Dynamic game interactive decision-making method for intelligent vehicles in mixed traffic based on subjective cognition

Abstract

The present invention discloses a mixed traffic intelligent vehicle dynamic game interactive decision-making method based on subjective cognition, including: sensing, acquiring and exchanging the driving information of each vehicle in the mixed traffic; judging whether the K instantaneous game system is started according to the situation risk assessment; if started, each vehicle participating in the game first makes a decision based on its own subjective cognition until the subjective cognition of each vehicle reaches a consensus; respectively establish and optimize the personalized embedded expected utility dynamic equation of each vehicle participating in the game, output its own acceleration at the K instant based on this decision and execute it; synchronously update the driving information and characteristic quantitative index of each vehicle in the mixed traffic at the K+1 instant; repeat the above steps until the potential mixed traffic conflict is completely resolved through continuous interactive decision-making. This method fully simulates human decision-making logic, improves the intelligence, interactivity, sustainability and integrity of decision-making, and realizes personalized driving and human-like decision-making of intelligent vehicles.

Description

Technical Field

[0001] The invention belongs to the technical field of interactive behavior decision-making of intelligent vehicles, in particular to a dynamic game interactive decision-making method of mixed traffic intelligent vehicles based on subjective cognition. Background Art

[0002] Intelligent vehicles are the forefront and inevitable trend of future automobile development. Decision-making is the difficulty and key to improving vehicle intelligence. For a long time in the future, intelligent vehicles with different intelligence levels and penetration rates will coexist with manually driven vehicles of different styles on the road, forming a mixed traffic situation. Mixed traffic has extremely strong time-varying, random and uncertain characteristics. At the same time, there is a dynamic coupling between intelligent vehicles, manually driven vehicles and environmental factors, and there is a strong interactivity.

[0003] At present, rule-based decision-making methods mostly make behavioral decisions from a macro and meso perspective, usually treating surrounding vehicles and environmental factors as fixed obstacles, lacking considerations from a micro perspective, and are unable to effectively respond to random and interactive processes in open and complex scenarios, resulting in intelligent vehicles being unable to produce driving behaviors like skilled drivers, and not truly integrating into the complex mixed traffic ecology. Limited by mixed traffic, strong interactivity, and human-like challenges, it is difficult to fully implement intelligent driving. The increasing trust, acceptance, and adaptability of drivers and passengers and vehicles participating in traffic have put forward higher requirements for intelligent technology.

[0004] It can be seen that interactive decision-making in mixed traffic that fully considers the interaction between multiple intelligent agents is a key scientific problem that needs to be solved urgently before truly realizing fully intelligent driving. Solving the problem is an important guarantee for the safe driving of intelligent vehicles. It is also beneficial to improve the safety of other vehicles participating in mixed traffic, improve the coordination and traffic efficiency of mixed traffic, and has very important practical significance. Summary of the invention

[0005] In view of the above problems, the present invention provides a dynamic game interactive decision-making method for mixed traffic intelligent vehicles based on subjective cognition, which at least solves some of the above technical problems. The method overcomes the shortcomings of existing decision-making methods based on single and fixed rules, and improves the anthropomorphism and personalization of intelligent vehicle autonomous decision-making while meeting the requirements for safe driving of intelligent vehicles in mixed traffic environments.

[0006] The embodiment of the present invention provides a dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition, including:

[0007] S1. The intelligent vehicle in mixed traffic obtains the driving information of surrounding vehicles and sends its own driving information to surrounding vehicles;

[0008] S2, based on the driving information of the intelligent vehicle and surrounding vehicles, each vehicle in the mixed traffic calculates the time difference between the K instantaneous self-vehicle and the surrounding vehicles passing through the conflict area; compares the calculation result with the risk threshold, and determines whether the game system is started according to the comparison result; if the game system is started, execute step S3;

[0009] S3. Each vehicle participating in the game first makes a decision based on its own subjective cognition until the subjective cognition of each vehicle participating in the game reaches a consensus;

[0010] S4. Establishing the expected utility dynamic equations of each vehicle participating in the game that are personalized and embedded in the intelligent agent game, and using the expected utility dynamic equations as the optimization objective function of each vehicle in the rolling time-varying decision, and outputting the acceleration of the vehicle at the instant K;

[0011] S5, the intelligent vehicle executes its own acceleration at instant K;

[0012] S6, instantaneously and synchronously updating the driving information and characteristic quantitative index of the intelligent vehicle and surrounding vehicles in mixed traffic at K+1;

[0013] S7. Repeat steps S1-S6 until the potential mixed traffic conflict is completely resolved through continuous interactive decision-making.

[0014] Furthermore, in the S1, the driving information includes: driving style information and real-time motion state information.

[0015] Furthermore, S1 also includes: when there is missing information in the driving information of the surrounding vehicles obtained by the smart vehicle, the smart vehicle infers the type of the surrounding vehicles based on the collected environmental historical information including the surrounding vehicles, and predicts the future driving status of the surrounding vehicles, and fills in the missing information based on this.

[0016] Furthermore, in S2, the time difference between the K instantaneous own vehicle and the surrounding vehicles passing through the conflict area is evaluated in the following manner:

[0017]

[0018] in, represents the time difference between vehicle m and vehicle n passing through the conflict area at K instant; v m (K) and v n (K) represents the instantaneous speed of vehicle m and vehicle n at K instants respectively; and They represent the driving distances of vehicle m and vehicle n heading to the conflict area at K instant; Indicates the width of the road; and Respectively represent the dimensions of vehicle m and vehicle n; A m (v m ) and A n (v n ) represent the enlargement factors of the dimensions of vehicle m and vehicle n, respectively, v m (K) and v n (K) function.

[0019] Furthermore, in S2, the basis for determining whether the gaming system is activated is expressed as:

[0020]

[0021] Among them, when When GT mode =ON, indicating that the vehicle-mounted gaming system is started, and the intelligent vehicle executes the output of the gaming controller; when When GT mode =OFF, indicating that the vehicle-mounted game system is not started, and each intelligent vehicle drives freely according to its expected driving mode.

[0022] Furthermore, the S3 specifically includes:

[0023] Each vehicle participating in the game defines itself as the leader, and based on its own subjective cognition, defines other vehicles participating in mixed traffic around it as followers.

[0024] Construct a feature quantification index to quantify the vehicle's driving characteristics through its external behavior, so that distributed decision-making gradually converges to integrated collaborative decision-making and achieves a stable multi-vehicle decision-making situation;

[0025] Constructing a sequential action MAP diagram, querying the instantaneously updated feature quantization index in real time from the sequential action MAP diagram, and realizing the sequential action order of the time-varying replanning stage game;

[0026] The feature quantification index is updated in real time as the rolling decision-making process progresses until the subjective cognition of each vehicle participating in the game reaches a consensus.

[0027] Furthermore, the characteristic quantification index quantifies the external behavior characteristics by using instantaneous acceleration, average acceleration, instantaneous speed, and average speed as main characteristic parameters.

[0028] Furthermore, the characteristic quantization index is expressed as:

[0029]

[0030] ||a m (K)||=δ 1 · m ​(K)>

[0031]

[0032] ||v m (K)||=δ 3 · <v m (K)>

[0033]

[0034]

[0035]

[0036] in, is the characteristic quantitative index of vehicle m; ||a m (K)|| is the instantaneous acceleration of vehicle m; is the normal form of the average acceleration of vehicle m; ||v m (K)|| is the instantaneous velocity of vehicle m; is the normal form of the average speed of vehicle m; δ 1 , δ 2 , δ 3 , δ 4 are the weight factors of each paradigm respectively; m (K)> is the standard instantaneous acceleration of vehicle m; is the standard average acceleration of vehicle m; <v m (K)> is the standard instantaneous speed of vehicle m; is the standard average speed of vehicle m; a m (i) is the instantaneous acceleration of vehicle m at instant i in the interval [0, K]; v m (i) is the instantaneous speed of vehicle m at moment i in the interval [0, K].

[0037] Furthermore, a transition area is provided in the sequential action MAP. When the difference in the characteristic quantization index between the vehicles exceeds the switching threshold, the switching of the sequential action is realized. The switching logic includes:

[0038] When the K instant decision point falls within the transition area, the position of the decision point at the K-1 instant of the previous decision step must be considered. At this time, the K instant will maintain the action sequence of the K-1 instant decision; if the K-1 instant decision point is still within the transition area, the action sequence of the K-2 instant decision will be maintained together, and so on. If the K instant decision point falls outside the transition area, then each vehicle participating in the game will make a decision according to the sequential action logic corresponding to the area where the decision point is located.

[0039] ​Furthermore, in S4, the expected utility dynamic equation includes: a safety model, an economy model and a comfort model of intelligent vehicle driving, and a coordination model and an efficiency model of a mixed traffic system.

[0040] Compared with the prior art, the dynamic game interactive decision-making method for mixed traffic intelligent vehicles based on subjective cognition recorded in the present invention has the following beneficial effects: the method fully simulates human decision-making logic, improves the intelligence, interactivity, sustainability and integrity of decision-making, and realizes personalized driving and human-like decision-making of intelligent vehicles.

[0041] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0042] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0044] Figure 1 A schematic flow chart of a method for interactive decision-making in dynamic game of hybrid traffic intelligent vehicles based on subjective cognition provided in an embodiment of the present invention.

[0045] Figure 2 Schematic diagram of the scenarios of intersection, ramp merging, and roundabout merging provided in an embodiment of the present invention.

[0046] Figure 3 A schematic diagram of the architecture of a dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition provided in an embodiment of the present invention.

[0047] Figure 4 A sequential action MAP diagram provided for an embodiment of the present invention.

[0048] Figure 5 This is a diagram of application results provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0050] See also Figure 1 As shown, the embodiment of the present invention provides a dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition, which specifically includes the following steps:

[0051] S1. The intelligent vehicle in mixed traffic obtains the driving information of surrounding vehicles and sends its own driving information to surrounding vehicles;

[0052] S2, based on the driving information of the intelligent vehicle and surrounding vehicles, each vehicle in the mixed traffic calculates the time difference between the K instantaneous self-vehicle and the surrounding vehicles passing through the conflict area; compares the calculation result with the risk threshold, and determines whether the game system is started according to the comparison result; if the game system is started, execute step S3;

[0053] S3. Each vehicle participating in the game first makes a decision based on its own subjective cognition until the subjective cognition of each vehicle participating in the game reaches a consensus;

[0054] S4. Establishing the expected utility dynamic equations of each vehicle participating in the game that are personalized and embedded in the intelligent agent game, and using the expected utility dynamic equations as the optimization objective function of each vehicle in the rolling time-varying decision, and outputting the acceleration of the vehicle at the instant K;

[0055] S5, the intelligent vehicle executes its own acceleration at instant K;

[0056] S6, instantaneously and synchronously updating the driving information and characteristic quantitative index of the intelligent vehicle and surrounding vehicles in mixed traffic at K+1;

[0057] S7. Repeat steps S1-S6 until the potential mixed traffic conflict is completely resolved through continuous interactive decision-making.

[0058] Next, each of the above steps will be described in detail.

[0059] In the above step S1, see Figure 2As shown in the figure, when there is an intersection area in the future driving trajectory of vehicles in a mixed traffic environment, the intelligent vehicle in the mixed traffic obtains the driving information of surrounding vehicles through the on-board sensing device or communication device, and sends its own driving information to the surrounding vehicles; the surrounding vehicles include other intelligent vehicles and manually driven vehicles in the mixed traffic environment; the driving information includes driving style information and real-time motion status information. When the information collected by the on-board sensing device or communication device is missing, the intelligent vehicle infers the type of surrounding vehicles based on the collected environmental historical information including surrounding vehicles, and predicts the future driving status of surrounding vehicles, and fills in the missing information based on the understanding of the environment.

[0060] In the above step S2, based on the driving information of the intelligent vehicle and the surrounding vehicles obtained above, each vehicle in the mixed traffic calculates the time difference between the K instantaneous self-vehicle and the surrounding vehicles passing through the conflict area respectively; the time difference between the K instantaneous self-vehicle and the surrounding vehicles passing through the conflict area respectively is evaluated in the following way:

[0061]

[0062] in, represents the time difference between vehicle m and vehicle n passing through the conflict area at K instant; v m (K) and v n (K) represents the instantaneous speed of vehicle m and vehicle n at K instants respectively; and They represent the driving distances of vehicle m and vehicle n heading to the conflict area at K instant; Indicates the width of the road; and Respectively represent the dimensions of vehicle m and vehicle n; A m (v m ) and A n (v n ) represent the enlargement factors of the dimensions of vehicle m and vehicle n, respectively, and are the instantaneous speed v m (K) and v n (K) is a function that is positively correlated and is a dimensionless index;

[0063] According to the above calculation, the time difference between the vehicle and the surrounding vehicles passing through the conflict area is With the risk threshold τ d Make a comparison; judge the game system GT based on the comparison results mode Whether to start; the judgment basis is expressed by the following formula:

[0064]

[0065] Among them, when When GTmode =ON, indicating that the vehicle-mounted gaming system is started, and the intelligent vehicle executes the output of the gaming controller; when When GT mode =OFF, indicating that the vehicle-mounted game system is not started, and each intelligent vehicle drives freely according to its expected driving mode.

[0066] In the above step S3, after the game system is started, based on the inherent psychological cognitive mechanism of the real driving process, the embodiment of the present invention proposes that each vehicle first carries out distributed decision-making based on its own subjective cognition. In the master-slave sequential behavior mode of the dynamic game, the intelligent vehicle defines itself as the leader's position, and defines other participating vehicles in the surrounding mixed traffic as followers' positions based on its own subjective cognition. At the same time, other participating vehicles in the mixed traffic also define themselves as the leader's position, and define the surrounding vehicles as followers' positions based on their own subjective cognition. In the above content, the leader refers to the vehicle that takes action first in the dynamic game; the follower refers to the vehicle that follows the action in the dynamic game.

[0067] In order to accurately describe the gradual process of evolutionary decision-making, the embodiment of the present invention proposes a characteristic quantification index, which quantifies the driving characteristics of the vehicle through the external behavior of the vehicle, and adopts the above-mentioned internal psychological cognition and external behavior performance The way is combined, the distributed decision-making gradually converges to an integrated collaborative decision-making, and achieves a stable multi-vehicle decision-making situation. Figure 3 shown.

[0068] Based on the feature quantization index proposed in the embodiment of the present invention, in order to clarify the process of evolution from distributed decision-making to integrated collaborative decision-making, the embodiment of the present invention further proposes a sequential action MAP diagram. In the constructed MAP diagram, the instantaneously updated feature quantization index is queried in real time to change the sequential action order of the game in the re-planning stage. The sequential action MAP diagram is as follows: Figure 4 shown.

[0069] The characteristic quantification index proposed in the above embodiment of the present invention is calculated in the following way, selecting instantaneous acceleration, average acceleration, instantaneous speed, and average speed as main characteristic parameters to quantify external behavior characteristics:

[0070]

[0071] ||a m (K)||=δ 1 · m (K)>

[0072]

[0073] ||v m (K)||=δ​3 · <v m (K)>

[0074]

[0075]

[0076]

[0077] in, is the characteristic quantitative index of vehicle m; ||a m (K)|| is the instantaneous acceleration of vehicle m; is the normal form of the average acceleration of vehicle m; ||v m (K)|| is the instantaneous velocity of vehicle m; is the normal form of the average speed of vehicle m; δ 1 , δ 2 , δ 3 , δ 4 are the weight factors of each paradigm respectively. In order to avoid the influence of different value ranges and different dimensions on different feature parameters, it is necessary to standardize the deviation of each feature parameter and linearly transform the original feature parameter data to the interval [0, 1]; m (K)> is the standard instantaneous acceleration of vehicle m; is the standard average acceleration of vehicle m; <v m (K)> is the standard instantaneous speed of vehicle m; is the standard average speed of vehicle m; a m (i) is the instantaneous acceleration of vehicle m at instant i in the interval [0, K]; v m (i) is the instantaneous speed of vehicle m at instant i in the interval [0, K]; the standardized calculation formula is as follows:

[0078]

[0079]

[0080]

[0081]

[0082] in, They represent the maximum values ​​of instantaneous acceleration, average acceleration, instantaneous velocity, and average velocity, respectively. They represent the minimum values ​​of instantaneous acceleration, average acceleration, instantaneous velocity, and average velocity respectively. ​

[0083] The sequential action MAP diagram proposed in the above-mentioned embodiment of the present invention has coordinates that are the characteristic quantitative index of the intelligent vehicle and the characteristic quantitative index of other traffic participating vehicles participating in the game in mixed traffic. In order to avoid frequent switching of the sequential action logic caused by directly comparing the characteristic quantitative indexes of each game participating vehicle, which has an adverse effect on the coordination of mixed traffic, the embodiment of the present invention specifically sets a transition area in the sequential action MAP diagram. When the difference in the characteristic quantitative index between the vehicles exceeds the switching threshold, the switching of sequential actions is realized. The switching logic is shown in the following formula. In the MAP diagram, when the K instantaneous decision point falls within the transition area, the position of the decision point of the previous decision step at the K-1 instantaneous moment must be considered. At this time, the K instantaneous moment will maintain the action sequence of the K-1 instantaneous decision; if the K-1 instantaneous decision point is still within the transition area, the action sequence of the K-2 instantaneous decision will be maintained together, and so on; if the decision point of the K instantaneous moment falls outside the transition area, then each game participating vehicle will make a decision according to the sequential action logic corresponding to the area where the decision point is located. Specifically, as Figure 4 As shown, if falls within region I, then the dynamic game follows The action sequence is carried out; if falls within region II, then the dynamic game follows The action sequence is carried out; if falls within the transition region III, and K-1 instantaneous falls within the transition region I, then the dynamic game remains The action sequence does not switch; if falls within the transition region III, and K-1 instantaneously falls within transition region II, then the dynamic game remains The action sequence does not switch; if falls within the transition region III, and K-1 instantaneously It also falls into transition zone III. At this time, it is necessary to continue to trace back to K-2 instant The area where the dynamic game is located is used to determine the type of action sequence that the dynamic game will maintain, and so on.

[0084]

[0085]

[0086] in, Represents the characteristic quantitative index of vehicle m; Represents the characteristic quantitative index of vehicle n; F π Indicates the switching threshold; when When , it means that K instantaneous vehicle makes distributed decisions based on its own subjective cognition and does not form an integrated collaborative decision-making situation; when When , it means that K needs to switch the sequential action instantly and initially form an instantaneous integrated collaborative decision-making situation. The specific switching direction and evolution process need to be determined in combination with the information of the above sequential action MAP diagram.

[0087] In the above step S4, in the process of each vehicle participating in the game interactive decision-making, the embodiment of the present invention respectively establishes a personalized embedded expected utility dynamic equation for each vehicle participating in the game in the intelligent agent game; specifically, combining the driving needs of the intelligent vehicle and the performance requirements of the mixed traffic, a model to ensure the safety, economy and comfort of the intelligent vehicle driving, as well as a model of the coordination and efficiency of the mixed traffic system is established; each vehicle considers the strategy set of other game participating vehicles and the predicted results of their possible response strategies in the future, and uses the established expected utility dynamic equation as the optimization objective function of each vehicle in the rolling time-varying decision; optimizes its own personalized embedded expected utility dynamic equation, and decides to output the acceleration of the intelligent vehicle itself at the instant K.

[0088] The above-mentioned expected utility dynamic equation takes a single intelligent vehicle as an example. The embodiment of the present invention comprehensively considers the instantaneous motion states of the K instantaneous vehicle and surrounding vehicles in mixed traffic, as well as the specific actions taken by each from its game strategy set, and incorporates the psychological safety acceptance closely related to the personalized driving style to establish a safety model; the safety model is expressed as follows:

[0089]

[0090]

[0091]

[0092]

[0093] in, and They represent the remaining time for K instantaneous vehicles m and n to reach the conflict area based on specific game strategies; and K represent the driving distances of vehicle m and vehicle n towards the conflict area at the instant; a m (K) and a n (K) represents the instantaneous acceleration of vehicle m and vehicle n at K; v m (K) and v n (K) represents the instantaneous speed of vehicle m and vehicle n at K instants respectively; represents the psychological safety acceptance of vehicle m; λ, μ, ξ represent the aggressiveness factors of personalized intelligent vehicles respectively.

[0094] The above expected utility dynamic equation takes a single smart vehicle as an example. In order to improve the adaptability of the method to traditional fuel vehicles and new energy vehicles, the embodiment of the present invention establishes an economic model to punish repeated acceleration and braking behaviors in the continuous interactive decision-making process, so as to improve the driving economy of the smart vehicle and reduce energy consumption; the economic model is expressed as follows:

[0095]

[0096] Among them, a m (K-1) represents the instantaneous acceleration of vehicle m at K-1; a m (K) represents the instantaneous acceleration of vehicle m at instant K.

[0097] The above-mentioned expected utility dynamic equation takes a single intelligent vehicle as an example. In order to improve the smoothness of vehicle driving and improve driving comfort, the comfort model established in the embodiment of the present invention is as follows:

[0098] CO m =-|a m (K)-a m (K-1)|

[0099] Among them, a m (K) represents the instantaneous acceleration of vehicle m at the instant K; a m (K-1) represents the instantaneous acceleration of vehicle m at instant K-1.

[0100] The above expected utility dynamic equation is also from the perspective of mixed traffic. In order to improve the coordination of the mixed traffic system and avoid frequent switching of the action sequence of the aforementioned game decision, the mixed traffic coordination model established in the embodiment of the present invention is as follows:

[0101]

[0102] in, represents the sequential function of K-1 instantaneous vehicles m and n; represents the sequential function of K instantaneous vehicles m and n.

[0103] The above-mentioned expected utility dynamic equation is also from the perspective of mixed traffic. In order to improve the efficiency of the mixed traffic system and comprehensively improve the traffic efficiency of the mixed traffic system, the mixed traffic efficiency model established in the embodiment of the present invention is as follows:

[0104]

[0105]

[0106]

[0107] in, and K represent the driving distances of vehicle m and vehicle n towards the conflict area at the instant; a m (K) and a n (K) represents the instantaneous acceleration of vehicle m and vehicle n at K; v m (K) and v n (K) represents the instantaneous speed of vehicle m and vehicle n at K instants respectively.

[0108] The safety, economy, comfort of the above-mentioned intelligent vehicle driving and the coordination and efficiency of the mixed traffic system are combined to form the expected utility dynamic equation of each intelligent agent in the interactive decision-making process of the game, which serves as the optimization objective function of each vehicle in the rolling time-varying decision. The above-mentioned performances constrain and complement each other. The expected utility dynamic equation of each game-participating vehicle is related to its own driving state and driving style, and is also affected by other traffic participating vehicles in the mixed traffic. In the continuous rolling optimization, each vehicle reaches the Nash equilibrium of the dynamic game and gradually converges to the above-mentioned integrated collaborative decision-making situation.

[0109] In the above step S5, the bottom control module of the intelligent vehicle controls the brake force and the throttle opening to execute the acceleration that the intelligent vehicle itself should execute at the instant K as output by the upper decision module K.

[0110] In the above step S6, the motion information, prediction information, cognitive information and feature quantization index of the intelligent vehicle and surrounding vehicles in the mixed traffic are synchronously updated at K+1 instant;

[0111] In the above step S7, steps S1-S6 are repeatedly executed until the potential mixed traffic conflict is completely resolved through continuous interactive decision-making.

[0112] In order to effectively solve the problem of safe passage of intelligent vehicles in mixed traffic environments, while fully considering the interactivity of behavioral decisions and deeply exploring the subjective cognition of intelligent agent decisions, the embodiment of the present invention provides a mixed traffic intelligent vehicle dynamic game interactive decision-making method based on subjective cognition, which is used in driving scenarios involving strong interactions of multiple intelligent agents such as intersections, ramp merging, and roundabout merging. The application result diagram of this method can be found in Figure 5As shown in the figure, this method fully understands the driving behavior of skilled drivers from the perspective of internal psychological cognition and external behavioral performance, and interacts with other vehicles participating in mixed traffic in a human-like manner to coordinate and resolve potential traffic conflicts and clarify the right of way for mixed traffic. It overcomes the shortcomings of existing decision-making methods based on single and fixed rules, and improves the intelligence, interactivity, sustainability and integrity of decision-making while meeting the requirements for safe driving of intelligent vehicles in mixed traffic environments, thus realizing personalized driving and human-like decision-making of intelligent vehicles.

[0113] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. Dynamic game interactive decision-making method for mixed traffic intelligent vehicles based on subjective cognition, It is characterized in that include: S1. The intelligent vehicle in mixed traffic obtains the driving information of surrounding vehicles and sends its own driving information to surrounding vehicles; S2, based on the driving information of the intelligent vehicle and surrounding vehicles, each vehicle in the mixed traffic calculates the time difference between the K instantaneous self-vehicle and the surrounding vehicles passing through the conflict area; compares the calculation result with the risk threshold, and determines whether the game system is started according to the comparison result; if the game system is started, execute step S3; S3. Each vehicle participating in the game first makes a decision based on its own subjective cognition until the subjective cognition of each vehicle participating in the game reaches a consensus; S4. Establishing the expected utility dynamic equations of each vehicle participating in the game that are personalized and embedded in the intelligent agent game, and using the expected utility dynamic equations as the optimization objective function of each vehicle in the rolling time-varying decision, and outputting the acceleration of the vehicle at the instant K; S5, the intelligent vehicle executes its own acceleration at instant K; S6, instantaneously and synchronously updating the driving information and characteristic quantitative index of the intelligent vehicle and surrounding vehicles in mixed traffic at K+1; S7, repeating steps S1-S6 until the potential mixed traffic conflict is completely resolved through continuous interactive decision-making; The S3 specifically includes: Each vehicle participating in the game defines itself as the leader, and based on its own subjective cognition, defines other vehicles participating in mixed traffic around it as followers. Construct a feature quantification index to quantify the vehicle's driving characteristics through its external behavior, so that distributed decision-making gradually converges to integrated collaborative decision-making and achieves a stable multi-vehicle decision-making situation; Constructing a sequential action MAP diagram, querying the instantaneously updated feature quantization index in real time from the sequential action MAP diagram, and realizing the sequential action order of the time-varying replanning stage game; The feature quantification index is updated in real time as the rolling decision-making process progresses until the subjective cognition of each vehicle participating in the game reaches a consensus.

2. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that In the step S1 , the driving information includes driving style information and real-time motion status information.

3. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that The S1 also includes: when there is missing information in the driving information of the surrounding vehicles obtained by the smart vehicle, the smart vehicle infers the type of the surrounding vehicles based on the collected environmental historical information including the surrounding vehicles, and predicts the future driving status of the surrounding vehicles, and fills in the missing information based on this.

4. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that In S2, the time difference between the K instantaneous own vehicle and the surrounding vehicles passing through the conflict area is evaluated in the following way: in, represents the time difference between vehicle m and vehicle n passing through the conflict area at K instant; v m (K) and v n (K) represents the instantaneous speed of vehicle m and vehicle n at K instants respectively; and They represent the driving distances of vehicle m and vehicle n heading to the conflict area at K instant; Indicates the width of the road; and Respectively represent the dimensions of vehicle m and vehicle n; A m (v m ) and A n (v n ) represent the enlargement factors of the dimensions of vehicle m and vehicle n, respectively, v m (K) and v n (K) function.

5. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 4, It is characterized in that In S2, according to With the risk threshold τ d The comparison results are used to judge the game system GT mode Whether to start or not, the judgment basis is expressed as: Among them, when When GT mode =ON, indicating that the vehicle-mounted gaming system is started, and the intelligent vehicle executes the output of the gaming controller; when When GT mode =OFF, indicating that the vehicle-mounted game system is not started, and each intelligent vehicle drives freely according to its expected driving mode.

6. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that The characteristic quantification index quantifies the external behavior characteristics by using instantaneous acceleration, average acceleration, instantaneous speed, and average speed as characteristic parameters.

7. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that The characteristic quantization index is expressed as: ||a m (K)||=δ 1 · m (K)>​ ||v m (K)||=δ 3 · <v m (K)> in, is the characteristic quantitative index of vehicle m; ||a m (K)|| is the instantaneous acceleration of vehicle m; is the normal form of the average acceleration of vehicle m; ||v m (K)|| is the instantaneous velocity of vehicle m; is the normal form of the average speed of vehicle m; δ 1 ,δ 2 ,δ 3 ,δ 4 are the weight factors of each paradigm respectively; m (K)> is the standard instantaneous acceleration of vehicle m; is the standard average acceleration of vehicle m; <v m (K)> is the standard instantaneous speed of vehicle m; is the standard average speed of vehicle m; a m (i) is the instantaneous acceleration of vehicle m at instant i in the interval [0, K]; v m (i) is the instantaneous speed of vehicle m at moment i in the interval [0, K].​ 8. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that The sequential action MAP is provided with a transition area. When the difference in the characteristic quantization index between the vehicles exceeds the switching threshold, the sequential action switching is realized. The switching logic includes: When the K instant decision point falls within the transition area, the position of the decision point at the K-1 instant of the previous decision step must be considered. At this time, the K instant will maintain the action sequence of the K-1 instant decision; if the K-1 instant decision point is still within the transition area, the action sequence of the K-2 instant decision will be maintained together, and so on. If the K instant decision point falls outside the transition area, then each vehicle participating in the game will make a decision according to the sequential action logic corresponding to the area where the decision point is located.

9. The dynamic game interactive decision-making method for hybrid traffic intelligent vehicles based on subjective cognition as claimed in claim 1, It is characterized in that In S4, the expected utility dynamic equation includes: a safety model, an economy model and a comfort model of intelligent vehicle driving, and a coordination model and an efficiency model of a mixed traffic system.

Images