ​ ​ ​ 2.The hierarchical game-based ramp-merging scenario intelligent car brain-like decision-making method of claim 1, wherein, ​ wherein, is a top-level space intelligent vehicle game decision objective function, is a bottom-level space main road vehicle game decision objective function, π OR ,σ OR ,μ OR is a characteristic factor of the intelligent vehicle individualized embedding module, ε MR ,δ MR ,γ MR is a characteristic factor of the main road vehicle individualized embedding module. 3.The method according to claim 2, wherein, ​ The behavior coordination constraint condition includes a top-level space intelligent vehicle behavior coordination constraint condition and a bottom-level space main road vehicle behavior coordination constraint condition and a bottom-level space main road vehicle behavior coordination constraint condition g and r are constraint condition numbers, g = 1, 2, 3…G, r = 1, 2, 3…R; G and R represent upper limits of the constraint condition numbers; wherein, and are maximum acceleration constraints for the intelligent vehicle and the host road vehicle, respectively, and are minimum acceleration constraints for the intelligent vehicle and the host road vehicle, respectively, and are road speed limit constraints for the ramp and the host road, respectively, and are minimum speed constraints for the ramp and the host road, respectively, and are maximum jerk constraints for the intelligent vehicle and the host road vehicle, respectively.

Images